Hugging Face's logo

Datasets:
AriaAICompany
/
dlp-lab

Tasks:
Token Classification
Other
Languages:
English
Persian
Size:
n<1K
Tags:
cybersecurity
pii-detection
redaction
document-analysis
Synthetic
License:
Dataset card Files
xet
 Community
alirezaaminzadeh commited on
Commit
eef4eed
·
verified ·
1 Parent(s): 9878e03

Publish AriaDLP synthetic Persian/English redaction lab

Browse files
Files changed (50) hide show
  1. README.md +49 -0
  2. data/documents.jsonl +80 -0
  3. data/evaluation.json +88 -0
  4. data/feature_card.md +8 -0
  5. data/gold_spans.json +2662 -0
  6. data/protocol.md +23 -0
  7. data/sample_preview.json +88 -0
  8. data/scenarios.jsonl +5 -0
  9. data/schema.json +25 -0
  10. data/splits.json +17 -0
  11. demo/.gitattributes +11 -0
  12. demo/.gitignore +8 -0
  13. demo/LICENSE +17 -0
  14. demo/README.md +55 -0
  15. demo/app.py +8 -0
  16. demo/evaluations/evaluation.json +88 -0
  17. demo/examples/documents.jsonl +80 -0
  18. demo/examples/evaluation.json +88 -0
  19. demo/examples/feature_card.md +8 -0
  20. demo/examples/gold_spans.json +2662 -0
  21. demo/examples/protocol.md +23 -0
  22. demo/examples/sample_preview.json +88 -0
  23. demo/examples/scenarios.jsonl +5 -0
  24. demo/examples/schema.json +25 -0
  25. demo/examples/splits.json +17 -0
  26. demo/requirements.txt +4 -0
  27. demo/src/__init__.py +1 -0
  28. demo/src/analyzer.py +47 -0
  29. demo/src/baseline.py +48 -0
  30. demo/src/checksums.py +106 -0
  31. demo/src/domain.py +122 -0
  32. demo/src/explain.py +68 -0
  33. demo/src/generator.py +462 -0
  34. demo/src/i18n.py +143 -0
  35. demo/src/metrics.py +212 -0
  36. demo/src/normalize.py +28 -0
  37. demo/src/parsers.py +62 -0
  38. demo/src/pdfio.py +99 -0
  39. demo/src/pipeline.py +325 -0
  40. demo/src/plots.py +101 -0
  41. demo/src/recognizers.py +303 -0
  42. demo/src/redaction.py +76 -0
  43. demo/src/reporting.py +70 -0
  44. demo/src/schemas.py +175 -0
  45. demo/src/ui.py +315 -0
  46. demo/src/validation.py +42 -0
  47. demo/tests/conftest.py +8 -0
  48. demo/tests/test_pipeline.py +159 -0
  49. showcase/README.md +18 -0
  50. showcase/index.html +64 -0
README.md ADDED
@@ -0,0 +1,49 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: cc-by-4.0
3
+ pretty_name: AriaDLP Persian/English Redaction Lab
4
+ task_categories:
5
+ - token-classification
6
+ - other
7
+ tags:
8
+ - cybersecurity
9
+ - pii-detection
10
+ - redaction
11
+ - document-analysis
12
+ - synthetic
13
+ language:
14
+ - en
15
+ - fa
16
+ size_categories:
17
+ - n<1K
18
+ ---
19
+
20
+ # AriaDLP Persian / English redaction lab
21
+
22
+ Synthetic 80-document bilingual corpus for the AriaDLP workstation: text, CSV, and generated text PDFs with gold character spans. Split is by **document template**, not random rows. Regex-only is the baseline. Policy uses checksums, Persian digit folding, and a closed name lexicon.
23
+
24
+ Organization dataset and collection are public. Live Gradio is [alirezaaminzadeh/aria-dlp](https://huggingface.co/spaces/alirezaaminzadeh/aria-dlp) and the organization card is [AriaAICompany/aria-dlp](https://huggingface.co/spaces/AriaAICompany/aria-dlp). Runnable Space source is stored in `demo/`. Collection: [Aria AI — Cybersecurity](https://huggingface.co/collections/AriaAICompany/aria-ai-cybersecurity-6ab01faa3995a3bb737b58a7).
25
+
26
+ Generated data, seed 24. Mailboxes use `aria-lab.test`. PANs are Luhn-valid laboratory numbers. Names come from a closed lexicon (`Labian`, `Fixture`, `نمونه‌ای`, `آزمایشی`). No real identity records are included. Passing the suite is not a claim of complete production redaction.
27
+
28
+ ## Files
29
+
30
+ - data/documents.jsonl
31
+ - data/gold_spans.json
32
+ - data/scenarios.jsonl
33
+ - data/schema.json
34
+ - data/splits.json
35
+ - data/protocol.md
36
+ - data/feature_card.md
37
+ - data/evaluation.json
38
+ - data/sample_preview.json
39
+ - demo/ — runnable Gradio Space source
40
+ - showcase/ — organization static card
41
+
42
+ ## Split
43
+
44
+ Development templates: English support email, Persian HR letter, invoice CSV, onboarding worksheet.
45
+ Test templates: Persian contract, PDF memo, hard-negative catalog, mixed ticket. Rewriting one document is not placed across the split.
46
+
47
+ ## License
48
+
49
+ CC-BY-4.0. Keep the synthetic-data and laboratory-label notices.
data/documents.jsonl ADDED
@@ -0,0 +1,80 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {"doc_id":"DLP-EN-EMAIL-01","template":"en_support_email","family":"en_support_email","split":"development","language":"en","format":"text","title":"English support mailbox","body":"From: Nima Fixture <Nima.Fixture01@aria-lab.test>\nPhone: 09124063383\nEmployee: ARIA-24001\nPlease reset the laboratory mailbox for Nima. Ticket ORD-2024-0001. Version 2.4.1.\n","metadata":{},"gold_spans":[{"start":6,"end":18,"entity_type":"PERSON","surface":"Nima Fixture","language":"en","field":null},{"start":20,"end":48,"entity_type":"EMAIL","surface":"Nima.Fixture01@aria-lab.test","language":"en","field":null},{"start":57,"end":68,"entity_type":"PHONE","surface":"09124063383","language":"en","field":null},{"start":79,"end":89,"entity_type":"EMPLOYEE_ID","surface":"ARIA-24001","language":"en","field":null}],"gold_meta":[],"hard_negatives":["ORD-2024-0001","2.4.1"],"description":"English support email with name, mailbox, phone, and employee id.","description_fa":"ایمیل پشتیبانی انگلیسی با نام، صندوق، تلفن و شناسه کارمند."}
2
+ {"doc_id":"DLP-EN-EMAIL-02","template":"en_support_email","family":"en_support_email","split":"development","language":"en","format":"text","title":"English support mailbox","body":"From: Noah Labian <Noah.Labian02@aria-lab.test>\nPhone: 09124256381\nEmployee: ARIA-24002\nPlease reset the laboratory mailbox for Noah. Ticket ORD-2024-0002. Version 2.4.1.\n","metadata":{},"gold_spans":[{"start":6,"end":17,"entity_type":"PERSON","surface":"Noah Labian","language":"en","field":null},{"start":19,"end":46,"entity_type":"EMAIL","surface":"Noah.Labian02@aria-lab.test","language":"en","field":null},{"start":55,"end":66,"entity_type":"PHONE","surface":"09124256381","language":"en","field":null},{"start":77,"end":87,"entity_type":"EMPLOYEE_ID","surface":"ARIA-24002","language":"en","field":null}],"gold_meta":[],"hard_negatives":["ORD-2024-0002","2.4.1"],"description":"English support email with name, mailbox, phone, and employee id.","description_fa":"ایمیل پشتیبانی انگلیسی با نام، صندوق، تلفن و شناسه کارمند."}
3
+ {"doc_id":"DLP-EN-EMAIL-03","template":"en_support_email","family":"en_support_email","split":"development","language":"en","format":"text","title":"English support mailbox","body":"From: Noah Labian <Noah.Labian03@aria-lab.test>\nPhone: 09123538397\nEmployee: ARIA-24003\nPlease reset the laboratory mailbox for Noah. Ticket ORD-2024-0003. Version 2.4.1.\n","metadata":{},"gold_spans":[{"start":6,"end":17,"entity_type":"PERSON","surface":"Noah Labian","language":"en","field":null},{"start":19,"end":46,"entity_type":"EMAIL","surface":"Noah.Labian03@aria-lab.test","language":"en","field":null},{"start":55,"end":66,"entity_type":"PHONE","surface":"09123538397","language":"en","field":null},{"start":77,"end":87,"entity_type":"EMPLOYEE_ID","surface":"ARIA-24003","language":"en","field":null}],"gold_meta":[],"hard_negatives":["ORD-2024-0003","2.4.1"],"description":"English support email with name, mailbox, phone, and employee id.","description_fa":"ایمیل پشتیبانی انگلیسی با نام، صندوق، تلفن و شناسه کارمند."}
4
+ {"doc_id":"DLP-EN-EMAIL-04","template":"en_support_email","family":"en_support_email","split":"development","language":"en","format":"text","title":"English support mailbox","body":"From: Nima Fixture <Nima.Fixture04@aria-lab.test>\nPhone: 09121213266\nEmployee: ARIA-24004\nPlease reset the laboratory mailbox for Nima. Ticket ORD-2024-0004. Version 2.4.1.\n","metadata":{},"gold_spans":[{"start":6,"end":18,"entity_type":"PERSON","surface":"Nima Fixture","language":"en","field":null},{"start":20,"end":48,"entity_type":"EMAIL","surface":"Nima.Fixture04@aria-lab.test","language":"en","field":null},{"start":57,"end":68,"entity_type":"PHONE","surface":"09121213266","language":"en","field":null},{"start":79,"end":89,"entity_type":"EMPLOYEE_ID","surface":"ARIA-24004","language":"en","field":null}],"gold_meta":[],"hard_negatives":["ORD-2024-0004","2.4.1"],"description":"English support email with name, mailbox, phone, and employee id.","description_fa":"ایمیل پشتیبانی انگلیسی با نام، صندوق، تلفن و شناسه کارمند."}
5
+ {"doc_id":"DLP-EN-EMAIL-05","template":"en_support_email","family":"en_support_email","split":"development","language":"en","format":"text","title":"English support mailbox","body":"From: Omar Fixture <Omar.Fixture05@aria-lab.test>\nPhone: 09122939341\nEmployee: ARIA-24005\nPlease reset the laboratory mailbox for Omar. Ticket ORD-2024-0005. Version 2.4.1.\n","metadata":{},"gold_spans":[{"start":6,"end":18,"entity_type":"PERSON","surface":"Omar Fixture","language":"en","field":null},{"start":20,"end":48,"entity_type":"EMAIL","surface":"Omar.Fixture05@aria-lab.test","language":"en","field":null},{"start":57,"end":68,"entity_type":"PHONE","surface":"09122939341","language":"en","field":null},{"start":79,"end":89,"entity_type":"EMPLOYEE_ID","surface":"ARIA-24005","language":"en","field":null}],"gold_meta":[],"hard_negatives":["ORD-2024-0005","2.4.1"],"description":"English support email with name, mailbox, phone, and employee id.","description_fa":"ایمیل پشتیبانی انگلیسی با نام، صندوق، تلفن و شناسه کارمند."}
6
+ {"doc_id":"DLP-EN-EMAIL-06","template":"en_support_email","family":"en_support_email","split":"development","language":"en","format":"text","title":"English support mailbox","body":"From: Ava Labian <Ava.Labian06@aria-lab.test>\nPhone: 09129242814\nEmployee: ARIA-24006\nPlease reset the laboratory mailbox for Ava. Ticket ORD-2024-0006. Version 2.4.1.\n","metadata":{},"gold_spans":[{"start":6,"end":16,"entity_type":"PERSON","surface":"Ava Labian","language":"en","field":null},{"start":18,"end":44,"entity_type":"EMAIL","surface":"Ava.Labian06@aria-lab.test","language":"en","field":null},{"start":53,"end":64,"entity_type":"PHONE","surface":"09129242814","language":"en","field":null},{"start":75,"end":85,"entity_type":"EMPLOYEE_ID","surface":"ARIA-24006","language":"en","field":null}],"gold_meta":[],"hard_negatives":["ORD-2024-0006","2.4.1"],"description":"English support email with name, mailbox, phone, and employee id.","description_fa":"ایمیل پشتیبانی انگلیسی با نام، صندوق، تلفن و شناسه کارمند."}
7
+ {"doc_id":"DLP-EN-EMAIL-07","template":"en_support_email","family":"en_support_email","split":"development","language":"en","format":"text","title":"English support mailbox","body":"From: Nima Fixture <Nima.Fixture07@aria-lab.test>\nPhone: 09126110356\nEmployee: ARIA-24007\nPlease reset the laboratory mailbox for Nima. Ticket ORD-2024-0007. Version 2.4.1.\n","metadata":{},"gold_spans":[{"start":6,"end":18,"entity_type":"PERSON","surface":"Nima Fixture","language":"en","field":null},{"start":20,"end":48,"entity_type":"EMAIL","surface":"Nima.Fixture07@aria-lab.test","language":"en","field":null},{"start":57,"end":68,"entity_type":"PHONE","surface":"09126110356","language":"en","field":null},{"start":79,"end":89,"entity_type":"EMPLOYEE_ID","surface":"ARIA-24007","language":"en","field":null}],"gold_meta":[],"hard_negatives":["ORD-2024-0007","2.4.1"],"description":"English support email with name, mailbox, phone, and employee id.","description_fa":"ایمیل پشتیبانی انگلیسی با نام، صندوق، تلفن و شناسه کارمند."}
8
+ {"doc_id":"DLP-EN-EMAIL-08","template":"en_support_email","family":"en_support_email","split":"development","language":"en","format":"text","title":"English support mailbox","body":"From: Omar Labian <Omar.Labian08@aria-lab.test>\nPhone: 09125259879\nEmployee: ARIA-24008\nPlease reset the laboratory mailbox for Omar. Ticket ORD-2024-0008. Version 2.4.1.\n","metadata":{},"gold_spans":[{"start":6,"end":17,"entity_type":"PERSON","surface":"Omar Labian","language":"en","field":null},{"start":19,"end":46,"entity_type":"EMAIL","surface":"Omar.Labian08@aria-lab.test","language":"en","field":null},{"start":55,"end":66,"entity_type":"PHONE","surface":"09125259879","language":"en","field":null},{"start":77,"end":87,"entity_type":"EMPLOYEE_ID","surface":"ARIA-24008","language":"en","field":null}],"gold_meta":[],"hard_negatives":["ORD-2024-0008","2.4.1"],"description":"English support email with name, mailbox, phone, and employee id.","description_fa":"ایمیل پشتیبانی انگلیسی با نام، صندوق، تلفن و شناسه کارمند."}
9
+ {"doc_id":"DLP-EN-EMAIL-09","template":"en_support_email","family":"en_support_email","split":"development","language":"en","format":"text","title":"English support mailbox","body":"From: Helen Labian <Helen.Labian09@aria-lab.test>\nPhone: 09126483978\nEmployee: ARIA-24009\nPlease reset the laboratory mailbox for Helen. Ticket ORD-2024-0009. Version 2.4.1.\n","metadata":{},"gold_spans":[{"start":6,"end":18,"entity_type":"PERSON","surface":"Helen Labian","language":"en","field":null},{"start":20,"end":48,"entity_type":"EMAIL","surface":"Helen.Labian09@aria-lab.test","language":"en","field":null},{"start":57,"end":68,"entity_type":"PHONE","surface":"09126483978","language":"en","field":null},{"start":79,"end":89,"entity_type":"EMPLOYEE_ID","surface":"ARIA-24009","language":"en","field":null}],"gold_meta":[],"hard_negatives":["ORD-2024-0009","2.4.1"],"description":"English support email with name, mailbox, phone, and employee id.","description_fa":"ایمیل پشتیبانی انگلیسی با نام، صندوق، تلفن و شناسه کارمند."}
10
+ {"doc_id":"DLP-EN-EMAIL-10","template":"en_support_email","family":"en_support_email","split":"development","language":"en","format":"text","title":"English support mailbox","body":"From: Leila Labian <Leila.Labian10@aria-lab.test>\nPhone: 09127157194\nEmployee: ARIA-24010\nPlease reset the laboratory mailbox for Leila. Ticket ORD-2024-0010. Version 2.4.1.\n","metadata":{},"gold_spans":[{"start":6,"end":18,"entity_type":"PERSON","surface":"Leila Labian","language":"en","field":null},{"start":20,"end":48,"entity_type":"EMAIL","surface":"Leila.Labian10@aria-lab.test","language":"en","field":null},{"start":57,"end":68,"entity_type":"PHONE","surface":"09127157194","language":"en","field":null},{"start":79,"end":89,"entity_type":"EMPLOYEE_ID","surface":"ARIA-24010","language":"en","field":null}],"gold_meta":[],"hard_negatives":["ORD-2024-0010","2.4.1"],"description":"English support email with name, mailbox, phone, and employee id.","description_fa":"ایمیل پشتیبانی انگلیسی با نام، صندوق، تلفن و شناسه کارمند."}
11
+ {"doc_id":"DLP-FA-HR-01","template":"fa_hr_letter","family":"fa_hr_letter","split":"development","language":"fa","format":"text","title":"Persian HR letter","body":"خانم سارا نمونه‌ای،\nکد‌ملی ۵۵۱۴۱۳۹۲۷۱\nتلفن همراه ۰۹۱۲۵۷۶۷۰۹۶\nایمیل user.01@aria-lab.test\nشماره پیگیری INV-1403-0001 و نسخه ۳٫۲٫۱ فقط ارجاع داخلی‌اند.\n","metadata":{},"gold_spans":[{"start":5,"end":18,"entity_type":"PERSON","surface":"سارا نمونه‌ای","language":"fa","field":null},{"start":27,"end":37,"entity_type":"IR_NATIONAL_ID","surface":"۵۵۱۴۱۳۹۲۷۱","language":"fa","field":null},{"start":49,"end":60,"entity_type":"PHONE","surface":"۰۹۱۲۵۷۶۷۰۹۶","language":"fa","field":null},{"start":67,"end":88,"entity_type":"EMAIL","surface":"user.01@aria-lab.test","language":"en","field":null}],"gold_meta":[],"hard_negatives":["INV-1403-0001"],"description":"Persian HR note with ZWNJ, Persian digits, national id, and phone.","description_fa":"نامه منابع انسانی با نیم‌فاصله، رقم فارسی، کد ملی و تلفن."}
12
+ {"doc_id":"DLP-FA-HR-02","template":"fa_hr_letter","family":"fa_hr_letter","split":"development","language":"fa","format":"text","title":"Persian HR letter","body":"خانم لیلا آزمایشی،\nکد‌ملی ۳۱۹۷۲۲۵۷۳۲\nتلفن همراه ۰۹۱۲۳۲۰۳۶۶۹\nایمیل user.02@aria-lab.test\nشماره پیگیری INV-1403-0002 و نسخه ۳٫۲٫۱ فقط ارجاع داخلی‌اند.\n","metadata":{},"gold_spans":[{"start":5,"end":17,"entity_type":"PERSON","surface":"لیلا آزمایشی","language":"fa","field":null},{"start":26,"end":36,"entity_type":"IR_NATIONAL_ID","surface":"۳۱۹۷۲۲۵۷۳۲","language":"fa","field":null},{"start":48,"end":59,"entity_type":"PHONE","surface":"۰۹۱۲۳۲۰۳۶۶۹","language":"fa","field":null},{"start":66,"end":87,"entity_type":"EMAIL","surface":"user.02@aria-lab.test","language":"en","field":null}],"gold_meta":[],"hard_negatives":["INV-1403-0002"],"description":"Persian HR note with ZWNJ, Persian digits, national id, and phone.","description_fa":"نامه منابع انسانی با نیم‌فاصله، رقم فارسی، کد ملی و تلفن."}
13
+ {"doc_id":"DLP-FA-HR-03","template":"fa_hr_letter","family":"fa_hr_letter","split":"development","language":"fa","format":"text","title":"Persian HR letter","body":"خانم امید آزمایشی،\nکد‌ملی ۸۳۸۳۱۲۴۵۸۹\nتلفن همراه ۰۹۱۲۴۸۶۵۴۴۱\nایمیل user.03@aria-lab.test\nشماره پیگیری INV-1403-0003 و نسخه ۳٫۲٫۱ فقط ارجاع داخلی‌اند.\n","metadata":{},"gold_spans":[{"start":5,"end":17,"entity_type":"PERSON","surface":"امید آزمایشی","language":"fa","field":null},{"start":26,"end":36,"entity_type":"IR_NATIONAL_ID","surface":"۸۳۸۳۱۲۴۵۸۹","language":"fa","field":null},{"start":48,"end":59,"entity_type":"PHONE","surface":"۰۹۱۲۴۸۶۵۴۴۱","language":"fa","field":null},{"start":66,"end":87,"entity_type":"EMAIL","surface":"user.03@aria-lab.test","language":"en","field":null}],"gold_meta":[],"hard_negatives":["INV-1403-0003"],"description":"Persian HR note with ZWNJ, Persian digits, national id, and phone.","description_fa":"نامه منابع انسانی با نیم‌فاصله، رقم فارسی، کد ملی و تلفن."}
14
+ {"doc_id":"DLP-FA-HR-04","template":"fa_hr_letter","family":"fa_hr_letter","split":"development","language":"fa","format":"text","title":"Persian HR letter","body":"خانم امید نمونه‌ای،\nکد‌ملی ۴۶۱۷۵۵۴۲۹۷\nتلفن همراه ۰۹۱۲۸۲۹۱۰۰۲\nایمیل user.04@aria-lab.test\nشماره پیگیری INV-1403-0004 و نسخه ۳٫۲٫۱ فقط ارجاع داخلی‌اند.\n","metadata":{},"gold_spans":[{"start":5,"end":18,"entity_type":"PERSON","surface":"امید نمونه‌ای","language":"fa","field":null},{"start":27,"end":37,"entity_type":"IR_NATIONAL_ID","surface":"۴۶۱۷۵۵۴۲۹۷","language":"fa","field":null},{"start":49,"end":60,"entity_type":"PHONE","surface":"۰۹۱۲۸۲۹۱۰۰۲","language":"fa","field":null},{"start":67,"end":88,"entity_type":"EMAIL","surface":"user.04@aria-lab.test","language":"en","field":null}],"gold_meta":[],"hard_negatives":["INV-1403-0004"],"description":"Persian HR note with ZWNJ, Persian digits, national id, and phone.","description_fa":"نامه منابع انسانی با نیم‌فاصله، رقم فارسی، کد ملی و تلفن."}
15
+ {"doc_id":"DLP-FA-HR-05","template":"fa_hr_letter","family":"fa_hr_letter","split":"development","language":"fa","format":"text","title":"Persian HR letter","body":"خانم لیلا نمونه‌ای،\nکد‌ملی ۵۰۰۷۵۷۷۱۰۳\nتلفن همراه ۰۹۱۲۲۳۶۷۲۱۴\nایمیل user.05@aria-lab.test\nشماره پیگیری INV-1403-0005 و نسخه ۳٫۲٫۱ فقط ارجاع داخلی‌اند.\n","metadata":{},"gold_spans":[{"start":5,"end":18,"entity_type":"PERSON","surface":"لیلا نمونه‌ای","language":"fa","field":null},{"start":27,"end":37,"entity_type":"IR_NATIONAL_ID","surface":"۵۰۰۷۵۷۷۱۰۳","language":"fa","field":null},{"start":49,"end":60,"entity_type":"PHONE","surface":"۰۹۱۲۲۳۶۷۲۱۴","language":"fa","field":null},{"start":67,"end":88,"entity_type":"EMAIL","surface":"user.05@aria-lab.test","language":"en","field":null}],"gold_meta":[],"hard_negatives":["INV-1403-0005"],"description":"Persian HR note with ZWNJ, Persian digits, national id, and phone.","description_fa":"نامه منابع انسانی با نیم‌فاصله، رقم فارسی، کد ملی و تلفن."}
16
+ {"doc_id":"DLP-FA-HR-06","template":"fa_hr_letter","family":"fa_hr_letter","split":"development","language":"fa","format":"text","title":"Persian HR letter","body":"خانم لیلا آزمایشی،\nکد‌ملی ۶۸۷۹۸۶۶۳۴۴\nتلفن همراه ۰۹۱۲۶۰۸۳۲۸۳\nایمیل user.06@aria-lab.test\nشماره پیگیری INV-1403-0006 و نسخه ۳٫۲٫۱ فقط ارجاع داخلی‌اند.\n","metadata":{},"gold_spans":[{"start":5,"end":17,"entity_type":"PERSON","surface":"لیلا آزمایشی","language":"fa","field":null},{"start":26,"end":36,"entity_type":"IR_NATIONAL_ID","surface":"۶۸۷۹۸۶۶۳۴۴","language":"fa","field":null},{"start":48,"end":59,"entity_type":"PHONE","surface":"۰۹۱۲۶۰۸۳۲۸۳","language":"fa","field":null},{"start":66,"end":87,"entity_type":"EMAIL","surface":"user.06@aria-lab.test","language":"en","field":null}],"gold_meta":[],"hard_negatives":["INV-1403-0006"],"description":"Persian HR note with ZWNJ, Persian digits, national id, and phone.","description_fa":"نامه منابع انسانی با نیم‌فاصله، رقم فارسی، کد ملی و تلفن."}
17
+ {"doc_id":"DLP-FA-HR-07","template":"fa_hr_letter","family":"fa_hr_letter","split":"development","language":"fa","format":"text","title":"Persian HR letter","body":"خانم لیلا آزمایشی،\nکد‌ملی ۲۵۱۹۰۴۸۰۶۹\nتلفن همراه ۰۹۱۲۷۷۱۴۶۰۱\nایمیل user.07@aria-lab.test\nشماره پیگیری INV-1403-0007 و نسخه ۳٫۲٫۱ فقط ارجاع داخلی‌اند.\n","metadata":{},"gold_spans":[{"start":5,"end":17,"entity_type":"PERSON","surface":"لیلا آزمایشی","language":"fa","field":null},{"start":26,"end":36,"entity_type":"IR_NATIONAL_ID","surface":"۲۵۱۹۰۴۸۰۶۹","language":"fa","field":null},{"start":48,"end":59,"entity_type":"PHONE","surface":"۰۹۱۲۷۷۱۴۶۰۱","language":"fa","field":null},{"start":66,"end":87,"entity_type":"EMAIL","surface":"user.07@aria-lab.test","language":"en","field":null}],"gold_meta":[],"hard_negatives":["INV-1403-0007"],"description":"Persian HR note with ZWNJ, Persian digits, national id, and phone.","description_fa":"نامه منابع انسانی با نیم‌فاصله، رقم فارسی، کد ملی و تلفن."}
18
+ {"doc_id":"DLP-FA-HR-08","template":"fa_hr_letter","family":"fa_hr_letter","split":"development","language":"fa","format":"text","title":"Persian HR letter","body":"خانم امید آزمایشی،\nکد‌ملی ۴۷۶۵۵۰۶۰۷۱\nتلفن همراه ۰۹۱۲۸۷۶۰۹۰۱\nایمیل user.08@aria-lab.test\nشماره پیگیری INV-1403-0008 و نسخه ۳٫۲٫۱ فقط ارجاع داخلی‌اند.\n","metadata":{},"gold_spans":[{"start":5,"end":17,"entity_type":"PERSON","surface":"امید آزمایشی","language":"fa","field":null},{"start":26,"end":36,"entity_type":"IR_NATIONAL_ID","surface":"۴۷۶۵۵۰۶۰۷۱","language":"fa","field":null},{"start":48,"end":59,"entity_type":"PHONE","surface":"۰۹۱۲۸۷۶۰۹۰۱","language":"fa","field":null},{"start":66,"end":87,"entity_type":"EMAIL","surface":"user.08@aria-lab.test","language":"en","field":null}],"gold_meta":[],"hard_negatives":["INV-1403-0008"],"description":"Persian HR note with ZWNJ, Persian digits, national id, and phone.","description_fa":"نامه منابع انسانی با نیم‌فاصله، رقم فارسی، کد ملی و تلفن."}
19
+ {"doc_id":"DLP-FA-HR-09","template":"fa_hr_letter","family":"fa_hr_letter","split":"development","language":"fa","format":"text","title":"Persian HR letter","body":"خانم سارا نمونه‌ای،\nکد‌ملی ۳۲۸۹۳۷۲۲۰۰\nتلفن همراه ۰۹۱۲۲۸۹۳۲۸۹\nایمیل user.09@aria-lab.test\nشماره پیگیری INV-1403-0009 و نسخه ۳٫۲٫۱ فقط ارجاع داخلی‌اند.\n","metadata":{},"gold_spans":[{"start":5,"end":18,"entity_type":"PERSON","surface":"سارا نمونه‌ای","language":"fa","field":null},{"start":27,"end":37,"entity_type":"IR_NATIONAL_ID","surface":"۳۲۸۹۳۷۲۲۰۰","language":"fa","field":null},{"start":49,"end":60,"entity_type":"PHONE","surface":"۰۹۱۲۲۸۹۳۲۸۹","language":"fa","field":null},{"start":67,"end":88,"entity_type":"EMAIL","surface":"user.09@aria-lab.test","language":"en","field":null}],"gold_meta":[],"hard_negatives":["INV-1403-0009"],"description":"Persian HR note with ZWNJ, Persian digits, national id, and phone.","description_fa":"نامه منابع انسانی با نیم‌فاصله، رقم فارسی، کد ملی و تلفن."}
20
+ {"doc_id":"DLP-FA-HR-10","template":"fa_hr_letter","family":"fa_hr_letter","split":"development","language":"fa","format":"text","title":"Persian HR letter","body":"خانم سارا نمونه‌ای،\nکد‌ملی ۳۰۸۷۰۲۵۴۱۰\nتلفن همراه ۰۹۱۲۴۰۶۷۵۴۸\nایمیل user.10@aria-lab.test\nشماره پیگیری INV-1403-0010 و نسخه ۳٫۲٫۱ فقط ارجاع داخلی‌اند.\n","metadata":{},"gold_spans":[{"start":5,"end":18,"entity_type":"PERSON","surface":"سارا نمونه‌ای","language":"fa","field":null},{"start":27,"end":37,"entity_type":"IR_NATIONAL_ID","surface":"۳۰۸۷۰۲۵۴۱۰","language":"fa","field":null},{"start":49,"end":60,"entity_type":"PHONE","surface":"۰۹۱۲۴۰۶۷۵۴۸","language":"fa","field":null},{"start":67,"end":88,"entity_type":"EMAIL","surface":"user.10@aria-lab.test","language":"en","field":null}],"gold_meta":[],"hard_negatives":["INV-1403-0010"],"description":"Persian HR note with ZWNJ, Persian digits, national id, and phone.","description_fa":"نامه منابع انسانی با نیم‌فاصله، رقم فارسی، کد ملی و تلفن."}
21
+ {"doc_id":"DLP-CSV-INV-01","template":"mixed_invoice_csv","family":"mixed_invoice_csv","split":"development","language":"mixed","format":"csv","title":"Invoice CSV","body":"invoice_id,customer,email,phone,lookup\nINV-2024-1001,Ava Labian,Ava.Labian01@aria-lab.test,09129456762,1987844778\n","metadata":{},"gold_spans":[{"start":53,"end":63,"entity_type":"PERSON","surface":"Ava Labian","language":"en","field":null},{"start":64,"end":90,"entity_type":"EMAIL","surface":"Ava.Labian01@aria-lab.test","language":"en","field":null},{"start":91,"end":102,"entity_type":"PHONE","surface":"09129456762","language":"en","field":null}],"gold_meta":[],"hard_negatives":["INV-2024-1001","1987844778"],"description":"CSV with a customer row and a checksum-failing lookup number.","description_fa":"سی‌اس‌وی مشتری با شماره جست‌وجوی بدون checksum معتبر."}
22
+ {"doc_id":"DLP-CSV-INV-02","template":"mixed_invoice_csv","family":"mixed_invoice_csv","split":"development","language":"mixed","format":"csv","title":"Invoice CSV","body":"invoice_id,customer,email,phone,lookup\nINV-2024-1002,Nima Fixture,Nima.Fixture02@aria-lab.test,09127891210,2513822278\n","metadata":{},"gold_spans":[{"start":53,"end":65,"entity_type":"PERSON","surface":"Nima Fixture","language":"en","field":null},{"start":66,"end":94,"entity_type":"EMAIL","surface":"Nima.Fixture02@aria-lab.test","language":"en","field":null},{"start":95,"end":106,"entity_type":"PHONE","surface":"09127891210","language":"en","field":null}],"gold_meta":[],"hard_negatives":["INV-2024-1002","2513822278"],"description":"CSV with a customer row and a checksum-failing lookup number.","description_fa":"سی‌اس‌وی مشتری با شماره جست‌وجوی بدون checksum معتبر."}
23
+ {"doc_id":"DLP-CSV-INV-03","template":"mixed_invoice_csv","family":"mixed_invoice_csv","split":"development","language":"mixed","format":"csv","title":"Invoice CSV","body":"invoice_id,customer,email,phone,lookup\nINV-2024-1003,Leila Fixture,Leila.Fixture03@aria-lab.test,09124895575,9231499344\n","metadata":{},"gold_spans":[{"start":53,"end":66,"entity_type":"PERSON","surface":"Leila Fixture","language":"en","field":null},{"start":67,"end":96,"entity_type":"EMAIL","surface":"Leila.Fixture03@aria-lab.test","language":"en","field":null},{"start":97,"end":108,"entity_type":"PHONE","surface":"09124895575","language":"en","field":null}],"gold_meta":[],"hard_negatives":["INV-2024-1003","9231499344"],"description":"CSV with a customer row and a checksum-failing lookup number.","description_fa":"سی‌اس‌وی مشتری با شماره جست‌وجوی بدون checksum معتبر."}
24
+ {"doc_id":"DLP-CSV-INV-04","template":"mixed_invoice_csv","family":"mixed_invoice_csv","split":"development","language":"mixed","format":"csv","title":"Invoice CSV","body":"invoice_id,customer,email,phone,lookup\nINV-2024-1004,Omar Labian,Omar.Labian04@aria-lab.test,09124151655,3699795722\n","metadata":{},"gold_spans":[{"start":53,"end":64,"entity_type":"PERSON","surface":"Omar Labian","language":"en","field":null},{"start":65,"end":92,"entity_type":"EMAIL","surface":"Omar.Labian04@aria-lab.test","language":"en","field":null},{"start":93,"end":104,"entity_type":"PHONE","surface":"09124151655","language":"en","field":null}],"gold_meta":[],"hard_negatives":["INV-2024-1004","3699795722"],"description":"CSV with a customer row and a checksum-failing lookup number.","description_fa":"سی‌اس‌وی مشتری با شماره جست‌وجوی بدون checksum معتبر."}
25
+ {"doc_id":"DLP-CSV-INV-05","template":"mixed_invoice_csv","family":"mixed_invoice_csv","split":"development","language":"mixed","format":"csv","title":"Invoice CSV","body":"invoice_id,customer,email,phone,lookup\nINV-2024-1005,Ava Labian,Ava.Labian05@aria-lab.test,09123831675,3303447994\n","metadata":{},"gold_spans":[{"start":53,"end":63,"entity_type":"PERSON","surface":"Ava Labian","language":"en","field":null},{"start":64,"end":90,"entity_type":"EMAIL","surface":"Ava.Labian05@aria-lab.test","language":"en","field":null},{"start":91,"end":102,"entity_type":"PHONE","surface":"09123831675","language":"en","field":null}],"gold_meta":[],"hard_negatives":["INV-2024-1005","3303447994"],"description":"CSV with a customer row and a checksum-failing lookup number.","description_fa":"سی‌اس‌وی مشتری با شماره جست‌وجوی بدون checksum معتبر."}
26
+ {"doc_id":"DLP-CSV-INV-06","template":"mixed_invoice_csv","family":"mixed_invoice_csv","split":"development","language":"mixed","format":"csv","title":"Invoice CSV","body":"invoice_id,customer,email,phone,lookup\nINV-2024-1006,Noah Fixture,Noah.Fixture06@aria-lab.test,09125428133,3478403597\n","metadata":{},"gold_spans":[{"start":53,"end":65,"entity_type":"PERSON","surface":"Noah Fixture","language":"en","field":null},{"start":66,"end":94,"entity_type":"EMAIL","surface":"Noah.Fixture06@aria-lab.test","language":"en","field":null},{"start":95,"end":106,"entity_type":"PHONE","surface":"09125428133","language":"en","field":null}],"gold_meta":[],"hard_negatives":["INV-2024-1006","3478403597"],"description":"CSV with a customer row and a checksum-failing lookup number.","description_fa":"سی‌اس‌وی مشتری با شماره جست‌وجوی بدون checksum معتبر."}
27
+ {"doc_id":"DLP-CSV-INV-07","template":"mixed_invoice_csv","family":"mixed_invoice_csv","split":"development","language":"mixed","format":"csv","title":"Invoice CSV","body":"invoice_id,customer,email,phone,lookup\nINV-2024-1007,Helen Fixture,Helen.Fixture07@aria-lab.test,09124672166,5386511573\n","metadata":{},"gold_spans":[{"start":53,"end":66,"entity_type":"PERSON","surface":"Helen Fixture","language":"en","field":null},{"start":67,"end":96,"entity_type":"EMAIL","surface":"Helen.Fixture07@aria-lab.test","language":"en","field":null},{"start":97,"end":108,"entity_type":"PHONE","surface":"09124672166","language":"en","field":null}],"gold_meta":[],"hard_negatives":["INV-2024-1007","5386511573"],"description":"CSV with a customer row and a checksum-failing lookup number.","description_fa":"سی‌اس‌وی مشتری با شماره جست‌وجوی بدون checksum معتبر."}
28
+ {"doc_id":"DLP-CSV-INV-08","template":"mixed_invoice_csv","family":"mixed_invoice_csv","split":"development","language":"mixed","format":"csv","title":"Invoice CSV","body":"invoice_id,customer,email,phone,lookup\nINV-2024-1008,Omar Fixture,Omar.Fixture08@aria-lab.test,09128527564,7963012808\n","metadata":{},"gold_spans":[{"start":53,"end":65,"entity_type":"PERSON","surface":"Omar Fixture","language":"en","field":null},{"start":66,"end":94,"entity_type":"EMAIL","surface":"Omar.Fixture08@aria-lab.test","language":"en","field":null},{"start":95,"end":106,"entity_type":"PHONE","surface":"09128527564","language":"en","field":null}],"gold_meta":[],"hard_negatives":["INV-2024-1008","7963012808"],"description":"CSV with a customer row and a checksum-failing lookup number.","description_fa":"سی‌اس‌وی مشتری با شماره جست‌وجوی بدون checksum معتبر."}
29
+ {"doc_id":"DLP-CSV-INV-09","template":"mixed_invoice_csv","family":"mixed_invoice_csv","split":"development","language":"mixed","format":"csv","title":"Invoice CSV","body":"invoice_id,customer,email,phone,lookup\nINV-2024-1009,Leila Labian,Leila.Labian09@aria-lab.test,09127031590,3935218380\n","metadata":{},"gold_spans":[{"start":53,"end":65,"entity_type":"PERSON","surface":"Leila Labian","language":"en","field":null},{"start":66,"end":94,"entity_type":"EMAIL","surface":"Leila.Labian09@aria-lab.test","language":"en","field":null},{"start":95,"end":106,"entity_type":"PHONE","surface":"09127031590","language":"en","field":null}],"gold_meta":[],"hard_negatives":["INV-2024-1009","3935218380"],"description":"CSV with a customer row and a checksum-failing lookup number.","description_fa":"سی‌اس‌وی مشتری با شماره جست‌وجوی بدون checksum معتبر."}
30
+ {"doc_id":"DLP-CSV-INV-10","template":"mixed_invoice_csv","family":"mixed_invoice_csv","split":"development","language":"mixed","format":"csv","title":"Invoice CSV","body":"invoice_id,customer,email,phone,lookup\nINV-2024-1010,Leila Fixture,Leila.Fixture10@aria-lab.test,09123631859,4874249265\n","metadata":{},"gold_spans":[{"start":53,"end":66,"entity_type":"PERSON","surface":"Leila Fixture","language":"en","field":null},{"start":67,"end":96,"entity_type":"EMAIL","surface":"Leila.Fixture10@aria-lab.test","language":"en","field":null},{"start":97,"end":108,"entity_type":"PHONE","surface":"09123631859","language":"en","field":null}],"gold_meta":[],"hard_negatives":["INV-2024-1010","4874249265"],"description":"CSV with a customer row and a checksum-failing lookup number.","description_fa":"سی‌اس‌وی مشتری با شماره جست‌وجوی بدون checksum معتبر."}
31
+ {"doc_id":"DLP-EN-ONB-01","template":"en_onboarding","family":"en_onboarding","split":"development","language":"en","format":"text","title":"Onboarding worksheet","body":"Name: Helen Labian\nWork email: Helen.Labian01@aria-lab.test\nTravel advance test PAN: 4111137038772834\nPayroll IBAN: IR308660110372211081968590\nISBN 978-3-16-148410-0 is a catalog example, not an account.\n","metadata":{},"gold_spans":[{"start":6,"end":18,"entity_type":"PERSON","surface":"Helen Labian","language":"en","field":null},{"start":31,"end":59,"entity_type":"EMAIL","surface":"Helen.Labian01@aria-lab.test","language":"en","field":null},{"start":85,"end":101,"entity_type":"CREDIT_CARD","surface":"4111137038772834","language":"en","field":null},{"start":116,"end":142,"entity_type":"IBAN","surface":"IR308660110372211081968590","language":"en","field":null}],"gold_meta":[],"hard_negatives":["978-3-16-148410-0","9783161484100"],"description":"Onboarding sheet with test PAN and IBAN plus an ISBN hard negative.","description_fa":"برگه ورود با کارت و شبا آزمایشی و ISBN به‌عنوان منفی سخت."}
32
+ {"doc_id":"DLP-EN-ONB-02","template":"en_onboarding","family":"en_onboarding","split":"development","language":"en","format":"text","title":"Onboarding worksheet","body":"Name: Helen Labian\nWork email: Helen.Labian02@aria-lab.test\nTravel advance test PAN: 4111529393529637\nPayroll IBAN: IR044280439609406102122129\nISBN 978-3-16-148410-0 is a catalog example, not an account.\n","metadata":{},"gold_spans":[{"start":6,"end":18,"entity_type":"PERSON","surface":"Helen Labian","language":"en","field":null},{"start":31,"end":59,"entity_type":"EMAIL","surface":"Helen.Labian02@aria-lab.test","language":"en","field":null},{"start":85,"end":101,"entity_type":"CREDIT_CARD","surface":"4111529393529637","language":"en","field":null},{"start":116,"end":142,"entity_type":"IBAN","surface":"IR044280439609406102122129","language":"en","field":null}],"gold_meta":[],"hard_negatives":["978-3-16-148410-0","9783161484100"],"description":"Onboarding sheet with test PAN and IBAN plus an ISBN hard negative.","description_fa":"برگه ورود با کارت و شبا آزمایشی و ISBN به‌عنوان منفی سخت."}
33
+ {"doc_id":"DLP-EN-ONB-03","template":"en_onboarding","family":"en_onboarding","split":"development","language":"en","format":"text","title":"Onboarding worksheet","body":"Name: Helen Labian\nWork email: Helen.Labian03@aria-lab.test\nTravel advance test PAN: 4111785437328044\nPayroll IBAN: IR019630083854359765553642\nISBN 978-3-16-148410-0 is a catalog example, not an account.\n","metadata":{},"gold_spans":[{"start":6,"end":18,"entity_type":"PERSON","surface":"Helen Labian","language":"en","field":null},{"start":31,"end":59,"entity_type":"EMAIL","surface":"Helen.Labian03@aria-lab.test","language":"en","field":null},{"start":85,"end":101,"entity_type":"CREDIT_CARD","surface":"4111785437328044","language":"en","field":null},{"start":116,"end":142,"entity_type":"IBAN","surface":"IR019630083854359765553642","language":"en","field":null}],"gold_meta":[],"hard_negatives":["978-3-16-148410-0","9783161484100"],"description":"Onboarding sheet with test PAN and IBAN plus an ISBN hard negative.","description_fa":"برگه ورود با کارت و شبا آزمایشی و ISBN به‌عنوان منفی سخت."}
34
+ {"doc_id":"DLP-EN-ONB-04","template":"en_onboarding","family":"en_onboarding","split":"development","language":"en","format":"text","title":"Onboarding worksheet","body":"Name: Helen Labian\nWork email: Helen.Labian04@aria-lab.test\nTravel advance test PAN: 4111759320597908\nPayroll IBAN: IR472630756281276293569333\nISBN 978-3-16-148410-0 is a catalog example, not an account.\n","metadata":{},"gold_spans":[{"start":6,"end":18,"entity_type":"PERSON","surface":"Helen Labian","language":"en","field":null},{"start":31,"end":59,"entity_type":"EMAIL","surface":"Helen.Labian04@aria-lab.test","language":"en","field":null},{"start":85,"end":101,"entity_type":"CREDIT_CARD","surface":"4111759320597908","language":"en","field":null},{"start":116,"end":142,"entity_type":"IBAN","surface":"IR472630756281276293569333","language":"en","field":null}],"gold_meta":[],"hard_negatives":["978-3-16-148410-0","9783161484100"],"description":"Onboarding sheet with test PAN and IBAN plus an ISBN hard negative.","description_fa":"برگه ورود با کارت و شبا آزمایشی و ISBN به‌عنوان منفی سخت."}
35
+ {"doc_id":"DLP-EN-ONB-05","template":"en_onboarding","family":"en_onboarding","split":"development","language":"en","format":"text","title":"Onboarding worksheet","body":"Name: Leila Fixture\nWork email: Leila.Fixture05@aria-lab.test\nTravel advance test PAN: 4111313694373346\nPayroll IBAN: IR466530650673099236853917\nISBN 978-3-16-148410-0 is a catalog example, not an account.\n","metadata":{},"gold_spans":[{"start":6,"end":19,"entity_type":"PERSON","surface":"Leila Fixture","language":"en","field":null},{"start":32,"end":61,"entity_type":"EMAIL","surface":"Leila.Fixture05@aria-lab.test","language":"en","field":null},{"start":87,"end":103,"entity_type":"CREDIT_CARD","surface":"4111313694373346","language":"en","field":null},{"start":118,"end":144,"entity_type":"IBAN","surface":"IR466530650673099236853917","language":"en","field":null}],"gold_meta":[],"hard_negatives":["978-3-16-148410-0","9783161484100"],"description":"Onboarding sheet with test PAN and IBAN plus an ISBN hard negative.","description_fa":"برگه ورود با کارت و شبا آزمایشی و ISBN به‌عنوان منفی سخت."}
36
+ {"doc_id":"DLP-EN-ONB-06","template":"en_onboarding","family":"en_onboarding","split":"development","language":"en","format":"text","title":"Onboarding worksheet","body":"Name: Nima Labian\nWork email: Nima.Labian06@aria-lab.test\nTravel advance test PAN: 4111814102596287\nPayroll IBAN: IR581060722681444648227055\nISBN 978-3-16-148410-0 is a catalog example, not an account.\n","metadata":{},"gold_spans":[{"start":6,"end":17,"entity_type":"PERSON","surface":"Nima Labian","language":"en","field":null},{"start":30,"end":57,"entity_type":"EMAIL","surface":"Nima.Labian06@aria-lab.test","language":"en","field":null},{"start":83,"end":99,"entity_type":"CREDIT_CARD","surface":"4111814102596287","language":"en","field":null},{"start":114,"end":140,"entity_type":"IBAN","surface":"IR581060722681444648227055","language":"en","field":null}],"gold_meta":[],"hard_negatives":["978-3-16-148410-0","9783161484100"],"description":"Onboarding sheet with test PAN and IBAN plus an ISBN hard negative.","description_fa":"برگه ورود با کارت و شبا آزمایشی و ISBN به‌عنوان منفی سخت."}
37
+ {"doc_id":"DLP-EN-ONB-07","template":"en_onboarding","family":"en_onboarding","split":"development","language":"en","format":"text","title":"Onboarding worksheet","body":"Name: Noah Fixture\nWork email: Noah.Fixture07@aria-lab.test\nTravel advance test PAN: 4111842413164297\nPayroll IBAN: IR895120978818064078339048\nISBN 978-3-16-148410-0 is a catalog example, not an account.\n","metadata":{},"gold_spans":[{"start":6,"end":18,"entity_type":"PERSON","surface":"Noah Fixture","language":"en","field":null},{"start":31,"end":59,"entity_type":"EMAIL","surface":"Noah.Fixture07@aria-lab.test","language":"en","field":null},{"start":85,"end":101,"entity_type":"CREDIT_CARD","surface":"4111842413164297","language":"en","field":null},{"start":116,"end":142,"entity_type":"IBAN","surface":"IR895120978818064078339048","language":"en","field":null}],"gold_meta":[],"hard_negatives":["978-3-16-148410-0","9783161484100"],"description":"Onboarding sheet with test PAN and IBAN plus an ISBN hard negative.","description_fa":"برگه ورود با کارت و شبا آزمایشی و ISBN به‌عنوان منفی سخت."}
38
+ {"doc_id":"DLP-EN-ONB-08","template":"en_onboarding","family":"en_onboarding","split":"development","language":"en","format":"text","title":"Onboarding worksheet","body":"Name: Noah Fixture\nWork email: Noah.Fixture08@aria-lab.test\nTravel advance test PAN: 4111329710194664\nPayroll IBAN: IR237680114663394226281259\nISBN 978-3-16-148410-0 is a catalog example, not an account.\n","metadata":{},"gold_spans":[{"start":6,"end":18,"entity_type":"PERSON","surface":"Noah Fixture","language":"en","field":null},{"start":31,"end":59,"entity_type":"EMAIL","surface":"Noah.Fixture08@aria-lab.test","language":"en","field":null},{"start":85,"end":101,"entity_type":"CREDIT_CARD","surface":"4111329710194664","language":"en","field":null},{"start":116,"end":142,"entity_type":"IBAN","surface":"IR237680114663394226281259","language":"en","field":null}],"gold_meta":[],"hard_negatives":["978-3-16-148410-0","9783161484100"],"description":"Onboarding sheet with test PAN and IBAN plus an ISBN hard negative.","description_fa":"برگه ورود با کارت و شبا آزمایشی و ISBN به‌عنوان منفی سخت."}
39
+ {"doc_id":"DLP-EN-ONB-09","template":"en_onboarding","family":"en_onboarding","split":"development","language":"en","format":"text","title":"Onboarding worksheet","body":"Name: Leila Fixture\nWork email: Leila.Fixture09@aria-lab.test\nTravel advance test PAN: 4111996662241709\nPayroll IBAN: IR772810906854799104860878\nISBN 978-3-16-148410-0 is a catalog example, not an account.\n","metadata":{},"gold_spans":[{"start":6,"end":19,"entity_type":"PERSON","surface":"Leila Fixture","language":"en","field":null},{"start":32,"end":61,"entity_type":"EMAIL","surface":"Leila.Fixture09@aria-lab.test","language":"en","field":null},{"start":87,"end":103,"entity_type":"CREDIT_CARD","surface":"4111996662241709","language":"en","field":null},{"start":118,"end":144,"entity_type":"IBAN","surface":"IR772810906854799104860878","language":"en","field":null}],"gold_meta":[],"hard_negatives":["978-3-16-148410-0","9783161484100"],"description":"Onboarding sheet with test PAN and IBAN plus an ISBN hard negative.","description_fa":"برگه ورود با کارت و شبا آزمایشی و ISBN به‌عنوان منفی سخت."}
40
+ {"doc_id":"DLP-EN-ONB-10","template":"en_onboarding","family":"en_onboarding","split":"development","language":"en","format":"text","title":"Onboarding worksheet","body":"Name: Nima Fixture\nWork email: Nima.Fixture10@aria-lab.test\nTravel advance test PAN: 4111041294625191\nPayroll IBAN: IR338550745309562340114615\nISBN 978-3-16-148410-0 is a catalog example, not an account.\n","metadata":{},"gold_spans":[{"start":6,"end":18,"entity_type":"PERSON","surface":"Nima Fixture","language":"en","field":null},{"start":31,"end":59,"entity_type":"EMAIL","surface":"Nima.Fixture10@aria-lab.test","language":"en","field":null},{"start":85,"end":101,"entity_type":"CREDIT_CARD","surface":"4111041294625191","language":"en","field":null},{"start":116,"end":142,"entity_type":"IBAN","surface":"IR338550745309562340114615","language":"en","field":null}],"gold_meta":[],"hard_negatives":["978-3-16-148410-0","9783161484100"],"description":"Onboarding sheet with test PAN and IBAN plus an ISBN hard negative.","description_fa":"برگه ورود با کارت و شبا آزمایشی و ISBN به‌عنوان منفی سخت."}
41
+ {"doc_id":"DLP-FA-CTR-01","template":"fa_contract_notice","family":"fa_contract_notice","split":"test","language":"fa","format":"text","title":"Persian contract notice","body":"موضوع: اخطار قراردادی\nنام: لیلا آزمایشی\nکد ملی 4985528323\nتلفن ۰۹۱۲۳۴۲۱۵۱۴\nکد ملی 3666916466 در سامانه پذیرفته نشد و برای بازبینی مانده است.\nتاریخ ۱۴۰۳/۰۵/۱۲ و نسخه ۱.۰.۰ ارجاع سند هستند.\n","metadata":{},"gold_spans":[{"start":27,"end":39,"entity_type":"PERSON","surface":"لیلا آزمایشی","language":"fa","field":null},{"start":47,"end":57,"entity_type":"IR_NATIONAL_ID","surface":"4985528323","language":"fa","field":null},{"start":63,"end":74,"entity_type":"PHONE","surface":"۰۹۱۲۳۴۲۱۵۱۴","language":"fa","field":null}],"gold_meta":[],"hard_negatives":["3666916466","1403/05/12","۱۴۰۳/۰۵/۱۲","1.0.0"],"description":"Persian contract with a failed national-id left for review.","description_fa":"اخطار قراردادی با کد ملی نامعتبر برای بازبینی."}
42
+ {"doc_id":"DLP-FA-CTR-02","template":"fa_contract_notice","family":"fa_contract_notice","split":"test","language":"fa","format":"text","title":"Persian contract notice","body":"موضوع: اخطار قراردادی\nنام: سارا آزمایشی\nکد ملی 6361423751\nتلفن ۰۹۱۲۳۶۳۰۹۶۹\nکد ملی 9032951212 در سامانه پذیرفته نشد و برای بازبینی مانده است.\nتاریخ ۱۴۰۳/۰۵/۱۲ و نسخه ۱.۰.۰ ارجاع سند هستند.\n","metadata":{},"gold_spans":[{"start":27,"end":39,"entity_type":"PERSON","surface":"سارا آزمایشی","language":"fa","field":null},{"start":47,"end":57,"entity_type":"IR_NATIONAL_ID","surface":"6361423751","language":"fa","field":null},{"start":63,"end":74,"entity_type":"PHONE","surface":"۰۹۱۲۳۶۳۰۹۶۹","language":"fa","field":null}],"gold_meta":[],"hard_negatives":["9032951212","1403/05/12","۱۴۰۳/۰۵/۱۲","1.0.0"],"description":"Persian contract with a failed national-id left for review.","description_fa":"اخطار قراردادی با کد ملی نامعتبر برای بازبینی."}
43
+ {"doc_id":"DLP-FA-CTR-03","template":"fa_contract_notice","family":"fa_contract_notice","split":"test","language":"fa","format":"text","title":"Persian contract notice","body":"موضوع: اخطار قراردادی\nنام: نیما نمونه‌ای\nکد ملی 3589737379\nتلفن ۰۹۱۲۵۹۲۸۹۵۰\nکد ملی 9386005221 در سامانه پذیرفته نشد و برای بازبینی مانده است.\nتاریخ ۱۴۰۳/۰۵/۱۲ و نسخه ۱.۰.۰ ارجاع سند هستند.\n","metadata":{},"gold_spans":[{"start":27,"end":40,"entity_type":"PERSON","surface":"نیما نمونه‌ای","language":"fa","field":null},{"start":48,"end":58,"entity_type":"IR_NATIONAL_ID","surface":"3589737379","language":"fa","field":null},{"start":64,"end":75,"entity_type":"PHONE","surface":"۰۹۱۲۵۹۲۸۹۵۰","language":"fa","field":null}],"gold_meta":[],"hard_negatives":["9386005221","1403/05/12","۱۴۰۳/۰۵/۱۲","1.0.0"],"description":"Persian contract with a failed national-id left for review.","description_fa":"اخطار قراردادی با کد ملی نامعتبر برای بازبینی."}
44
+ {"doc_id":"DLP-FA-CTR-04","template":"fa_contract_notice","family":"fa_contract_notice","split":"test","language":"fa","format":"text","title":"Persian contract notice","body":"موضوع: اخطار قراردادی\nنام: نیما آزمایشی\nکد ملی 2341457452\nتلفن ۰۹۱۲۲۳۶۵۹۰۷\nکد ملی 2506786750 در سامانه پذیرفته نشد و برای بازبینی مانده است.\nتار��خ ۱۴۰۳/۰۵/۱۲ و نسخه ۱.۰.۰ ارجاع سند هستند.\n","metadata":{},"gold_spans":[{"start":27,"end":39,"entity_type":"PERSON","surface":"نیما آزمایشی","language":"fa","field":null},{"start":47,"end":57,"entity_type":"IR_NATIONAL_ID","surface":"2341457452","language":"fa","field":null},{"start":63,"end":74,"entity_type":"PHONE","surface":"۰۹۱۲۲۳۶۵۹۰۷","language":"fa","field":null}],"gold_meta":[],"hard_negatives":["2506786750","1403/05/12","۱۴۰۳/۰۵/۱۲","1.0.0"],"description":"Persian contract with a failed national-id left for review.","description_fa":"اخطار قراردادی با کد ملی نامعتبر برای بازبینی."}
45
+ {"doc_id":"DLP-FA-CTR-05","template":"fa_contract_notice","family":"fa_contract_notice","split":"test","language":"fa","format":"text","title":"Persian contract notice","body":"موضوع: اخطار قراردادی\nنام: نیما نمونه‌ای\nکد ملی 4061666126\nتلفن ۰۹۱۲۴۹۶۶۴۱۱\nکد ملی 9044212126 در سامانه پذیرفته نشد و برای بازبینی مانده است.\nتاریخ ۱۴۰۳/۰۵/۱۲ و نسخه ۱.۰.۰ ارجاع سند هستند.\n","metadata":{},"gold_spans":[{"start":27,"end":40,"entity_type":"PERSON","surface":"نیما نمونه‌ای","language":"fa","field":null},{"start":48,"end":58,"entity_type":"IR_NATIONAL_ID","surface":"4061666126","language":"fa","field":null},{"start":64,"end":75,"entity_type":"PHONE","surface":"۰۹۱۲۴۹۶۶۴۱۱","language":"fa","field":null}],"gold_meta":[],"hard_negatives":["9044212126","1403/05/12","۱۴۰۳/۰۵/۱۲","1.0.0"],"description":"Persian contract with a failed national-id left for review.","description_fa":"اخطار قراردادی با کد ملی نامعتبر برای بازبینی."}
46
+ {"doc_id":"DLP-FA-CTR-06","template":"fa_contract_notice","family":"fa_contract_notice","split":"test","language":"fa","format":"text","title":"Persian contract notice","body":"موضوع: اخطار قراردادی\nنام: لیلا نمونه‌ای\nکد ملی 3179136243\nتلفن ۰۹۱۲۸۱۳۵۴۲۶\nکد ملی 4511025429 در سامانه پذیرفته نشد و برای بازبینی مانده است.\nتاریخ ۱۴۰۳/۰۵/۱۲ و نسخه ۱.۰.۰ ارجاع سند هستند.\n","metadata":{},"gold_spans":[{"start":27,"end":40,"entity_type":"PERSON","surface":"لیلا نمونه‌ای","language":"fa","field":null},{"start":48,"end":58,"entity_type":"IR_NATIONAL_ID","surface":"3179136243","language":"fa","field":null},{"start":64,"end":75,"entity_type":"PHONE","surface":"۰۹۱۲۸۱۳۵۴۲۶","language":"fa","field":null}],"gold_meta":[],"hard_negatives":["4511025429","1403/05/12","۱۴۰۳/۰۵/۱۲","1.0.0"],"description":"Persian contract with a failed national-id left for review.","description_fa":"اخطار قراردادی با کد ملی نامعتبر برای بازبینی."}
47
+ {"doc_id":"DLP-FA-CTR-07","template":"fa_contract_notice","family":"fa_contract_notice","split":"test","language":"fa","format":"text","title":"Persian contract notice","body":"موضوع: اخطار قراردادی\nنام: امید آزمایشی\nکد ملی 2154052452\nتلفن ۰۹۱۲۸۸۱۷۱۸۵\nکد ملی 2976140040 در سامانه پذیرفته نشد و برای بازبینی مانده است.\nتاریخ ۱۴۰۳/۰۵/۱۲ و نسخه ۱.۰.۰ ارجاع سند هستند.\n","metadata":{},"gold_spans":[{"start":27,"end":39,"entity_type":"PERSON","surface":"امید آزمایشی","language":"fa","field":null},{"start":47,"end":57,"entity_type":"IR_NATIONAL_ID","surface":"2154052452","language":"fa","field":null},{"start":63,"end":74,"entity_type":"PHONE","surface":"۰۹۱۲۸۸۱۷۱۸۵","language":"fa","field":null}],"gold_meta":[],"hard_negatives":["2976140040","1403/05/12","۱۴۰۳/۰۵/۱۲","1.0.0"],"description":"Persian contract with a failed national-id left for review.","description_fa":"اخطار قراردادی با کد ملی نامعتبر برای بازبینی."}
48
+ {"doc_id":"DLP-FA-CTR-08","template":"fa_contract_notice","family":"fa_contract_notice","split":"test","language":"fa","format":"text","title":"Persian contract notice","body":"موضوع: اخطار قراردادی\nنام: سارا آزمایشی\nکد ملی 2313920291\nتلفن ۰۹۱۲۱۳۴۴۶۷۵\nکد ملی 7199959037 در سامانه پذیرفته نشد و برای بازبینی مانده است.\nتاریخ ۱۴۰۳/۰۵/۱۲ و نسخه ۱.۰.۰ ارجاع سند هستند.\n","metadata":{},"gold_spans":[{"start":27,"end":39,"entity_type":"PERSON","surface":"سارا آزمایشی","language":"fa","field":null},{"start":47,"end":57,"entity_type":"IR_NATIONAL_ID","surface":"2313920291","language":"fa","field":null},{"start":63,"end":74,"entity_type":"PHONE","surface":"۰۹۱۲۱۳۴۴۶۷۵","language":"fa","field":null}],"gold_meta":[],"hard_negatives":["7199959037","1403/05/12","۱۴۰۳/۰۵/۱۲","1.0.0"],"description":"Persian contract with a failed national-id left for review.","description_fa":"اخطار قراردادی با کد ملی نامعتبر برای بازبینی."}
49
+ {"doc_id":"DLP-FA-CTR-09","template":"fa_contract_notice","family":"fa_contract_notice","split":"test","language":"fa","format":"text","title":"Persian contract notice","body":"موضوع: اخطار قراردادی\nنام: نیما نمونه‌ای\nکد ملی 0190065516\nتلفن ۰۹۱۲۷۲۰۹۰۸۸\nکد ملی 2601360047 در سامانه پذیرفته نشد و برای بازبینی مانده است.\nتاریخ ۱۴۰۳/۰۵/۱۲ و نسخه ۱.۰.۰ ارجاع سند هستند.\n","metadata":{},"gold_spans":[{"start":27,"end":40,"entity_type":"PERSON","surface":"نیما نمونه‌ای","language":"fa","field":null},{"start":48,"end":58,"entity_type":"IR_NATIONAL_ID","surface":"0190065516","language":"fa","field":null},{"start":64,"end":75,"entity_type":"PHONE","surface":"۰۹۱۲۷۲۰۹۰۸۸","language":"fa","field":null}],"gold_meta":[],"hard_negatives":["2601360047","1403/05/12","۱۴۰۳/۰۵/۱۲","1.0.0"],"description":"Persian contract with a failed national-id left for review.","description_fa":"اخطار قراردادی با کد ملی نامعتبر برای بازبینی."}
50
+ {"doc_id":"DLP-FA-CTR-10","template":"fa_contract_notice","family":"fa_contract_notice","split":"test","language":"fa","format":"text","title":"Persian contract notice","body":"موضوع: اخطار قراردادی\nنام: لیلا آزمایشی\nکد ملی 8212633349\nتلفن ۰۹۱۲۷۴۸۴۳۶۷\nکد ملی 4063697593 در سامانه پذیرفته نشد و برای بازبینی مانده است.\nتاریخ ۱۴۰۳/۰۵/۱۲ و نسخه ۱.۰.۰ ارجاع سند هستند.\n","metadata":{},"gold_spans":[{"start":27,"end":39,"entity_type":"PERSON","surface":"لیلا آزمایشی","language":"fa","field":null},{"start":47,"end":57,"entity_type":"IR_NATIONAL_ID","surface":"8212633349","language":"fa","field":null},{"start":63,"end":74,"entity_type":"PHONE","surface":"۰۹۱۲۷۴۸۴۳۶۷","language":"fa","field":null}],"gold_meta":[],"hard_negatives":["4063697593","1403/05/12","۱۴۰۳/۰۵/۱۲","1.0.0"],"description":"Persian contract with a failed national-id left for review.","description_fa":"اخطار قراردادی با کد ملی نامعتبر برای بازبینی."}
51
+ {"doc_id":"DLP-PDF-MEMO-01","template":"en_pdf_memo","family":"en_pdf_memo","split":"test","language":"en","format":"pdf","title":"PDF memo","body":"Internal memo\nAuthor: Omar Fixture\nContact: Omar.Fixture01@aria-lab.test / 09127692383\nMr. Protocol is the software module, not a person.\n","metadata":{"author":"Omar Fixture","keywords":"Omar.Fixture01@aria-lab.test","title":"Memo 1"},"gold_spans":[{"start":22,"end":34,"entity_type":"PERSON","surface":"Omar Fixture","language":"en","field":null},{"start":44,"end":72,"entity_type":"EMAIL","surface":"Omar.Fixture01@aria-lab.test","language":"en","field":null},{"start":75,"end":86,"entity_type":"PHONE","surface":"09127692383","language":"en","field":null}],"gold_meta":[{"field":"author","entity_type":"PERSON","surface":"Omar Fixture"},{"field":"keywords","entity_type":"EMAIL","surface":"Omar.Fixture01@aria-lab.test"}],"hard_negatives":["Mr. Protocol"],"description":"Text PDF whose metadata also repeats the author mailbox.","description_fa":"پی‌دی‌اف متنی که فراداده هم نام و ایمیل را تکرار می‌کند."}
52
+ {"doc_id":"DLP-PDF-MEMO-02","template":"en_pdf_memo","family":"en_pdf_memo","split":"test","language":"en","format":"pdf","title":"PDF memo","body":"Internal memo\nAuthor: Noah Fixture\nContact: Noah.Fixture02@aria-lab.test / 09125328205\nMr. Protocol is the software module, not a person.\n","metadata":{"author":"Noah Fixture","keywords":"Noah.Fixture02@aria-lab.test","title":"Memo 2"},"gold_spans":[{"start":22,"end":34,"entity_type":"PERSON","surface":"Noah Fixture","language":"en","field":null},{"start":44,"end":72,"entity_type":"EMAIL","surface":"Noah.Fixture02@aria-lab.test","language":"en","field":null},{"start":75,"end":86,"entity_type":"PHONE","surface":"09125328205","language":"en","field":null}],"gold_meta":[{"field":"author","entity_type":"PERSON","surface":"Noah Fixture"},{"field":"keywords","entity_type":"EMAIL","surface":"Noah.Fixture02@aria-lab.test"}],"hard_negatives":["Mr. Protocol"],"description":"Text PDF whose metadata also repeats the author mailbox.","description_fa":"پی‌دی‌اف متنی که فراداده هم نام و ایمیل را تکرار می‌کند."}
53
+ {"doc_id":"DLP-PDF-MEMO-03","template":"en_pdf_memo","family":"en_pdf_memo","split":"test","language":"en","format":"pdf","title":"PDF memo","body":"Internal memo\nAuthor: Omar Labian\nContact: Omar.Labian03@aria-lab.test / 09125603855\nMr. Protocol is the software module, not a person.\n","metadata":{"author":"Omar Labian","keywords":"Omar.Labian03@aria-lab.test","title":"Memo 3"},"gold_spans":[{"start":22,"end":33,"entity_type":"PERSON","surface":"Omar Labian","language":"en","field":null},{"start":43,"end":70,"entity_type":"EMAIL","surface":"Omar.Labian03@aria-lab.test","language":"en","field":null},{"start":73,"end":84,"entity_type":"PHONE","surface":"09125603855","language":"en","field":null}],"gold_meta":[{"field":"author","entity_type":"PERSON","surface":"Omar Labian"},{"field":"keywords","entity_type":"EMAIL","surface":"Omar.Labian03@aria-lab.test"}],"hard_negatives":["Mr. Protocol"],"description":"Text PDF whose metadata also repeats the author mailbox.","description_fa":"پی‌دی‌اف متنی که فراداده هم نام و ایمیل را تکرار می‌کند."}
54
+ {"doc_id":"DLP-PDF-MEMO-04","template":"en_pdf_memo","family":"en_pdf_memo","split":"test","language":"en","format":"pdf","title":"PDF memo","body":"Internal memo\nAuthor: Noah Labian\nContact: Noah.Labian04@aria-lab.test / 09127591902\nMr. Protocol is the software module, not a person.\n","metadata":{"author":"Noah Labian","keywords":"Noah.Labian04@aria-lab.test","title":"Memo 4"},"gold_spans":[{"start":22,"end":33,"entity_type":"PERSON","surface":"Noah Labian","language":"en","field":null},{"start":43,"end":70,"entity_type":"EMAIL","surface":"Noah.Labian04@aria-lab.test","language":"en","field":null},{"start":73,"end":84,"entity_type":"PHONE","surface":"09127591902","language":"en","field":null}],"gold_meta":[{"field":"author","entity_type":"PERSON","surface":"Noah Labian"},{"field":"keywords","entity_type":"EMAIL","surface":"Noah.Labian04@aria-lab.test"}],"hard_negatives":["Mr. Protocol"],"description":"Text PDF whose metadata also repeats the author mailbox.","description_fa":"پی‌دی‌اف متنی که فراداده هم نام و ایمیل را تکرار می‌کند."}
55
+ {"doc_id":"DLP-PDF-MEMO-05","template":"en_pdf_memo","family":"en_pdf_memo","split":"test","language":"en","format":"pdf","title":"PDF memo","body":"Internal memo\nAuthor: Nima Labian\nContact: Nima.Labian05@aria-lab.test / 09129473675\nMr. Protocol is the software module, not a person.\n","metadata":{"author":"Nima Labian","keywords":"Nima.Labian05@aria-lab.test","title":"Memo 5"},"gold_spans":[{"start":22,"end":33,"entity_type":"PERSON","surface":"Nima Labian","language":"en","field":null},{"start":43,"end":70,"entity_type":"EMAIL","surface":"Nima.Labian05@aria-lab.test","language":"en","field":null},{"start":73,"end":84,"entity_type":"PHONE","surface":"09129473675","language":"en","field":null}],"gold_meta":[{"field":"author","entity_type":"PERSON","surface":"Nima Labian"},{"field":"keywords","entity_type":"EMAIL","surface":"Nima.Labian05@aria-lab.test"}],"hard_negatives":["Mr. Protocol"],"description":"Text PDF whose metadata also repeats the author mailbox.","description_fa":"پی‌دی‌اف متنی که فراداده هم نام و ایمیل را تکرار می‌کند."}
56
+ {"doc_id":"DLP-PDF-MEMO-06","template":"en_pdf_memo","family":"en_pdf_memo","split":"test","language":"en","format":"pdf","title":"PDF memo","body":"Internal memo\nAuthor: Helen Fixture\nContact: Helen.Fixture06@aria-lab.test / 09126506157\nMr. Protocol is the software module, not a person.\n","metadata":{"author":"Helen Fixture","keywords":"Helen.Fixture06@aria-lab.test","title":"Memo 6"},"gold_spans":[{"start":22,"end":35,"entity_type":"PERSON","surface":"Helen Fixture","language":"en","field":null},{"start":45,"end":74,"entity_type":"EMAIL","surface":"Helen.Fixture06@aria-lab.test","language":"en","field":null},{"start":77,"end":88,"entity_type":"PHONE","surface":"09126506157","language":"en","field":null}],"gold_meta":[{"field":"author","entity_type":"PERSON","surface":"Helen Fixture"},{"field":"keywords","entity_type":"EMAIL","surface":"Helen.Fixture06@aria-lab.test"}],"hard_negatives":["Mr. Protocol"],"description":"Text PDF whose metadata also repeats the author mailbox.","description_fa":"پی‌دی‌اف متنی که فراداده هم نام و ایمیل را تکرار می‌کند."}
57
+ {"doc_id":"DLP-PDF-MEMO-07","template":"en_pdf_memo","family":"en_pdf_memo","split":"test","language":"en","format":"pdf","title":"PDF memo","body":"Internal memo\nAuthor: Ava Fixture\nContact: Ava.Fixture07@aria-lab.test / 09126633882\nMr. Protocol is the software module, not a person.\n","metadata":{"author":"Ava Fixture","keywords":"Ava.Fixture07@aria-lab.test","title":"Memo 7"},"gold_spans":[{"start":22,"end":33,"entity_type":"PERSON","surface":"Ava Fixture","language":"en","field":null},{"start":43,"end":70,"entity_type":"EMAIL","surface":"Ava.Fixture07@aria-lab.test","language":"en","field":null},{"start":73,"end":84,"entity_type":"PHONE","surface":"09126633882","language":"en","field":null}],"gold_meta":[{"field":"author","entity_type":"PERSON","surface":"Ava Fixture"},{"field":"keywords","entity_type":"EMAIL","surface":"Ava.Fixture07@aria-lab.test"}],"hard_negatives":["Mr. Protocol"],"description":"Text PDF whose metadata also repeats the author mailbox.","description_fa":"پی‌دی‌اف متنی که فراداده هم نام و ایمیل را تکرار می‌کند."}
58
+ {"doc_id":"DLP-PDF-MEMO-08","template":"en_pdf_memo","family":"en_pdf_memo","split":"test","language":"en","format":"pdf","title":"PDF memo","body":"Internal memo\nAuthor: Nima Labian\nContact: Nima.Labian08@aria-lab.test / 09128661790\nMr. Protocol is the software module, not a person.\n","metadata":{"author":"Nima Labian","keywords":"Nima.Labian08@aria-lab.test","title":"Memo 8"},"gold_spans":[{"start":22,"end":33,"entity_type":"PERSON","surface":"Nima Labian","language":"en","field":null},{"start":43,"end":70,"entity_type":"EMAIL","surface":"Nima.Labian08@aria-lab.test","language":"en","field":null},{"start":73,"end":84,"entity_type":"PHONE","surface":"09128661790","language":"en","field":null}],"gold_meta":[{"field":"author","entity_type":"PERSON","surface":"Nima Labian"},{"field":"keywords","entity_type":"EMAIL","surface":"Nima.Labian08@aria-lab.test"}],"hard_negatives":["Mr. Protocol"],"description":"Text PDF whose metadata also repeats the author mailbox.","description_fa":"پی‌دی‌اف متنی که فراداده هم نام و ایمیل را تکرار می‌کند."}
59
+ {"doc_id":"DLP-PDF-MEMO-09","template":"en_pdf_memo","family":"en_pdf_memo","split":"test","language":"en","format":"pdf","title":"PDF memo","body":"Internal memo\nAuthor: Ava Fixture\nContact: Ava.Fixture09@aria-lab.test / 09122509645\nMr. Protocol is the software module, not a person.\n","metadata":{"author":"Ava Fixture","keywords":"Ava.Fixture09@aria-lab.test","title":"Memo 9"},"gold_spans":[{"start":22,"end":33,"entity_type":"PERSON","surface":"Ava Fixture","language":"en","field":null},{"start":43,"end":70,"entity_type":"EMAIL","surface":"Ava.Fixture09@aria-lab.test","language":"en","field":null},{"start":73,"end":84,"entity_type":"PHONE","surface":"09122509645","language":"en","field":null}],"gold_meta":[{"field":"author","entity_type":"PERSON","surface":"Ava Fixture"},{"field":"keywords","entity_type":"EMAIL","surface":"Ava.Fixture09@aria-lab.test"}],"hard_negatives":["Mr. Protocol"],"description":"Text PDF whose metadata also repeats the author mailbox.","description_fa":"پی‌دی‌اف متنی که فراداده هم نام و ایمیل را تکرار می‌کند."}
60
+ {"doc_id":"DLP-PDF-MEMO-10","template":"en_pdf_memo","family":"en_pdf_memo","split":"test","language":"en","format":"pdf","title":"PDF memo","body":"Internal memo\nAuthor: Nima Fixture\nContact: Nima.Fixture10@aria-lab.test / 09122528399\nMr. Protocol is the software module, not a person.\n","metadata":{"author":"Nima Fixture","keywords":"Nima.Fixture10@aria-lab.test","title":"Memo 10"},"gold_spans":[{"start":22,"end":34,"entity_type":"PERSON","surface":"Nima Fixture","language":"en","field":null},{"start":44,"end":72,"entity_type":"EMAIL","surface":"Nima.Fixture10@aria-lab.test","language":"en","field":null},{"start":75,"end":86,"entity_type":"PHONE","surface":"09122528399","language":"en","field":null}],"gold_meta":[{"field":"author","entity_type":"PERSON","surface":"Nima Fixture"},{"field":"keywords","entity_type":"EMAIL","surface":"Nima.Fixture10@aria-lab.test"}],"hard_negatives":["Mr. Protocol"],"description":"Text PDF whose metadata also repeats the author mailbox.","description_fa":"پی‌دی‌اف متنی که فراداده هم نام و ایمیل را تکرار می‌کند."}
61
+ {"doc_id":"DLP-HARDNEG-01","template":"hardneg_catalog","family":"hardneg_catalog","split":"test","language":"en","format":"text","title":"Hard-negative catalog","body":"Catalog row 1\nISBN 978-0-306-40615-7\nInvoice 9882096604\nCard-like 4111799672516487\nIBAN-like IR969760340568087700433687\nMr. Protocol handler v2.4.1\nDate 2026/09/21 SKU-9001\n","metadata":{},"gold_spans":[],"gold_meta":[],"hard_negatives":["4111799672516487","IR969760340568087700433687","9882096604","Mr. Protocol","2.4.1","978-0-306-40615-7"],"description":"Lookalikes that must not be redacted: failed checksums, ISBN, protocol title.","description_fa":"شبیه‌سازی‌هایی که نباید پوشانده شوند: checksum نامعتبر، شابک، عنوان پروتکل."}
62
+ {"doc_id":"DLP-HARDNEG-02","template":"hardneg_catalog","family":"hardneg_catalog","split":"test","language":"en","format":"text","title":"Hard-negative catalog","body":"Catalog row 2\nISBN 978-0-306-40615-7\nInvoice 7209502842\nCard-like 4111134559932047\nIBAN-like IR859360346783598885359083\nMr. Protocol handler v2.4.1\nDate 2026/09/21 SKU-9002\n","metadata":{},"gold_spans":[],"gold_meta":[],"hard_negatives":["4111134559932047","IR859360346783598885359083","7209502842","Mr. Protocol","2.4.1","978-0-306-40615-7"],"description":"Lookalikes that must not be redacted: failed checksums, ISBN, protocol title.","description_fa":"شبیه‌سازی‌هایی که نباید پوشانده شوند: checksum نامعتبر، شابک، عنوان پروتکل."}
63
+ {"doc_id":"DLP-HARDNEG-03","template":"hardneg_catalog","family":"hardneg_catalog","split":"test","language":"en","format":"text","title":"Hard-negative catalog","body":"Catalog row 3\nISBN 978-0-306-40615-7\nInvoice 8163562111\nCard-like 4111842764173344\nIBAN-like IR189190290458319906920450\nMr. Protocol handler v2.4.1\nDate 2026/09/21 SKU-9003\n","metadata":{},"gold_spans":[],"gold_meta":[],"hard_negatives":["4111842764173344","IR189190290458319906920450","8163562111","Mr. Protocol","2.4.1","978-0-306-40615-7"],"description":"Lookalikes that must not be redacted: failed checksums, ISBN, protocol title.","description_fa":"شبیه‌سازی‌هایی که نباید پوشانده شوند: checksum نامعتبر، شابک، عنوان پروتکل."}
64
+ {"doc_id":"DLP-HARDNEG-04","template":"hardneg_catalog","family":"hardneg_catalog","split":"test","language":"en","format":"text","title":"Hard-negative catalog","body":"Catalog row 4\nISBN 978-0-306-40615-7\nInvoice 2665961572\nCard-like 4111391694332034\nIBAN-like IR289390013050842973101189\nMr. Protocol handler v2.4.1\nDate 2026/09/21 SKU-9004\n","metadata":{},"gold_spans":[],"gold_meta":[],"hard_negatives":["4111391694332034","IR289390013050842973101189","2665961572","Mr. Protocol","2.4.1","978-0-306-40615-7"],"description":"Lookalikes that must not be redacted: failed checksums, ISBN, protocol title.","description_fa":"شبیه‌سازی‌هایی که نباید پوشانده شوند: checksum نامعتبر، شابک، عنوان پروتکل."}
65
+ {"doc_id":"DLP-HARDNEG-05","template":"hardneg_catalog","family":"hardneg_catalog","split":"test","language":"en","format":"text","title":"Hard-negative catalog","body":"Catalog row 5\nISBN 978-0-306-40615-7\nInvoice 5490470241\nCard-like 4111874194171432\nIBAN-like IR549770672881619226418819\nMr. Protocol handler v2.4.1\nDate 2026/09/21 SKU-9005\n","metadata":{},"gold_spans":[],"gold_meta":[],"hard_negatives":["4111874194171432","IR549770672881619226418819","5490470241","Mr. Protocol","2.4.1","978-0-306-40615-7"],"description":"Lookalikes that must not be redacted: failed checksums, ISBN, protocol title.","description_fa":"شبیه‌سازی‌هایی که نباید پوشانده شوند: checksum نامعتبر، شابک، عنوان پروتکل."}
66
+ {"doc_id":"DLP-HARDNEG-06","template":"hardneg_catalog","family":"hardneg_catalog","split":"test","language":"en","format":"text","title":"Hard-negative catalog","body":"Catalog row 6\nISBN 978-0-306-40615-7\nInvoice 9997521600\nCard-like 4111260024722695\nIBAN-like IR299310569176970687717687\nMr. Protocol handler v2.4.1\nDate 2026/09/21 SKU-9006\n","metadata":{},"gold_spans":[],"gold_meta":[],"hard_negatives":["4111260024722695","IR299310569176970687717687","9997521600","Mr. Protocol","2.4.1","978-0-306-40615-7"],"description":"Lookalikes that must not be redacted: failed checksums, ISBN, protocol title.","description_fa":"شبیه‌سازی‌هایی که نباید پوشانده شوند: checksum نامعتبر، شابک، عنوان پروتکل."}
67
+ {"doc_id":"DLP-HARDNEG-07","template":"hardneg_catalog","family":"hardneg_catalog","split":"test","language":"en","format":"text","title":"Hard-negative catalog","body":"Catalog row 7\nISBN 978-0-306-40615-7\nInvoice 4413142056\nCard-like 4111156662872018\nIBAN-like IR849380980760628231287991\nMr. Protocol handler v2.4.1\nDate 2026/09/21 SKU-9007\n","metadata":{},"gold_spans":[],"gold_meta":[],"hard_negatives":["4111156662872018","IR849380980760628231287991","4413142056","Mr. Protocol","2.4.1","978-0-306-40615-7"],"description":"Lookalikes that must not be redacted: failed checksums, ISBN, protocol title.","description_fa":"شبیه‌سازی‌هایی که نباید پوشانده شوند: checksum نامعتبر، شابک، عنوان پروتکل."}
68
+ {"doc_id":"DLP-HARDNEG-08","template":"hardneg_catalog","family":"hardneg_catalog","split":"test","language":"en","format":"text","title":"Hard-negative catalog","body":"Catalog row 8\nISBN 978-0-306-40615-7\nInvoice 9844497419\nCard-like 4111658027974926\nIBAN-like IR749310432272162828796518\nMr. Protocol handler v2.4.1\nDate 2026/09/21 SKU-9008\n","metadata":{},"gold_spans":[],"gold_meta":[],"hard_negatives":["4111658027974926","IR749310432272162828796518","9844497419","Mr. Protocol","2.4.1","978-0-306-40615-7"],"description":"Lookalikes that must not be redacted: failed checksums, ISBN, protocol title.","description_fa":"شبیه‌سازی‌هایی که نباید پوشانده شوند: checksum نامعتبر، شابک، عنوان پروتکل."}
69
+ {"doc_id":"DLP-HARDNEG-09","template":"hardneg_catalog","family":"hardneg_catalog","split":"test","language":"en","format":"text","title":"Hard-negative catalog","body":"Catalog row 9\nISBN 978-0-306-40615-7\nInvoice 5948674852\nCard-like 4111784726661737\nIBAN-like IR529710074259549882292605\nMr. Protocol handler v2.4.1\nDate 2026/09/21 SKU-9009\n","metadata":{},"gold_spans":[],"gold_meta":[],"hard_negatives":["4111784726661737","IR529710074259549882292605","5948674852","Mr. Protocol","2.4.1","978-0-306-40615-7"],"description":"Lookalikes that must not be redacted: failed checksums, ISBN, protocol title.","description_fa":"شبیه‌سازی‌هایی که نباید پوشانده شوند: checksum نامعتبر، شابک، عنوان پروتکل."}
70
+ {"doc_id":"DLP-HARDNEG-10","template":"hardneg_catalog","family":"hardneg_catalog","split":"test","language":"en","format":"text","title":"Hard-negative catalog","body":"Catalog row 10\nISBN 978-0-306-40615-7\nInvoice 2672086444\nCard-like 4111733944863765\nIBAN-like IR739740106542912708354792\nMr. Protocol handler v2.4.1\nDate 2026/09/21 SKU-9010\n","metadata":{},"gold_spans":[],"gold_meta":[],"hard_negatives":["4111733944863765","IR739740106542912708354792","2672086444","Mr. Protocol","2.4.1","978-0-306-40615-7"],"description":"Lookalikes that must not be redacted: failed checksums, ISBN, protocol title.","description_fa":"شبیه‌سازی‌هایی که نباید پوشانده شوند: checksum نامعتبر، شابک، عنوان پروتکل."}
71
+ {"doc_id":"DLP-TICKET-01","template":"mixed_ticket","family":"mixed_ticket","split":"test","language":"mixed","format":"csv","title":"Mixed-language ticket CSV","body":"ticket_id,reporter,email,phone,national_id\nINV-MIX-0001,Leila Fixture / سارا آزمایشی,Leila.Fixture01@aria-lab.test,09126663672,8810500075\n","metadata":{},"gold_spans":[{"start":56,"end":69,"entity_type":"PERSON","surface":"Leila Fixture","language":"en","field":null},{"start":72,"end":84,"entity_type":"PERSON","surface":"سارا آزمایشی","language":"fa","field":null},{"start":85,"end":114,"entity_type":"EMAIL","surface":"Leila.Fixture01@aria-lab.test","language":"en","field":null},{"start":115,"end":126,"entity_type":"PHONE","surface":"09126663672","language":"en","field":null},{"start":127,"end":137,"entity_type":"IR_NATIONAL_ID","surface":"8810500075","language":"fa","field":null}],"gold_meta":[],"hard_negatives":["INV-MIX-0001"],"description":"Mixed English/Persian ticket row.","description_fa":"ردیف تیکت دوزبانه."}
72
+ {"doc_id":"DLP-TICKET-02","template":"mixed_ticket","family":"mixed_ticket","split":"test","language":"mixed","format":"csv","title":"Mixed-language ticket CSV","body":"ticket_id,reporter,email,phone,national_id\nINV-MIX-0002,Leila Fixture / امید نمونه‌ای,Leila.Fixture02@aria-lab.test,09124960362,7433177132\n","metadata":{},"gold_spans":[{"start":56,"end":69,"entity_type":"PERSON","surface":"Leila Fixture","language":"en","field":null},{"start":72,"end":85,"entity_type":"PERSON","surface":"امید نمونه‌ای","language":"fa","field":null},{"start":86,"end":115,"entity_type":"EMAIL","surface":"Leila.Fixture02@aria-lab.test","language":"en","field":null},{"start":116,"end":127,"entity_type":"PHONE","surface":"09124960362","language":"en","field":null},{"start":128,"end":138,"entity_type":"IR_NATIONAL_ID","surface":"7433177132","language":"fa","field":null}],"gold_meta":[],"hard_negatives":["INV-MIX-0002"],"description":"Mixed English/Persian ticket row.","description_fa":"ردیف تیکت دوزبانه."}
73
+ {"doc_id":"DLP-TICKET-03","template":"mixed_ticket","family":"mixed_ticket","split":"test","language":"mixed","format":"csv","title":"Mixed-language ticket CSV","body":"ticket_id,reporter,email,phone,national_id\nINV-MIX-0003,Ava Fixture / نیما نمونه‌ای,Ava.Fixture03@aria-lab.test,09129013939,4751024639\n","metadata":{},"gold_spans":[{"start":56,"end":67,"entity_type":"PERSON","surface":"Ava Fixture","language":"en","field":null},{"start":70,"end":83,"entity_type":"PERSON","surface":"نیما نمونه‌ای","language":"fa","field":null},{"start":84,"end":111,"entity_type":"EMAIL","surface":"Ava.Fixture03@aria-lab.test","language":"en","field":null},{"start":112,"end":123,"entity_type":"PHONE","surface":"09129013939","language":"en","field":null},{"start":124,"end":134,"entity_type":"IR_NATIONAL_ID","surface":"4751024639","language":"fa","field":null}],"gold_meta":[],"hard_negatives":["INV-MIX-0003"],"description":"Mixed English/Persian ticket row.","description_fa":"ردیف تیکت دوزبانه."}
74
+ {"doc_id":"DLP-TICKET-04","template":"mixed_ticket","family":"mixed_ticket","split":"test","language":"mixed","format":"csv","title":"Mixed-language ticket CSV","body":"ticket_id,reporter,email,phone,national_id\nINV-MIX-0004,Omar Fixture / سارا نمونه‌ای,Omar.Fixture04@aria-lab.test,09124265160,0843397276\n","metadata":{},"gold_spans":[{"start":56,"end":68,"entity_type":"PERSON","surface":"Omar Fixture","language":"en","field":null},{"start":71,"end":84,"entity_type":"PERSON","surface":"سارا نمونه‌ای","language":"fa","field":null},{"start":85,"end":113,"entity_type":"EMAIL","surface":"Omar.Fixture04@aria-lab.test","language":"en","field":null},{"start":114,"end":125,"entity_type":"PHONE","surface":"09124265160","language":"en","field":null},{"start":126,"end":136,"entity_type":"IR_NATIONAL_ID","surface":"0843397276","language":"fa","field":null}],"gold_meta":[],"hard_negatives":["INV-MIX-0004"],"description":"Mixed English/Persian ticket row.","description_fa":"ردیف تیکت دوزبانه."}
75
+ {"doc_id":"DLP-TICKET-05","template":"mixed_ticket","family":"mixed_ticket","split":"test","language":"mixed","format":"csv","title":"Mixed-language ticket CSV","body":"ticket_id,reporter,email,phone,national_id\nINV-MIX-0005,Ava Fixture / امید نمونه‌ای,Ava.Fixture05@aria-lab.test,09127385597,6714749521\n","metadata":{},"gold_spans":[{"start":56,"end":67,"entity_type":"PERSON","surface":"Ava Fixture","language":"en","field":null},{"start":70,"end":83,"entity_type":"PERSON","surface":"امید نمونه‌ای","language":"fa","field":null},{"start":84,"end":111,"entity_type":"EMAIL","surface":"Ava.Fixture05@aria-lab.test","language":"en","field":null},{"start":112,"end":123,"entity_type":"PHONE","surface":"09127385597","language":"en","field":null},{"start":124,"end":134,"entity_type":"IR_NATIONAL_ID","surface":"6714749521","language":"fa","field":null}],"gold_meta":[],"hard_negatives":["INV-MIX-0005"],"description":"Mixed English/Persian ticket row.","description_fa":"ردیف تیکت دوزبانه."}
76
+ {"doc_id":"DLP-TICKET-06","template":"mixed_ticket","family":"mixed_ticket","split":"test","language":"mixed","format":"csv","title":"Mixed-language ticket CSV","body":"ticket_id,reporter,email,phone,national_id\nINV-MIX-0006,Omar Fixture / لیلا آزمایشی,Omar.Fixture06@aria-lab.test,09126454625,7904846901\n","metadata":{},"gold_spans":[{"start":56,"end":68,"entity_type":"PERSON","surface":"Omar Fixture","language":"en","field":null},{"start":71,"end":83,"entity_type":"PERSON","surface":"لیلا آزمایشی","language":"fa","field":null},{"start":84,"end":112,"entity_type":"EMAIL","surface":"Omar.Fixture06@aria-lab.test","language":"en","field":null},{"start":113,"end":124,"entity_type":"PHONE","surface":"09126454625","language":"en","field":null},{"start":125,"end":135,"entity_type":"IR_NATIONAL_ID","surface":"7904846901","language":"fa","field":null}],"gold_meta":[],"hard_negatives":["INV-MIX-0006"],"description":"Mixed English/Persian ticket row.","description_fa":"ردیف تیکت دوزبانه."}
77
+ {"doc_id":"DLP-TICKET-07","template":"mixed_ticket","family":"mixed_ticket","split":"test","language":"mixed","format":"csv","title":"Mixed-language ticket CSV","body":"ticket_id,reporter,email,phone,national_id\nINV-MIX-0007,Helen Fixture / سارا آزمایشی,Helen.Fixture07@aria-lab.test,09127938081,6114411269\n","metadata":{},"gold_spans":[{"start":56,"end":69,"entity_type":"PERSON","surface":"Helen Fixture","language":"en","field":null},{"start":72,"end":84,"entity_type":"PERSON","surface":"سارا آزمایشی","language":"fa","field":null},{"start":85,"end":114,"entity_type":"EMAIL","surface":"Helen.Fixture07@aria-lab.test","language":"en","field":null},{"start":115,"end":126,"entity_type":"PHONE","surface":"09127938081","language":"en","field":null},{"start":127,"end":137,"entity_type":"IR_NATIONAL_ID","surface":"6114411269","language":"fa","field":null}],"gold_meta":[],"hard_negatives":["INV-MIX-0007"],"description":"Mixed English/Persian ticket row.","description_fa":"ردیف تیکت دوزبانه."}
78
+ {"doc_id":"DLP-TICKET-08","template":"mixed_ticket","family":"mixed_ticket","split":"test","language":"mixed","format":"csv","title":"Mixed-language ticket CSV","body":"ticket_id,reporter,email,phone,national_id\nINV-MIX-0008,Ava Fixture / لیلا نمونه‌ای,Ava.Fixture08@aria-lab.test,09125317904,2799826970\n","metadata":{},"gold_spans":[{"start":56,"end":67,"entity_type":"PERSON","surface":"Ava Fixture","language":"en","field":null},{"start":70,"end":83,"entity_type":"PERSON","surface":"لیلا نمونه‌ای","language":"fa","field":null},{"start":84,"end":111,"entity_type":"EMAIL","surface":"Ava.Fixture08@aria-lab.test","language":"en","field":null},{"start":112,"end":123,"entity_type":"PHONE","surface":"09125317904","language":"en","field":null},{"start":124,"end":134,"entity_type":"IR_NATIONAL_ID","surface":"2799826970","language":"fa","field":null}],"gold_meta":[],"hard_negatives":["INV-MIX-0008"],"description":"Mixed English/Persian ticket row.","description_fa":"ردیف تیکت دوزبانه."}
79
+ {"doc_id":"DLP-TICKET-09","template":"mixed_ticket","family":"mixed_ticket","split":"test","language":"mixed","format":"csv","title":"Mixed-language ticket CSV","body":"ticket_id,reporter,email,phone,national_id\nINV-MIX-0009,Omar Fixture / نیما نمونه‌ای,Omar.Fixture09@aria-lab.test,09125714005,0468470395\n","metadata":{},"gold_spans":[{"start":56,"end":68,"entity_type":"PERSON","surface":"Omar Fixture","language":"en","field":null},{"start":71,"end":84,"entity_type":"PERSON","surface":"نیما نمونه‌ای","language":"fa","field":null},{"start":85,"end":113,"entity_type":"EMAIL","surface":"Omar.Fixture09@aria-lab.test","language":"en","field":null},{"start":114,"end":125,"entity_type":"PHONE","surface":"09125714005","language":"en","field":null},{"start":126,"end":136,"entity_type":"IR_NATIONAL_ID","surface":"0468470395","language":"fa","field":null}],"gold_meta":[],"hard_negatives":["INV-MIX-0009"],"description":"Mixed English/Persian ticket row.","description_fa":"ردیف تیکت دوزبانه."}
80
+ {"doc_id":"DLP-TICKET-10","template":"mixed_ticket","family":"mixed_ticket","split":"test","language":"mixed","format":"csv","title":"Mixed-language ticket CSV","body":"ticket_id,reporter,email,phone,national_id\nINV-MIX-0010,Helen Labian / سارا نمونه‌ای,Helen.Labian10@aria-lab.test,09126818844,5049585570\n","metadata":{},"gold_spans":[{"start":56,"end":68,"entity_type":"PERSON","surface":"Helen Labian","language":"en","field":null},{"start":71,"end":84,"entity_type":"PERSON","surface":"سارا نمونه‌ای","language":"fa","field":null},{"start":85,"end":113,"entity_type":"EMAIL","surface":"Helen.Labian10@aria-lab.test","language":"en","field":null},{"start":114,"end":125,"entity_type":"PHONE","surface":"09126818844","language":"en","field":null},{"start":126,"end":136,"entity_type":"IR_NATIONAL_ID","surface":"5049585570","language":"fa","field":null}],"gold_meta":[],"hard_negatives":["INV-MIX-0010"],"description":"Mixed English/Persian ticket row.","description_fa":"ردیف تیکت دوزبانه."}
data/evaluation.json ADDED
@@ -0,0 +1,88 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "seed": 24,
3
+ "n_documents": 80,
4
+ "n_scenarios": 5,
5
+ "engine_profile": "presidio-style-checksum-policy",
6
+ "rows": [
7
+ {
8
+ "scenario_id": "S01",
9
+ "method": "checksum_policy",
10
+ "precision": 1.0,
11
+ "recall": 1.0,
12
+ "over_redaction": 0.0,
13
+ "residual_leak": 0.0,
14
+ "leftover_exports": 0,
15
+ "metadata_leaks": 0,
16
+ "hardneg_hits": 0,
17
+ "n_docs": 20
18
+ },
19
+ {
20
+ "scenario_id": "S02",
21
+ "method": "checksum_policy",
22
+ "precision": 1.0,
23
+ "recall": 1.0,
24
+ "over_redaction": 0.0,
25
+ "residual_leak": 0.0,
26
+ "leftover_exports": 0,
27
+ "metadata_leaks": 0,
28
+ "hardneg_hits": 0,
29
+ "n_docs": 20
30
+ },
31
+ {
32
+ "scenario_id": "S03",
33
+ "method": "checksum_policy",
34
+ "precision": 1.0,
35
+ "recall": 1.0,
36
+ "over_redaction": 0.0,
37
+ "residual_leak": 0.0,
38
+ "leftover_exports": 0,
39
+ "metadata_leaks": 0,
40
+ "hardneg_hits": 0,
41
+ "n_docs": 10
42
+ },
43
+ {
44
+ "scenario_id": "S04",
45
+ "method": "checksum_policy",
46
+ "precision": 1.0,
47
+ "recall": 1.0,
48
+ "over_redaction": 0.0,
49
+ "residual_leak": 0.0,
50
+ "leftover_exports": 0,
51
+ "metadata_leaks": 0,
52
+ "hardneg_hits": 10,
53
+ "n_docs": 20
54
+ },
55
+ {
56
+ "scenario_id": "S05",
57
+ "method": "checksum_policy",
58
+ "precision": 1.0,
59
+ "recall": 1.0,
60
+ "over_redaction": 0.0,
61
+ "residual_leak": 0.0,
62
+ "leftover_exports": 0,
63
+ "metadata_leaks": 0,
64
+ "hardneg_hits": 10,
65
+ "n_docs": 80
66
+ }
67
+ ],
68
+ "sample_kpis": {
69
+ "n_docs": 20,
70
+ "n_spans": 70,
71
+ "n_review": 0,
72
+ "n_redacted": 70,
73
+ "policy_precision": 1.0,
74
+ "policy_recall": 1.0,
75
+ "baseline_precision": 0.8333333333333334,
76
+ "baseline_recall": 0.7142857142857143,
77
+ "residual_leak": 0.0,
78
+ "over_redaction": 0.0,
79
+ "leftover_exports": 0,
80
+ "metadata_leaks": 0,
81
+ "elapsed_ms": 52,
82
+ "engine_profile": "presidio-style-checksum-policy",
83
+ "policy_action": "mask",
84
+ "pages": 0,
85
+ "chars": 2897
86
+ },
87
+ "note": "Synthetic laboratory corpus. A clean suite is not complete production redaction."
88
+ }
data/feature_card.md ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ # Feature card
2
+
3
+ - Task: span detection and reversible laboratory redaction
4
+ - Languages: English, Persian (digits, ZWNJ)
5
+ - Formats: text, CSV, text PDF (≤10 pages)
6
+ - Baseline: ASCII regex without checksum
7
+ - Policy: Presidio-style recognizer registry + checksum + lexicon
8
+ - Split: template-disjoint
data/gold_spans.json ADDED
@@ -0,0 +1,2662 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ [
2
+ {
3
+ "doc_id": "DLP-EN-EMAIL-01",
4
+ "spans": [
5
+ {
6
+ "start": 6,
7
+ "end": 18,
8
+ "entity_type": "PERSON",
9
+ "surface": "Nima Fixture",
10
+ "language": "en",
11
+ "field": null
12
+ },
13
+ {
14
+ "start": 20,
15
+ "end": 48,
16
+ "entity_type": "EMAIL",
17
+ "surface": "Nima.Fixture01@aria-lab.test",
18
+ "language": "en",
19
+ "field": null
20
+ },
21
+ {
22
+ "start": 57,
23
+ "end": 68,
24
+ "entity_type": "PHONE",
25
+ "surface": "09124063383",
26
+ "language": "en",
27
+ "field": null
28
+ },
29
+ {
30
+ "start": 79,
31
+ "end": 89,
32
+ "entity_type": "EMPLOYEE_ID",
33
+ "surface": "ARIA-24001",
34
+ "language": "en",
35
+ "field": null
36
+ }
37
+ ],
38
+ "meta": []
39
+ },
40
+ {
41
+ "doc_id": "DLP-EN-EMAIL-02",
42
+ "spans": [
43
+ {
44
+ "start": 6,
45
+ "end": 17,
46
+ "entity_type": "PERSON",
47
+ "surface": "Noah Labian",
48
+ "language": "en",
49
+ "field": null
50
+ },
51
+ {
52
+ "start": 19,
53
+ "end": 46,
54
+ "entity_type": "EMAIL",
55
+ "surface": "Noah.Labian02@aria-lab.test",
56
+ "language": "en",
57
+ "field": null
58
+ },
59
+ {
60
+ "start": 55,
61
+ "end": 66,
62
+ "entity_type": "PHONE",
63
+ "surface": "09124256381",
64
+ "language": "en",
65
+ "field": null
66
+ },
67
+ {
68
+ "start": 77,
69
+ "end": 87,
70
+ "entity_type": "EMPLOYEE_ID",
71
+ "surface": "ARIA-24002",
72
+ "language": "en",
73
+ "field": null
74
+ }
75
+ ],
76
+ "meta": []
77
+ },
78
+ {
79
+ "doc_id": "DLP-EN-EMAIL-03",
80
+ "spans": [
81
+ {
82
+ "start": 6,
83
+ "end": 17,
84
+ "entity_type": "PERSON",
85
+ "surface": "Noah Labian",
86
+ "language": "en",
87
+ "field": null
88
+ },
89
+ {
90
+ "start": 19,
91
+ "end": 46,
92
+ "entity_type": "EMAIL",
93
+ "surface": "Noah.Labian03@aria-lab.test",
94
+ "language": "en",
95
+ "field": null
96
+ },
97
+ {
98
+ "start": 55,
99
+ "end": 66,
100
+ "entity_type": "PHONE",
101
+ "surface": "09123538397",
102
+ "language": "en",
103
+ "field": null
104
+ },
105
+ {
106
+ "start": 77,
107
+ "end": 87,
108
+ "entity_type": "EMPLOYEE_ID",
109
+ "surface": "ARIA-24003",
110
+ "language": "en",
111
+ "field": null
112
+ }
113
+ ],
114
+ "meta": []
115
+ },
116
+ {
117
+ "doc_id": "DLP-EN-EMAIL-04",
118
+ "spans": [
119
+ {
120
+ "start": 6,
121
+ "end": 18,
122
+ "entity_type": "PERSON",
123
+ "surface": "Nima Fixture",
124
+ "language": "en",
125
+ "field": null
126
+ },
127
+ {
128
+ "start": 20,
129
+ "end": 48,
130
+ "entity_type": "EMAIL",
131
+ "surface": "Nima.Fixture04@aria-lab.test",
132
+ "language": "en",
133
+ "field": null
134
+ },
135
+ {
136
+ "start": 57,
137
+ "end": 68,
138
+ "entity_type": "PHONE",
139
+ "surface": "09121213266",
140
+ "language": "en",
141
+ "field": null
142
+ },
143
+ {
144
+ "start": 79,
145
+ "end": 89,
146
+ "entity_type": "EMPLOYEE_ID",
147
+ "surface": "ARIA-24004",
148
+ "language": "en",
149
+ "field": null
150
+ }
151
+ ],
152
+ "meta": []
153
+ },
154
+ {
155
+ "doc_id": "DLP-EN-EMAIL-05",
156
+ "spans": [
157
+ {
158
+ "start": 6,
159
+ "end": 18,
160
+ "entity_type": "PERSON",
161
+ "surface": "Omar Fixture",
162
+ "language": "en",
163
+ "field": null
164
+ },
165
+ {
166
+ "start": 20,
167
+ "end": 48,
168
+ "entity_type": "EMAIL",
169
+ "surface": "Omar.Fixture05@aria-lab.test",
170
+ "language": "en",
171
+ "field": null
172
+ },
173
+ {
174
+ "start": 57,
175
+ "end": 68,
176
+ "entity_type": "PHONE",
177
+ "surface": "09122939341",
178
+ "language": "en",
179
+ "field": null
180
+ },
181
+ {
182
+ "start": 79,
183
+ "end": 89,
184
+ "entity_type": "EMPLOYEE_ID",
185
+ "surface": "ARIA-24005",
186
+ "language": "en",
187
+ "field": null
188
+ }
189
+ ],
190
+ "meta": []
191
+ },
192
+ {
193
+ "doc_id": "DLP-EN-EMAIL-06",
194
+ "spans": [
195
+ {
196
+ "start": 6,
197
+ "end": 16,
198
+ "entity_type": "PERSON",
199
+ "surface": "Ava Labian",
200
+ "language": "en",
201
+ "field": null
202
+ },
203
+ {
204
+ "start": 18,
205
+ "end": 44,
206
+ "entity_type": "EMAIL",
207
+ "surface": "Ava.Labian06@aria-lab.test",
208
+ "language": "en",
209
+ "field": null
210
+ },
211
+ {
212
+ "start": 53,
213
+ "end": 64,
214
+ "entity_type": "PHONE",
215
+ "surface": "09129242814",
216
+ "language": "en",
217
+ "field": null
218
+ },
219
+ {
220
+ "start": 75,
221
+ "end": 85,
222
+ "entity_type": "EMPLOYEE_ID",
223
+ "surface": "ARIA-24006",
224
+ "language": "en",
225
+ "field": null
226
+ }
227
+ ],
228
+ "meta": []
229
+ },
230
+ {
231
+ "doc_id": "DLP-EN-EMAIL-07",
232
+ "spans": [
233
+ {
234
+ "start": 6,
235
+ "end": 18,
236
+ "entity_type": "PERSON",
237
+ "surface": "Nima Fixture",
238
+ "language": "en",
239
+ "field": null
240
+ },
241
+ {
242
+ "start": 20,
243
+ "end": 48,
244
+ "entity_type": "EMAIL",
245
+ "surface": "Nima.Fixture07@aria-lab.test",
246
+ "language": "en",
247
+ "field": null
248
+ },
249
+ {
250
+ "start": 57,
251
+ "end": 68,
252
+ "entity_type": "PHONE",
253
+ "surface": "09126110356",
254
+ "language": "en",
255
+ "field": null
256
+ },
257
+ {
258
+ "start": 79,
259
+ "end": 89,
260
+ "entity_type": "EMPLOYEE_ID",
261
+ "surface": "ARIA-24007",
262
+ "language": "en",
263
+ "field": null
264
+ }
265
+ ],
266
+ "meta": []
267
+ },
268
+ {
269
+ "doc_id": "DLP-EN-EMAIL-08",
270
+ "spans": [
271
+ {
272
+ "start": 6,
273
+ "end": 17,
274
+ "entity_type": "PERSON",
275
+ "surface": "Omar Labian",
276
+ "language": "en",
277
+ "field": null
278
+ },
279
+ {
280
+ "start": 19,
281
+ "end": 46,
282
+ "entity_type": "EMAIL",
283
+ "surface": "Omar.Labian08@aria-lab.test",
284
+ "language": "en",
285
+ "field": null
286
+ },
287
+ {
288
+ "start": 55,
289
+ "end": 66,
290
+ "entity_type": "PHONE",
291
+ "surface": "09125259879",
292
+ "language": "en",
293
+ "field": null
294
+ },
295
+ {
296
+ "start": 77,
297
+ "end": 87,
298
+ "entity_type": "EMPLOYEE_ID",
299
+ "surface": "ARIA-24008",
300
+ "language": "en",
301
+ "field": null
302
+ }
303
+ ],
304
+ "meta": []
305
+ },
306
+ {
307
+ "doc_id": "DLP-EN-EMAIL-09",
308
+ "spans": [
309
+ {
310
+ "start": 6,
311
+ "end": 18,
312
+ "entity_type": "PERSON",
313
+ "surface": "Helen Labian",
314
+ "language": "en",
315
+ "field": null
316
+ },
317
+ {
318
+ "start": 20,
319
+ "end": 48,
320
+ "entity_type": "EMAIL",
321
+ "surface": "Helen.Labian09@aria-lab.test",
322
+ "language": "en",
323
+ "field": null
324
+ },
325
+ {
326
+ "start": 57,
327
+ "end": 68,
328
+ "entity_type": "PHONE",
329
+ "surface": "09126483978",
330
+ "language": "en",
331
+ "field": null
332
+ },
333
+ {
334
+ "start": 79,
335
+ "end": 89,
336
+ "entity_type": "EMPLOYEE_ID",
337
+ "surface": "ARIA-24009",
338
+ "language": "en",
339
+ "field": null
340
+ }
341
+ ],
342
+ "meta": []
343
+ },
344
+ {
345
+ "doc_id": "DLP-EN-EMAIL-10",
346
+ "spans": [
347
+ {
348
+ "start": 6,
349
+ "end": 18,
350
+ "entity_type": "PERSON",
351
+ "surface": "Leila Labian",
352
+ "language": "en",
353
+ "field": null
354
+ },
355
+ {
356
+ "start": 20,
357
+ "end": 48,
358
+ "entity_type": "EMAIL",
359
+ "surface": "Leila.Labian10@aria-lab.test",
360
+ "language": "en",
361
+ "field": null
362
+ },
363
+ {
364
+ "start": 57,
365
+ "end": 68,
366
+ "entity_type": "PHONE",
367
+ "surface": "09127157194",
368
+ "language": "en",
369
+ "field": null
370
+ },
371
+ {
372
+ "start": 79,
373
+ "end": 89,
374
+ "entity_type": "EMPLOYEE_ID",
375
+ "surface": "ARIA-24010",
376
+ "language": "en",
377
+ "field": null
378
+ }
379
+ ],
380
+ "meta": []
381
+ },
382
+ {
383
+ "doc_id": "DLP-FA-HR-01",
384
+ "spans": [
385
+ {
386
+ "start": 5,
387
+ "end": 18,
388
+ "entity_type": "PERSON",
389
+ "surface": "سارا نمونه‌ای",
390
+ "language": "fa",
391
+ "field": null
392
+ },
393
+ {
394
+ "start": 27,
395
+ "end": 37,
396
+ "entity_type": "IR_NATIONAL_ID",
397
+ "surface": "۵۵۱۴۱۳۹۲۷۱",
398
+ "language": "fa",
399
+ "field": null
400
+ },
401
+ {
402
+ "start": 49,
403
+ "end": 60,
404
+ "entity_type": "PHONE",
405
+ "surface": "۰۹۱۲۵۷۶۷۰۹۶",
406
+ "language": "fa",
407
+ "field": null
408
+ },
409
+ {
410
+ "start": 67,
411
+ "end": 88,
412
+ "entity_type": "EMAIL",
413
+ "surface": "user.01@aria-lab.test",
414
+ "language": "en",
415
+ "field": null
416
+ }
417
+ ],
418
+ "meta": []
419
+ },
420
+ {
421
+ "doc_id": "DLP-FA-HR-02",
422
+ "spans": [
423
+ {
424
+ "start": 5,
425
+ "end": 17,
426
+ "entity_type": "PERSON",
427
+ "surface": "لیلا آزمایشی",
428
+ "language": "fa",
429
+ "field": null
430
+ },
431
+ {
432
+ "start": 26,
433
+ "end": 36,
434
+ "entity_type": "IR_NATIONAL_ID",
435
+ "surface": "۳۱۹۷۲۲۵۷۳۲",
436
+ "language": "fa",
437
+ "field": null
438
+ },
439
+ {
440
+ "start": 48,
441
+ "end": 59,
442
+ "entity_type": "PHONE",
443
+ "surface": "۰۹۱۲۳۲۰۳۶۶۹",
444
+ "language": "fa",
445
+ "field": null
446
+ },
447
+ {
448
+ "start": 66,
449
+ "end": 87,
450
+ "entity_type": "EMAIL",
451
+ "surface": "user.02@aria-lab.test",
452
+ "language": "en",
453
+ "field": null
454
+ }
455
+ ],
456
+ "meta": []
457
+ },
458
+ {
459
+ "doc_id": "DLP-FA-HR-03",
460
+ "spans": [
461
+ {
462
+ "start": 5,
463
+ "end": 17,
464
+ "entity_type": "PERSON",
465
+ "surface": "امید آزمایشی",
466
+ "language": "fa",
467
+ "field": null
468
+ },
469
+ {
470
+ "start": 26,
471
+ "end": 36,
472
+ "entity_type": "IR_NATIONAL_ID",
473
+ "surface": "۸۳۸۳۱۲۴۵۸۹",
474
+ "language": "fa",
475
+ "field": null
476
+ },
477
+ {
478
+ "start": 48,
479
+ "end": 59,
480
+ "entity_type": "PHONE",
481
+ "surface": "۰۹۱۲۴۸۶۵۴۴۱",
482
+ "language": "fa",
483
+ "field": null
484
+ },
485
+ {
486
+ "start": 66,
487
+ "end": 87,
488
+ "entity_type": "EMAIL",
489
+ "surface": "user.03@aria-lab.test",
490
+ "language": "en",
491
+ "field": null
492
+ }
493
+ ],
494
+ "meta": []
495
+ },
496
+ {
497
+ "doc_id": "DLP-FA-HR-04",
498
+ "spans": [
499
+ {
500
+ "start": 5,
501
+ "end": 18,
502
+ "entity_type": "PERSON",
503
+ "surface": "امید نمونه‌ای",
504
+ "language": "fa",
505
+ "field": null
506
+ },
507
+ {
508
+ "start": 27,
509
+ "end": 37,
510
+ "entity_type": "IR_NATIONAL_ID",
511
+ "surface": "۴۶۱۷۵۵۴۲۹۷",
512
+ "language": "fa",
513
+ "field": null
514
+ },
515
+ {
516
+ "start": 49,
517
+ "end": 60,
518
+ "entity_type": "PHONE",
519
+ "surface": "۰۹۱۲۸۲۹۱۰۰۲",
520
+ "language": "fa",
521
+ "field": null
522
+ },
523
+ {
524
+ "start": 67,
525
+ "end": 88,
526
+ "entity_type": "EMAIL",
527
+ "surface": "user.04@aria-lab.test",
528
+ "language": "en",
529
+ "field": null
530
+ }
531
+ ],
532
+ "meta": []
533
+ },
534
+ {
535
+ "doc_id": "DLP-FA-HR-05",
536
+ "spans": [
537
+ {
538
+ "start": 5,
539
+ "end": 18,
540
+ "entity_type": "PERSON",
541
+ "surface": "لیلا نمونه‌ای",
542
+ "language": "fa",
543
+ "field": null
544
+ },
545
+ {
546
+ "start": 27,
547
+ "end": 37,
548
+ "entity_type": "IR_NATIONAL_ID",
549
+ "surface": "۵۰۰۷۵۷۷۱۰۳",
550
+ "language": "fa",
551
+ "field": null
552
+ },
553
+ {
554
+ "start": 49,
555
+ "end": 60,
556
+ "entity_type": "PHONE",
557
+ "surface": "۰۹۱۲۲۳۶۷۲۱۴",
558
+ "language": "fa",
559
+ "field": null
560
+ },
561
+ {
562
+ "start": 67,
563
+ "end": 88,
564
+ "entity_type": "EMAIL",
565
+ "surface": "user.05@aria-lab.test",
566
+ "language": "en",
567
+ "field": null
568
+ }
569
+ ],
570
+ "meta": []
571
+ },
572
+ {
573
+ "doc_id": "DLP-FA-HR-06",
574
+ "spans": [
575
+ {
576
+ "start": 5,
577
+ "end": 17,
578
+ "entity_type": "PERSON",
579
+ "surface": "لیلا آزمایشی",
580
+ "language": "fa",
581
+ "field": null
582
+ },
583
+ {
584
+ "start": 26,
585
+ "end": 36,
586
+ "entity_type": "IR_NATIONAL_ID",
587
+ "surface": "۶۸۷۹۸۶۶۳۴۴",
588
+ "language": "fa",
589
+ "field": null
590
+ },
591
+ {
592
+ "start": 48,
593
+ "end": 59,
594
+ "entity_type": "PHONE",
595
+ "surface": "۰۹۱۲۶۰۸۳۲۸۳",
596
+ "language": "fa",
597
+ "field": null
598
+ },
599
+ {
600
+ "start": 66,
601
+ "end": 87,
602
+ "entity_type": "EMAIL",
603
+ "surface": "user.06@aria-lab.test",
604
+ "language": "en",
605
+ "field": null
606
+ }
607
+ ],
608
+ "meta": []
609
+ },
610
+ {
611
+ "doc_id": "DLP-FA-HR-07",
612
+ "spans": [
613
+ {
614
+ "start": 5,
615
+ "end": 17,
616
+ "entity_type": "PERSON",
617
+ "surface": "لیلا آزمایشی",
618
+ "language": "fa",
619
+ "field": null
620
+ },
621
+ {
622
+ "start": 26,
623
+ "end": 36,
624
+ "entity_type": "IR_NATIONAL_ID",
625
+ "surface": "۲۵۱۹۰۴۸۰۶۹",
626
+ "language": "fa",
627
+ "field": null
628
+ },
629
+ {
630
+ "start": 48,
631
+ "end": 59,
632
+ "entity_type": "PHONE",
633
+ "surface": "۰۹۱۲۷۷۱۴۶۰۱",
634
+ "language": "fa",
635
+ "field": null
636
+ },
637
+ {
638
+ "start": 66,
639
+ "end": 87,
640
+ "entity_type": "EMAIL",
641
+ "surface": "user.07@aria-lab.test",
642
+ "language": "en",
643
+ "field": null
644
+ }
645
+ ],
646
+ "meta": []
647
+ },
648
+ {
649
+ "doc_id": "DLP-FA-HR-08",
650
+ "spans": [
651
+ {
652
+ "start": 5,
653
+ "end": 17,
654
+ "entity_type": "PERSON",
655
+ "surface": "امید آزمایشی",
656
+ "language": "fa",
657
+ "field": null
658
+ },
659
+ {
660
+ "start": 26,
661
+ "end": 36,
662
+ "entity_type": "IR_NATIONAL_ID",
663
+ "surface": "۴۷۶۵۵۰۶۰۷۱",
664
+ "language": "fa",
665
+ "field": null
666
+ },
667
+ {
668
+ "start": 48,
669
+ "end": 59,
670
+ "entity_type": "PHONE",
671
+ "surface": "۰۹۱۲۸۷۶۰۹۰۱",
672
+ "language": "fa",
673
+ "field": null
674
+ },
675
+ {
676
+ "start": 66,
677
+ "end": 87,
678
+ "entity_type": "EMAIL",
679
+ "surface": "user.08@aria-lab.test",
680
+ "language": "en",
681
+ "field": null
682
+ }
683
+ ],
684
+ "meta": []
685
+ },
686
+ {
687
+ "doc_id": "DLP-FA-HR-09",
688
+ "spans": [
689
+ {
690
+ "start": 5,
691
+ "end": 18,
692
+ "entity_type": "PERSON",
693
+ "surface": "سارا نمونه‌ای",
694
+ "language": "fa",
695
+ "field": null
696
+ },
697
+ {
698
+ "start": 27,
699
+ "end": 37,
700
+ "entity_type": "IR_NATIONAL_ID",
701
+ "surface": "۳۲۸۹۳۷۲۲۰۰",
702
+ "language": "fa",
703
+ "field": null
704
+ },
705
+ {
706
+ "start": 49,
707
+ "end": 60,
708
+ "entity_type": "PHONE",
709
+ "surface": "۰۹۱۲۲۸۹۳۲۸۹",
710
+ "language": "fa",
711
+ "field": null
712
+ },
713
+ {
714
+ "start": 67,
715
+ "end": 88,
716
+ "entity_type": "EMAIL",
717
+ "surface": "user.09@aria-lab.test",
718
+ "language": "en",
719
+ "field": null
720
+ }
721
+ ],
722
+ "meta": []
723
+ },
724
+ {
725
+ "doc_id": "DLP-FA-HR-10",
726
+ "spans": [
727
+ {
728
+ "start": 5,
729
+ "end": 18,
730
+ "entity_type": "PERSON",
731
+ "surface": "سارا نمونه‌ای",
732
+ "language": "fa",
733
+ "field": null
734
+ },
735
+ {
736
+ "start": 27,
737
+ "end": 37,
738
+ "entity_type": "IR_NATIONAL_ID",
739
+ "surface": "۳۰۸۷۰۲۵۴۱۰",
740
+ "language": "fa",
741
+ "field": null
742
+ },
743
+ {
744
+ "start": 49,
745
+ "end": 60,
746
+ "entity_type": "PHONE",
747
+ "surface": "۰۹۱۲۴۰۶۷۵۴۸",
748
+ "language": "fa",
749
+ "field": null
750
+ },
751
+ {
752
+ "start": 67,
753
+ "end": 88,
754
+ "entity_type": "EMAIL",
755
+ "surface": "user.10@aria-lab.test",
756
+ "language": "en",
757
+ "field": null
758
+ }
759
+ ],
760
+ "meta": []
761
+ },
762
+ {
763
+ "doc_id": "DLP-CSV-INV-01",
764
+ "spans": [
765
+ {
766
+ "start": 53,
767
+ "end": 63,
768
+ "entity_type": "PERSON",
769
+ "surface": "Ava Labian",
770
+ "language": "en",
771
+ "field": null
772
+ },
773
+ {
774
+ "start": 64,
775
+ "end": 90,
776
+ "entity_type": "EMAIL",
777
+ "surface": "Ava.Labian01@aria-lab.test",
778
+ "language": "en",
779
+ "field": null
780
+ },
781
+ {
782
+ "start": 91,
783
+ "end": 102,
784
+ "entity_type": "PHONE",
785
+ "surface": "09129456762",
786
+ "language": "en",
787
+ "field": null
788
+ }
789
+ ],
790
+ "meta": []
791
+ },
792
+ {
793
+ "doc_id": "DLP-CSV-INV-02",
794
+ "spans": [
795
+ {
796
+ "start": 53,
797
+ "end": 65,
798
+ "entity_type": "PERSON",
799
+ "surface": "Nima Fixture",
800
+ "language": "en",
801
+ "field": null
802
+ },
803
+ {
804
+ "start": 66,
805
+ "end": 94,
806
+ "entity_type": "EMAIL",
807
+ "surface": "Nima.Fixture02@aria-lab.test",
808
+ "language": "en",
809
+ "field": null
810
+ },
811
+ {
812
+ "start": 95,
813
+ "end": 106,
814
+ "entity_type": "PHONE",
815
+ "surface": "09127891210",
816
+ "language": "en",
817
+ "field": null
818
+ }
819
+ ],
820
+ "meta": []
821
+ },
822
+ {
823
+ "doc_id": "DLP-CSV-INV-03",
824
+ "spans": [
825
+ {
826
+ "start": 53,
827
+ "end": 66,
828
+ "entity_type": "PERSON",
829
+ "surface": "Leila Fixture",
830
+ "language": "en",
831
+ "field": null
832
+ },
833
+ {
834
+ "start": 67,
835
+ "end": 96,
836
+ "entity_type": "EMAIL",
837
+ "surface": "Leila.Fixture03@aria-lab.test",
838
+ "language": "en",
839
+ "field": null
840
+ },
841
+ {
842
+ "start": 97,
843
+ "end": 108,
844
+ "entity_type": "PHONE",
845
+ "surface": "09124895575",
846
+ "language": "en",
847
+ "field": null
848
+ }
849
+ ],
850
+ "meta": []
851
+ },
852
+ {
853
+ "doc_id": "DLP-CSV-INV-04",
854
+ "spans": [
855
+ {
856
+ "start": 53,
857
+ "end": 64,
858
+ "entity_type": "PERSON",
859
+ "surface": "Omar Labian",
860
+ "language": "en",
861
+ "field": null
862
+ },
863
+ {
864
+ "start": 65,
865
+ "end": 92,
866
+ "entity_type": "EMAIL",
867
+ "surface": "Omar.Labian04@aria-lab.test",
868
+ "language": "en",
869
+ "field": null
870
+ },
871
+ {
872
+ "start": 93,
873
+ "end": 104,
874
+ "entity_type": "PHONE",
875
+ "surface": "09124151655",
876
+ "language": "en",
877
+ "field": null
878
+ }
879
+ ],
880
+ "meta": []
881
+ },
882
+ {
883
+ "doc_id": "DLP-CSV-INV-05",
884
+ "spans": [
885
+ {
886
+ "start": 53,
887
+ "end": 63,
888
+ "entity_type": "PERSON",
889
+ "surface": "Ava Labian",
890
+ "language": "en",
891
+ "field": null
892
+ },
893
+ {
894
+ "start": 64,
895
+ "end": 90,
896
+ "entity_type": "EMAIL",
897
+ "surface": "Ava.Labian05@aria-lab.test",
898
+ "language": "en",
899
+ "field": null
900
+ },
901
+ {
902
+ "start": 91,
903
+ "end": 102,
904
+ "entity_type": "PHONE",
905
+ "surface": "09123831675",
906
+ "language": "en",
907
+ "field": null
908
+ }
909
+ ],
910
+ "meta": []
911
+ },
912
+ {
913
+ "doc_id": "DLP-CSV-INV-06",
914
+ "spans": [
915
+ {
916
+ "start": 53,
917
+ "end": 65,
918
+ "entity_type": "PERSON",
919
+ "surface": "Noah Fixture",
920
+ "language": "en",
921
+ "field": null
922
+ },
923
+ {
924
+ "start": 66,
925
+ "end": 94,
926
+ "entity_type": "EMAIL",
927
+ "surface": "Noah.Fixture06@aria-lab.test",
928
+ "language": "en",
929
+ "field": null
930
+ },
931
+ {
932
+ "start": 95,
933
+ "end": 106,
934
+ "entity_type": "PHONE",
935
+ "surface": "09125428133",
936
+ "language": "en",
937
+ "field": null
938
+ }
939
+ ],
940
+ "meta": []
941
+ },
942
+ {
943
+ "doc_id": "DLP-CSV-INV-07",
944
+ "spans": [
945
+ {
946
+ "start": 53,
947
+ "end": 66,
948
+ "entity_type": "PERSON",
949
+ "surface": "Helen Fixture",
950
+ "language": "en",
951
+ "field": null
952
+ },
953
+ {
954
+ "start": 67,
955
+ "end": 96,
956
+ "entity_type": "EMAIL",
957
+ "surface": "Helen.Fixture07@aria-lab.test",
958
+ "language": "en",
959
+ "field": null
960
+ },
961
+ {
962
+ "start": 97,
963
+ "end": 108,
964
+ "entity_type": "PHONE",
965
+ "surface": "09124672166",
966
+ "language": "en",
967
+ "field": null
968
+ }
969
+ ],
970
+ "meta": []
971
+ },
972
+ {
973
+ "doc_id": "DLP-CSV-INV-08",
974
+ "spans": [
975
+ {
976
+ "start": 53,
977
+ "end": 65,
978
+ "entity_type": "PERSON",
979
+ "surface": "Omar Fixture",
980
+ "language": "en",
981
+ "field": null
982
+ },
983
+ {
984
+ "start": 66,
985
+ "end": 94,
986
+ "entity_type": "EMAIL",
987
+ "surface": "Omar.Fixture08@aria-lab.test",
988
+ "language": "en",
989
+ "field": null
990
+ },
991
+ {
992
+ "start": 95,
993
+ "end": 106,
994
+ "entity_type": "PHONE",
995
+ "surface": "09128527564",
996
+ "language": "en",
997
+ "field": null
998
+ }
999
+ ],
1000
+ "meta": []
1001
+ },
1002
+ {
1003
+ "doc_id": "DLP-CSV-INV-09",
1004
+ "spans": [
1005
+ {
1006
+ "start": 53,
1007
+ "end": 65,
1008
+ "entity_type": "PERSON",
1009
+ "surface": "Leila Labian",
1010
+ "language": "en",
1011
+ "field": null
1012
+ },
1013
+ {
1014
+ "start": 66,
1015
+ "end": 94,
1016
+ "entity_type": "EMAIL",
1017
+ "surface": "Leila.Labian09@aria-lab.test",
1018
+ "language": "en",
1019
+ "field": null
1020
+ },
1021
+ {
1022
+ "start": 95,
1023
+ "end": 106,
1024
+ "entity_type": "PHONE",
1025
+ "surface": "09127031590",
1026
+ "language": "en",
1027
+ "field": null
1028
+ }
1029
+ ],
1030
+ "meta": []
1031
+ },
1032
+ {
1033
+ "doc_id": "DLP-CSV-INV-10",
1034
+ "spans": [
1035
+ {
1036
+ "start": 53,
1037
+ "end": 66,
1038
+ "entity_type": "PERSON",
1039
+ "surface": "Leila Fixture",
1040
+ "language": "en",
1041
+ "field": null
1042
+ },
1043
+ {
1044
+ "start": 67,
1045
+ "end": 96,
1046
+ "entity_type": "EMAIL",
1047
+ "surface": "Leila.Fixture10@aria-lab.test",
1048
+ "language": "en",
1049
+ "field": null
1050
+ },
1051
+ {
1052
+ "start": 97,
1053
+ "end": 108,
1054
+ "entity_type": "PHONE",
1055
+ "surface": "09123631859",
1056
+ "language": "en",
1057
+ "field": null
1058
+ }
1059
+ ],
1060
+ "meta": []
1061
+ },
1062
+ {
1063
+ "doc_id": "DLP-EN-ONB-01",
1064
+ "spans": [
1065
+ {
1066
+ "start": 6,
1067
+ "end": 18,
1068
+ "entity_type": "PERSON",
1069
+ "surface": "Helen Labian",
1070
+ "language": "en",
1071
+ "field": null
1072
+ },
1073
+ {
1074
+ "start": 31,
1075
+ "end": 59,
1076
+ "entity_type": "EMAIL",
1077
+ "surface": "Helen.Labian01@aria-lab.test",
1078
+ "language": "en",
1079
+ "field": null
1080
+ },
1081
+ {
1082
+ "start": 85,
1083
+ "end": 101,
1084
+ "entity_type": "CREDIT_CARD",
1085
+ "surface": "4111137038772834",
1086
+ "language": "en",
1087
+ "field": null
1088
+ },
1089
+ {
1090
+ "start": 116,
1091
+ "end": 142,
1092
+ "entity_type": "IBAN",
1093
+ "surface": "IR308660110372211081968590",
1094
+ "language": "en",
1095
+ "field": null
1096
+ }
1097
+ ],
1098
+ "meta": []
1099
+ },
1100
+ {
1101
+ "doc_id": "DLP-EN-ONB-02",
1102
+ "spans": [
1103
+ {
1104
+ "start": 6,
1105
+ "end": 18,
1106
+ "entity_type": "PERSON",
1107
+ "surface": "Helen Labian",
1108
+ "language": "en",
1109
+ "field": null
1110
+ },
1111
+ {
1112
+ "start": 31,
1113
+ "end": 59,
1114
+ "entity_type": "EMAIL",
1115
+ "surface": "Helen.Labian02@aria-lab.test",
1116
+ "language": "en",
1117
+ "field": null
1118
+ },
1119
+ {
1120
+ "start": 85,
1121
+ "end": 101,
1122
+ "entity_type": "CREDIT_CARD",
1123
+ "surface": "4111529393529637",
1124
+ "language": "en",
1125
+ "field": null
1126
+ },
1127
+ {
1128
+ "start": 116,
1129
+ "end": 142,
1130
+ "entity_type": "IBAN",
1131
+ "surface": "IR044280439609406102122129",
1132
+ "language": "en",
1133
+ "field": null
1134
+ }
1135
+ ],
1136
+ "meta": []
1137
+ },
1138
+ {
1139
+ "doc_id": "DLP-EN-ONB-03",
1140
+ "spans": [
1141
+ {
1142
+ "start": 6,
1143
+ "end": 18,
1144
+ "entity_type": "PERSON",
1145
+ "surface": "Helen Labian",
1146
+ "language": "en",
1147
+ "field": null
1148
+ },
1149
+ {
1150
+ "start": 31,
1151
+ "end": 59,
1152
+ "entity_type": "EMAIL",
1153
+ "surface": "Helen.Labian03@aria-lab.test",
1154
+ "language": "en",
1155
+ "field": null
1156
+ },
1157
+ {
1158
+ "start": 85,
1159
+ "end": 101,
1160
+ "entity_type": "CREDIT_CARD",
1161
+ "surface": "4111785437328044",
1162
+ "language": "en",
1163
+ "field": null
1164
+ },
1165
+ {
1166
+ "start": 116,
1167
+ "end": 142,
1168
+ "entity_type": "IBAN",
1169
+ "surface": "IR019630083854359765553642",
1170
+ "language": "en",
1171
+ "field": null
1172
+ }
1173
+ ],
1174
+ "meta": []
1175
+ },
1176
+ {
1177
+ "doc_id": "DLP-EN-ONB-04",
1178
+ "spans": [
1179
+ {
1180
+ "start": 6,
1181
+ "end": 18,
1182
+ "entity_type": "PERSON",
1183
+ "surface": "Helen Labian",
1184
+ "language": "en",
1185
+ "field": null
1186
+ },
1187
+ {
1188
+ "start": 31,
1189
+ "end": 59,
1190
+ "entity_type": "EMAIL",
1191
+ "surface": "Helen.Labian04@aria-lab.test",
1192
+ "language": "en",
1193
+ "field": null
1194
+ },
1195
+ {
1196
+ "start": 85,
1197
+ "end": 101,
1198
+ "entity_type": "CREDIT_CARD",
1199
+ "surface": "4111759320597908",
1200
+ "language": "en",
1201
+ "field": null
1202
+ },
1203
+ {
1204
+ "start": 116,
1205
+ "end": 142,
1206
+ "entity_type": "IBAN",
1207
+ "surface": "IR472630756281276293569333",
1208
+ "language": "en",
1209
+ "field": null
1210
+ }
1211
+ ],
1212
+ "meta": []
1213
+ },
1214
+ {
1215
+ "doc_id": "DLP-EN-ONB-05",
1216
+ "spans": [
1217
+ {
1218
+ "start": 6,
1219
+ "end": 19,
1220
+ "entity_type": "PERSON",
1221
+ "surface": "Leila Fixture",
1222
+ "language": "en",
1223
+ "field": null
1224
+ },
1225
+ {
1226
+ "start": 32,
1227
+ "end": 61,
1228
+ "entity_type": "EMAIL",
1229
+ "surface": "Leila.Fixture05@aria-lab.test",
1230
+ "language": "en",
1231
+ "field": null
1232
+ },
1233
+ {
1234
+ "start": 87,
1235
+ "end": 103,
1236
+ "entity_type": "CREDIT_CARD",
1237
+ "surface": "4111313694373346",
1238
+ "language": "en",
1239
+ "field": null
1240
+ },
1241
+ {
1242
+ "start": 118,
1243
+ "end": 144,
1244
+ "entity_type": "IBAN",
1245
+ "surface": "IR466530650673099236853917",
1246
+ "language": "en",
1247
+ "field": null
1248
+ }
1249
+ ],
1250
+ "meta": []
1251
+ },
1252
+ {
1253
+ "doc_id": "DLP-EN-ONB-06",
1254
+ "spans": [
1255
+ {
1256
+ "start": 6,
1257
+ "end": 17,
1258
+ "entity_type": "PERSON",
1259
+ "surface": "Nima Labian",
1260
+ "language": "en",
1261
+ "field": null
1262
+ },
1263
+ {
1264
+ "start": 30,
1265
+ "end": 57,
1266
+ "entity_type": "EMAIL",
1267
+ "surface": "Nima.Labian06@aria-lab.test",
1268
+ "language": "en",
1269
+ "field": null
1270
+ },
1271
+ {
1272
+ "start": 83,
1273
+ "end": 99,
1274
+ "entity_type": "CREDIT_CARD",
1275
+ "surface": "4111814102596287",
1276
+ "language": "en",
1277
+ "field": null
1278
+ },
1279
+ {
1280
+ "start": 114,
1281
+ "end": 140,
1282
+ "entity_type": "IBAN",
1283
+ "surface": "IR581060722681444648227055",
1284
+ "language": "en",
1285
+ "field": null
1286
+ }
1287
+ ],
1288
+ "meta": []
1289
+ },
1290
+ {
1291
+ "doc_id": "DLP-EN-ONB-07",
1292
+ "spans": [
1293
+ {
1294
+ "start": 6,
1295
+ "end": 18,
1296
+ "entity_type": "PERSON",
1297
+ "surface": "Noah Fixture",
1298
+ "language": "en",
1299
+ "field": null
1300
+ },
1301
+ {
1302
+ "start": 31,
1303
+ "end": 59,
1304
+ "entity_type": "EMAIL",
1305
+ "surface": "Noah.Fixture07@aria-lab.test",
1306
+ "language": "en",
1307
+ "field": null
1308
+ },
1309
+ {
1310
+ "start": 85,
1311
+ "end": 101,
1312
+ "entity_type": "CREDIT_CARD",
1313
+ "surface": "4111842413164297",
1314
+ "language": "en",
1315
+ "field": null
1316
+ },
1317
+ {
1318
+ "start": 116,
1319
+ "end": 142,
1320
+ "entity_type": "IBAN",
1321
+ "surface": "IR895120978818064078339048",
1322
+ "language": "en",
1323
+ "field": null
1324
+ }
1325
+ ],
1326
+ "meta": []
1327
+ },
1328
+ {
1329
+ "doc_id": "DLP-EN-ONB-08",
1330
+ "spans": [
1331
+ {
1332
+ "start": 6,
1333
+ "end": 18,
1334
+ "entity_type": "PERSON",
1335
+ "surface": "Noah Fixture",
1336
+ "language": "en",
1337
+ "field": null
1338
+ },
1339
+ {
1340
+ "start": 31,
1341
+ "end": 59,
1342
+ "entity_type": "EMAIL",
1343
+ "surface": "Noah.Fixture08@aria-lab.test",
1344
+ "language": "en",
1345
+ "field": null
1346
+ },
1347
+ {
1348
+ "start": 85,
1349
+ "end": 101,
1350
+ "entity_type": "CREDIT_CARD",
1351
+ "surface": "4111329710194664",
1352
+ "language": "en",
1353
+ "field": null
1354
+ },
1355
+ {
1356
+ "start": 116,
1357
+ "end": 142,
1358
+ "entity_type": "IBAN",
1359
+ "surface": "IR237680114663394226281259",
1360
+ "language": "en",
1361
+ "field": null
1362
+ }
1363
+ ],
1364
+ "meta": []
1365
+ },
1366
+ {
1367
+ "doc_id": "DLP-EN-ONB-09",
1368
+ "spans": [
1369
+ {
1370
+ "start": 6,
1371
+ "end": 19,
1372
+ "entity_type": "PERSON",
1373
+ "surface": "Leila Fixture",
1374
+ "language": "en",
1375
+ "field": null
1376
+ },
1377
+ {
1378
+ "start": 32,
1379
+ "end": 61,
1380
+ "entity_type": "EMAIL",
1381
+ "surface": "Leila.Fixture09@aria-lab.test",
1382
+ "language": "en",
1383
+ "field": null
1384
+ },
1385
+ {
1386
+ "start": 87,
1387
+ "end": 103,
1388
+ "entity_type": "CREDIT_CARD",
1389
+ "surface": "4111996662241709",
1390
+ "language": "en",
1391
+ "field": null
1392
+ },
1393
+ {
1394
+ "start": 118,
1395
+ "end": 144,
1396
+ "entity_type": "IBAN",
1397
+ "surface": "IR772810906854799104860878",
1398
+ "language": "en",
1399
+ "field": null
1400
+ }
1401
+ ],
1402
+ "meta": []
1403
+ },
1404
+ {
1405
+ "doc_id": "DLP-EN-ONB-10",
1406
+ "spans": [
1407
+ {
1408
+ "start": 6,
1409
+ "end": 18,
1410
+ "entity_type": "PERSON",
1411
+ "surface": "Nima Fixture",
1412
+ "language": "en",
1413
+ "field": null
1414
+ },
1415
+ {
1416
+ "start": 31,
1417
+ "end": 59,
1418
+ "entity_type": "EMAIL",
1419
+ "surface": "Nima.Fixture10@aria-lab.test",
1420
+ "language": "en",
1421
+ "field": null
1422
+ },
1423
+ {
1424
+ "start": 85,
1425
+ "end": 101,
1426
+ "entity_type": "CREDIT_CARD",
1427
+ "surface": "4111041294625191",
1428
+ "language": "en",
1429
+ "field": null
1430
+ },
1431
+ {
1432
+ "start": 116,
1433
+ "end": 142,
1434
+ "entity_type": "IBAN",
1435
+ "surface": "IR338550745309562340114615",
1436
+ "language": "en",
1437
+ "field": null
1438
+ }
1439
+ ],
1440
+ "meta": []
1441
+ },
1442
+ {
1443
+ "doc_id": "DLP-FA-CTR-01",
1444
+ "spans": [
1445
+ {
1446
+ "start": 27,
1447
+ "end": 39,
1448
+ "entity_type": "PERSON",
1449
+ "surface": "لیلا آزمایشی",
1450
+ "language": "fa",
1451
+ "field": null
1452
+ },
1453
+ {
1454
+ "start": 47,
1455
+ "end": 57,
1456
+ "entity_type": "IR_NATIONAL_ID",
1457
+ "surface": "4985528323",
1458
+ "language": "fa",
1459
+ "field": null
1460
+ },
1461
+ {
1462
+ "start": 63,
1463
+ "end": 74,
1464
+ "entity_type": "PHONE",
1465
+ "surface": "۰۹۱۲۳۴۲۱۵۱۴",
1466
+ "language": "fa",
1467
+ "field": null
1468
+ }
1469
+ ],
1470
+ "meta": []
1471
+ },
1472
+ {
1473
+ "doc_id": "DLP-FA-CTR-02",
1474
+ "spans": [
1475
+ {
1476
+ "start": 27,
1477
+ "end": 39,
1478
+ "entity_type": "PERSON",
1479
+ "surface": "سارا آزمایشی",
1480
+ "language": "fa",
1481
+ "field": null
1482
+ },
1483
+ {
1484
+ "start": 47,
1485
+ "end": 57,
1486
+ "entity_type": "IR_NATIONAL_ID",
1487
+ "surface": "6361423751",
1488
+ "language": "fa",
1489
+ "field": null
1490
+ },
1491
+ {
1492
+ "start": 63,
1493
+ "end": 74,
1494
+ "entity_type": "PHONE",
1495
+ "surface": "۰۹۱۲۳۶۳۰۹۶۹",
1496
+ "language": "fa",
1497
+ "field": null
1498
+ }
1499
+ ],
1500
+ "meta": []
1501
+ },
1502
+ {
1503
+ "doc_id": "DLP-FA-CTR-03",
1504
+ "spans": [
1505
+ {
1506
+ "start": 27,
1507
+ "end": 40,
1508
+ "entity_type": "PERSON",
1509
+ "surface": "نیما نمونه‌ای",
1510
+ "language": "fa",
1511
+ "field": null
1512
+ },
1513
+ {
1514
+ "start": 48,
1515
+ "end": 58,
1516
+ "entity_type": "IR_NATIONAL_ID",
1517
+ "surface": "3589737379",
1518
+ "language": "fa",
1519
+ "field": null
1520
+ },
1521
+ {
1522
+ "start": 64,
1523
+ "end": 75,
1524
+ "entity_type": "PHONE",
1525
+ "surface": "۰۹۱۲۵۹۲۸۹۵۰",
1526
+ "language": "fa",
1527
+ "field": null
1528
+ }
1529
+ ],
1530
+ "meta": []
1531
+ },
1532
+ {
1533
+ "doc_id": "DLP-FA-CTR-04",
1534
+ "spans": [
1535
+ {
1536
+ "start": 27,
1537
+ "end": 39,
1538
+ "entity_type": "PERSON",
1539
+ "surface": "نیما آزمایشی",
1540
+ "language": "fa",
1541
+ "field": null
1542
+ },
1543
+ {
1544
+ "start": 47,
1545
+ "end": 57,
1546
+ "entity_type": "IR_NATIONAL_ID",
1547
+ "surface": "2341457452",
1548
+ "language": "fa",
1549
+ "field": null
1550
+ },
1551
+ {
1552
+ "start": 63,
1553
+ "end": 74,
1554
+ "entity_type": "PHONE",
1555
+ "surface": "۰۹۱۲۲۳۶۵۹۰۷",
1556
+ "language": "fa",
1557
+ "field": null
1558
+ }
1559
+ ],
1560
+ "meta": []
1561
+ },
1562
+ {
1563
+ "doc_id": "DLP-FA-CTR-05",
1564
+ "spans": [
1565
+ {
1566
+ "start": 27,
1567
+ "end": 40,
1568
+ "entity_type": "PERSON",
1569
+ "surface": "نیما نمونه‌ای",
1570
+ "language": "fa",
1571
+ "field": null
1572
+ },
1573
+ {
1574
+ "start": 48,
1575
+ "end": 58,
1576
+ "entity_type": "IR_NATIONAL_ID",
1577
+ "surface": "4061666126",
1578
+ "language": "fa",
1579
+ "field": null
1580
+ },
1581
+ {
1582
+ "start": 64,
1583
+ "end": 75,
1584
+ "entity_type": "PHONE",
1585
+ "surface": "۰۹۱۲۴۹۶۶۴۱۱",
1586
+ "language": "fa",
1587
+ "field": null
1588
+ }
1589
+ ],
1590
+ "meta": []
1591
+ },
1592
+ {
1593
+ "doc_id": "DLP-FA-CTR-06",
1594
+ "spans": [
1595
+ {
1596
+ "start": 27,
1597
+ "end": 40,
1598
+ "entity_type": "PERSON",
1599
+ "surface": "لیلا نمونه‌ای",
1600
+ "language": "fa",
1601
+ "field": null
1602
+ },
1603
+ {
1604
+ "start": 48,
1605
+ "end": 58,
1606
+ "entity_type": "IR_NATIONAL_ID",
1607
+ "surface": "3179136243",
1608
+ "language": "fa",
1609
+ "field": null
1610
+ },
1611
+ {
1612
+ "start": 64,
1613
+ "end": 75,
1614
+ "entity_type": "PHONE",
1615
+ "surface": "۰۹۱۲۸۱۳۵۴۲۶",
1616
+ "language": "fa",
1617
+ "field": null
1618
+ }
1619
+ ],
1620
+ "meta": []
1621
+ },
1622
+ {
1623
+ "doc_id": "DLP-FA-CTR-07",
1624
+ "spans": [
1625
+ {
1626
+ "start": 27,
1627
+ "end": 39,
1628
+ "entity_type": "PERSON",
1629
+ "surface": "امید آزمایشی",
1630
+ "language": "fa",
1631
+ "field": null
1632
+ },
1633
+ {
1634
+ "start": 47,
1635
+ "end": 57,
1636
+ "entity_type": "IR_NATIONAL_ID",
1637
+ "surface": "2154052452",
1638
+ "language": "fa",
1639
+ "field": null
1640
+ },
1641
+ {
1642
+ "start": 63,
1643
+ "end": 74,
1644
+ "entity_type": "PHONE",
1645
+ "surface": "۰۹۱۲۸۸۱۷۱۸۵",
1646
+ "language": "fa",
1647
+ "field": null
1648
+ }
1649
+ ],
1650
+ "meta": []
1651
+ },
1652
+ {
1653
+ "doc_id": "DLP-FA-CTR-08",
1654
+ "spans": [
1655
+ {
1656
+ "start": 27,
1657
+ "end": 39,
1658
+ "entity_type": "PERSON",
1659
+ "surface": "سارا آزمایشی",
1660
+ "language": "fa",
1661
+ "field": null
1662
+ },
1663
+ {
1664
+ "start": 47,
1665
+ "end": 57,
1666
+ "entity_type": "IR_NATIONAL_ID",
1667
+ "surface": "2313920291",
1668
+ "language": "fa",
1669
+ "field": null
1670
+ },
1671
+ {
1672
+ "start": 63,
1673
+ "end": 74,
1674
+ "entity_type": "PHONE",
1675
+ "surface": "۰۹۱۲۱۳۴۴۶۷۵",
1676
+ "language": "fa",
1677
+ "field": null
1678
+ }
1679
+ ],
1680
+ "meta": []
1681
+ },
1682
+ {
1683
+ "doc_id": "DLP-FA-CTR-09",
1684
+ "spans": [
1685
+ {
1686
+ "start": 27,
1687
+ "end": 40,
1688
+ "entity_type": "PERSON",
1689
+ "surface": "نیما نمونه‌ای",
1690
+ "language": "fa",
1691
+ "field": null
1692
+ },
1693
+ {
1694
+ "start": 48,
1695
+ "end": 58,
1696
+ "entity_type": "IR_NATIONAL_ID",
1697
+ "surface": "0190065516",
1698
+ "language": "fa",
1699
+ "field": null
1700
+ },
1701
+ {
1702
+ "start": 64,
1703
+ "end": 75,
1704
+ "entity_type": "PHONE",
1705
+ "surface": "۰۹۱۲۷۲۰۹۰۸۸",
1706
+ "language": "fa",
1707
+ "field": null
1708
+ }
1709
+ ],
1710
+ "meta": []
1711
+ },
1712
+ {
1713
+ "doc_id": "DLP-FA-CTR-10",
1714
+ "spans": [
1715
+ {
1716
+ "start": 27,
1717
+ "end": 39,
1718
+ "entity_type": "PERSON",
1719
+ "surface": "لیلا آزمایشی",
1720
+ "language": "fa",
1721
+ "field": null
1722
+ },
1723
+ {
1724
+ "start": 47,
1725
+ "end": 57,
1726
+ "entity_type": "IR_NATIONAL_ID",
1727
+ "surface": "8212633349",
1728
+ "language": "fa",
1729
+ "field": null
1730
+ },
1731
+ {
1732
+ "start": 63,
1733
+ "end": 74,
1734
+ "entity_type": "PHONE",
1735
+ "surface": "۰۹۱۲۷۴۸۴۳۶۷",
1736
+ "language": "fa",
1737
+ "field": null
1738
+ }
1739
+ ],
1740
+ "meta": []
1741
+ },
1742
+ {
1743
+ "doc_id": "DLP-PDF-MEMO-01",
1744
+ "spans": [
1745
+ {
1746
+ "start": 22,
1747
+ "end": 34,
1748
+ "entity_type": "PERSON",
1749
+ "surface": "Omar Fixture",
1750
+ "language": "en",
1751
+ "field": null
1752
+ },
1753
+ {
1754
+ "start": 44,
1755
+ "end": 72,
1756
+ "entity_type": "EMAIL",
1757
+ "surface": "Omar.Fixture01@aria-lab.test",
1758
+ "language": "en",
1759
+ "field": null
1760
+ },
1761
+ {
1762
+ "start": 75,
1763
+ "end": 86,
1764
+ "entity_type": "PHONE",
1765
+ "surface": "09127692383",
1766
+ "language": "en",
1767
+ "field": null
1768
+ }
1769
+ ],
1770
+ "meta": [
1771
+ {
1772
+ "field": "author",
1773
+ "entity_type": "PERSON",
1774
+ "surface": "Omar Fixture"
1775
+ },
1776
+ {
1777
+ "field": "keywords",
1778
+ "entity_type": "EMAIL",
1779
+ "surface": "Omar.Fixture01@aria-lab.test"
1780
+ }
1781
+ ]
1782
+ },
1783
+ {
1784
+ "doc_id": "DLP-PDF-MEMO-02",
1785
+ "spans": [
1786
+ {
1787
+ "start": 22,
1788
+ "end": 34,
1789
+ "entity_type": "PERSON",
1790
+ "surface": "Noah Fixture",
1791
+ "language": "en",
1792
+ "field": null
1793
+ },
1794
+ {
1795
+ "start": 44,
1796
+ "end": 72,
1797
+ "entity_type": "EMAIL",
1798
+ "surface": "Noah.Fixture02@aria-lab.test",
1799
+ "language": "en",
1800
+ "field": null
1801
+ },
1802
+ {
1803
+ "start": 75,
1804
+ "end": 86,
1805
+ "entity_type": "PHONE",
1806
+ "surface": "09125328205",
1807
+ "language": "en",
1808
+ "field": null
1809
+ }
1810
+ ],
1811
+ "meta": [
1812
+ {
1813
+ "field": "author",
1814
+ "entity_type": "PERSON",
1815
+ "surface": "Noah Fixture"
1816
+ },
1817
+ {
1818
+ "field": "keywords",
1819
+ "entity_type": "EMAIL",
1820
+ "surface": "Noah.Fixture02@aria-lab.test"
1821
+ }
1822
+ ]
1823
+ },
1824
+ {
1825
+ "doc_id": "DLP-PDF-MEMO-03",
1826
+ "spans": [
1827
+ {
1828
+ "start": 22,
1829
+ "end": 33,
1830
+ "entity_type": "PERSON",
1831
+ "surface": "Omar Labian",
1832
+ "language": "en",
1833
+ "field": null
1834
+ },
1835
+ {
1836
+ "start": 43,
1837
+ "end": 70,
1838
+ "entity_type": "EMAIL",
1839
+ "surface": "Omar.Labian03@aria-lab.test",
1840
+ "language": "en",
1841
+ "field": null
1842
+ },
1843
+ {
1844
+ "start": 73,
1845
+ "end": 84,
1846
+ "entity_type": "PHONE",
1847
+ "surface": "09125603855",
1848
+ "language": "en",
1849
+ "field": null
1850
+ }
1851
+ ],
1852
+ "meta": [
1853
+ {
1854
+ "field": "author",
1855
+ "entity_type": "PERSON",
1856
+ "surface": "Omar Labian"
1857
+ },
1858
+ {
1859
+ "field": "keywords",
1860
+ "entity_type": "EMAIL",
1861
+ "surface": "Omar.Labian03@aria-lab.test"
1862
+ }
1863
+ ]
1864
+ },
1865
+ {
1866
+ "doc_id": "DLP-PDF-MEMO-04",
1867
+ "spans": [
1868
+ {
1869
+ "start": 22,
1870
+ "end": 33,
1871
+ "entity_type": "PERSON",
1872
+ "surface": "Noah Labian",
1873
+ "language": "en",
1874
+ "field": null
1875
+ },
1876
+ {
1877
+ "start": 43,
1878
+ "end": 70,
1879
+ "entity_type": "EMAIL",
1880
+ "surface": "Noah.Labian04@aria-lab.test",
1881
+ "language": "en",
1882
+ "field": null
1883
+ },
1884
+ {
1885
+ "start": 73,
1886
+ "end": 84,
1887
+ "entity_type": "PHONE",
1888
+ "surface": "09127591902",
1889
+ "language": "en",
1890
+ "field": null
1891
+ }
1892
+ ],
1893
+ "meta": [
1894
+ {
1895
+ "field": "author",
1896
+ "entity_type": "PERSON",
1897
+ "surface": "Noah Labian"
1898
+ },
1899
+ {
1900
+ "field": "keywords",
1901
+ "entity_type": "EMAIL",
1902
+ "surface": "Noah.Labian04@aria-lab.test"
1903
+ }
1904
+ ]
1905
+ },
1906
+ {
1907
+ "doc_id": "DLP-PDF-MEMO-05",
1908
+ "spans": [
1909
+ {
1910
+ "start": 22,
1911
+ "end": 33,
1912
+ "entity_type": "PERSON",
1913
+ "surface": "Nima Labian",
1914
+ "language": "en",
1915
+ "field": null
1916
+ },
1917
+ {
1918
+ "start": 43,
1919
+ "end": 70,
1920
+ "entity_type": "EMAIL",
1921
+ "surface": "Nima.Labian05@aria-lab.test",
1922
+ "language": "en",
1923
+ "field": null
1924
+ },
1925
+ {
1926
+ "start": 73,
1927
+ "end": 84,
1928
+ "entity_type": "PHONE",
1929
+ "surface": "09129473675",
1930
+ "language": "en",
1931
+ "field": null
1932
+ }
1933
+ ],
1934
+ "meta": [
1935
+ {
1936
+ "field": "author",
1937
+ "entity_type": "PERSON",
1938
+ "surface": "Nima Labian"
1939
+ },
1940
+ {
1941
+ "field": "keywords",
1942
+ "entity_type": "EMAIL",
1943
+ "surface": "Nima.Labian05@aria-lab.test"
1944
+ }
1945
+ ]
1946
+ },
1947
+ {
1948
+ "doc_id": "DLP-PDF-MEMO-06",
1949
+ "spans": [
1950
+ {
1951
+ "start": 22,
1952
+ "end": 35,
1953
+ "entity_type": "PERSON",
1954
+ "surface": "Helen Fixture",
1955
+ "language": "en",
1956
+ "field": null
1957
+ },
1958
+ {
1959
+ "start": 45,
1960
+ "end": 74,
1961
+ "entity_type": "EMAIL",
1962
+ "surface": "Helen.Fixture06@aria-lab.test",
1963
+ "language": "en",
1964
+ "field": null
1965
+ },
1966
+ {
1967
+ "start": 77,
1968
+ "end": 88,
1969
+ "entity_type": "PHONE",
1970
+ "surface": "09126506157",
1971
+ "language": "en",
1972
+ "field": null
1973
+ }
1974
+ ],
1975
+ "meta": [
1976
+ {
1977
+ "field": "author",
1978
+ "entity_type": "PERSON",
1979
+ "surface": "Helen Fixture"
1980
+ },
1981
+ {
1982
+ "field": "keywords",
1983
+ "entity_type": "EMAIL",
1984
+ "surface": "Helen.Fixture06@aria-lab.test"
1985
+ }
1986
+ ]
1987
+ },
1988
+ {
1989
+ "doc_id": "DLP-PDF-MEMO-07",
1990
+ "spans": [
1991
+ {
1992
+ "start": 22,
1993
+ "end": 33,
1994
+ "entity_type": "PERSON",
1995
+ "surface": "Ava Fixture",
1996
+ "language": "en",
1997
+ "field": null
1998
+ },
1999
+ {
2000
+ "start": 43,
2001
+ "end": 70,
2002
+ "entity_type": "EMAIL",
2003
+ "surface": "Ava.Fixture07@aria-lab.test",
2004
+ "language": "en",
2005
+ "field": null
2006
+ },
2007
+ {
2008
+ "start": 73,
2009
+ "end": 84,
2010
+ "entity_type": "PHONE",
2011
+ "surface": "09126633882",
2012
+ "language": "en",
2013
+ "field": null
2014
+ }
2015
+ ],
2016
+ "meta": [
2017
+ {
2018
+ "field": "author",
2019
+ "entity_type": "PERSON",
2020
+ "surface": "Ava Fixture"
2021
+ },
2022
+ {
2023
+ "field": "keywords",
2024
+ "entity_type": "EMAIL",
2025
+ "surface": "Ava.Fixture07@aria-lab.test"
2026
+ }
2027
+ ]
2028
+ },
2029
+ {
2030
+ "doc_id": "DLP-PDF-MEMO-08",
2031
+ "spans": [
2032
+ {
2033
+ "start": 22,
2034
+ "end": 33,
2035
+ "entity_type": "PERSON",
2036
+ "surface": "Nima Labian",
2037
+ "language": "en",
2038
+ "field": null
2039
+ },
2040
+ {
2041
+ "start": 43,
2042
+ "end": 70,
2043
+ "entity_type": "EMAIL",
2044
+ "surface": "Nima.Labian08@aria-lab.test",
2045
+ "language": "en",
2046
+ "field": null
2047
+ },
2048
+ {
2049
+ "start": 73,
2050
+ "end": 84,
2051
+ "entity_type": "PHONE",
2052
+ "surface": "09128661790",
2053
+ "language": "en",
2054
+ "field": null
2055
+ }
2056
+ ],
2057
+ "meta": [
2058
+ {
2059
+ "field": "author",
2060
+ "entity_type": "PERSON",
2061
+ "surface": "Nima Labian"
2062
+ },
2063
+ {
2064
+ "field": "keywords",
2065
+ "entity_type": "EMAIL",
2066
+ "surface": "Nima.Labian08@aria-lab.test"
2067
+ }
2068
+ ]
2069
+ },
2070
+ {
2071
+ "doc_id": "DLP-PDF-MEMO-09",
2072
+ "spans": [
2073
+ {
2074
+ "start": 22,
2075
+ "end": 33,
2076
+ "entity_type": "PERSON",
2077
+ "surface": "Ava Fixture",
2078
+ "language": "en",
2079
+ "field": null
2080
+ },
2081
+ {
2082
+ "start": 43,
2083
+ "end": 70,
2084
+ "entity_type": "EMAIL",
2085
+ "surface": "Ava.Fixture09@aria-lab.test",
2086
+ "language": "en",
2087
+ "field": null
2088
+ },
2089
+ {
2090
+ "start": 73,
2091
+ "end": 84,
2092
+ "entity_type": "PHONE",
2093
+ "surface": "09122509645",
2094
+ "language": "en",
2095
+ "field": null
2096
+ }
2097
+ ],
2098
+ "meta": [
2099
+ {
2100
+ "field": "author",
2101
+ "entity_type": "PERSON",
2102
+ "surface": "Ava Fixture"
2103
+ },
2104
+ {
2105
+ "field": "keywords",
2106
+ "entity_type": "EMAIL",
2107
+ "surface": "Ava.Fixture09@aria-lab.test"
2108
+ }
2109
+ ]
2110
+ },
2111
+ {
2112
+ "doc_id": "DLP-PDF-MEMO-10",
2113
+ "spans": [
2114
+ {
2115
+ "start": 22,
2116
+ "end": 34,
2117
+ "entity_type": "PERSON",
2118
+ "surface": "Nima Fixture",
2119
+ "language": "en",
2120
+ "field": null
2121
+ },
2122
+ {
2123
+ "start": 44,
2124
+ "end": 72,
2125
+ "entity_type": "EMAIL",
2126
+ "surface": "Nima.Fixture10@aria-lab.test",
2127
+ "language": "en",
2128
+ "field": null
2129
+ },
2130
+ {
2131
+ "start": 75,
2132
+ "end": 86,
2133
+ "entity_type": "PHONE",
2134
+ "surface": "09122528399",
2135
+ "language": "en",
2136
+ "field": null
2137
+ }
2138
+ ],
2139
+ "meta": [
2140
+ {
2141
+ "field": "author",
2142
+ "entity_type": "PERSON",
2143
+ "surface": "Nima Fixture"
2144
+ },
2145
+ {
2146
+ "field": "keywords",
2147
+ "entity_type": "EMAIL",
2148
+ "surface": "Nima.Fixture10@aria-lab.test"
2149
+ }
2150
+ ]
2151
+ },
2152
+ {
2153
+ "doc_id": "DLP-HARDNEG-01",
2154
+ "spans": [],
2155
+ "meta": []
2156
+ },
2157
+ {
2158
+ "doc_id": "DLP-HARDNEG-02",
2159
+ "spans": [],
2160
+ "meta": []
2161
+ },
2162
+ {
2163
+ "doc_id": "DLP-HARDNEG-03",
2164
+ "spans": [],
2165
+ "meta": []
2166
+ },
2167
+ {
2168
+ "doc_id": "DLP-HARDNEG-04",
2169
+ "spans": [],
2170
+ "meta": []
2171
+ },
2172
+ {
2173
+ "doc_id": "DLP-HARDNEG-05",
2174
+ "spans": [],
2175
+ "meta": []
2176
+ },
2177
+ {
2178
+ "doc_id": "DLP-HARDNEG-06",
2179
+ "spans": [],
2180
+ "meta": []
2181
+ },
2182
+ {
2183
+ "doc_id": "DLP-HARDNEG-07",
2184
+ "spans": [],
2185
+ "meta": []
2186
+ },
2187
+ {
2188
+ "doc_id": "DLP-HARDNEG-08",
2189
+ "spans": [],
2190
+ "meta": []
2191
+ },
2192
+ {
2193
+ "doc_id": "DLP-HARDNEG-09",
2194
+ "spans": [],
2195
+ "meta": []
2196
+ },
2197
+ {
2198
+ "doc_id": "DLP-HARDNEG-10",
2199
+ "spans": [],
2200
+ "meta": []
2201
+ },
2202
+ {
2203
+ "doc_id": "DLP-TICKET-01",
2204
+ "spans": [
2205
+ {
2206
+ "start": 56,
2207
+ "end": 69,
2208
+ "entity_type": "PERSON",
2209
+ "surface": "Leila Fixture",
2210
+ "language": "en",
2211
+ "field": null
2212
+ },
2213
+ {
2214
+ "start": 72,
2215
+ "end": 84,
2216
+ "entity_type": "PERSON",
2217
+ "surface": "سارا آزمایشی",
2218
+ "language": "fa",
2219
+ "field": null
2220
+ },
2221
+ {
2222
+ "start": 85,
2223
+ "end": 114,
2224
+ "entity_type": "EMAIL",
2225
+ "surface": "Leila.Fixture01@aria-lab.test",
2226
+ "language": "en",
2227
+ "field": null
2228
+ },
2229
+ {
2230
+ "start": 115,
2231
+ "end": 126,
2232
+ "entity_type": "PHONE",
2233
+ "surface": "09126663672",
2234
+ "language": "en",
2235
+ "field": null
2236
+ },
2237
+ {
2238
+ "start": 127,
2239
+ "end": 137,
2240
+ "entity_type": "IR_NATIONAL_ID",
2241
+ "surface": "8810500075",
2242
+ "language": "fa",
2243
+ "field": null
2244
+ }
2245
+ ],
2246
+ "meta": []
2247
+ },
2248
+ {
2249
+ "doc_id": "DLP-TICKET-02",
2250
+ "spans": [
2251
+ {
2252
+ "start": 56,
2253
+ "end": 69,
2254
+ "entity_type": "PERSON",
2255
+ "surface": "Leila Fixture",
2256
+ "language": "en",
2257
+ "field": null
2258
+ },
2259
+ {
2260
+ "start": 72,
2261
+ "end": 85,
2262
+ "entity_type": "PERSON",
2263
+ "surface": "امید نمونه‌ای",
2264
+ "language": "fa",
2265
+ "field": null
2266
+ },
2267
+ {
2268
+ "start": 86,
2269
+ "end": 115,
2270
+ "entity_type": "EMAIL",
2271
+ "surface": "Leila.Fixture02@aria-lab.test",
2272
+ "language": "en",
2273
+ "field": null
2274
+ },
2275
+ {
2276
+ "start": 116,
2277
+ "end": 127,
2278
+ "entity_type": "PHONE",
2279
+ "surface": "09124960362",
2280
+ "language": "en",
2281
+ "field": null
2282
+ },
2283
+ {
2284
+ "start": 128,
2285
+ "end": 138,
2286
+ "entity_type": "IR_NATIONAL_ID",
2287
+ "surface": "7433177132",
2288
+ "language": "fa",
2289
+ "field": null
2290
+ }
2291
+ ],
2292
+ "meta": []
2293
+ },
2294
+ {
2295
+ "doc_id": "DLP-TICKET-03",
2296
+ "spans": [
2297
+ {
2298
+ "start": 56,
2299
+ "end": 67,
2300
+ "entity_type": "PERSON",
2301
+ "surface": "Ava Fixture",
2302
+ "language": "en",
2303
+ "field": null
2304
+ },
2305
+ {
2306
+ "start": 70,
2307
+ "end": 83,
2308
+ "entity_type": "PERSON",
2309
+ "surface": "نیما نمونه‌ای",
2310
+ "language": "fa",
2311
+ "field": null
2312
+ },
2313
+ {
2314
+ "start": 84,
2315
+ "end": 111,
2316
+ "entity_type": "EMAIL",
2317
+ "surface": "Ava.Fixture03@aria-lab.test",
2318
+ "language": "en",
2319
+ "field": null
2320
+ },
2321
+ {
2322
+ "start": 112,
2323
+ "end": 123,
2324
+ "entity_type": "PHONE",
2325
+ "surface": "09129013939",
2326
+ "language": "en",
2327
+ "field": null
2328
+ },
2329
+ {
2330
+ "start": 124,
2331
+ "end": 134,
2332
+ "entity_type": "IR_NATIONAL_ID",
2333
+ "surface": "4751024639",
2334
+ "language": "fa",
2335
+ "field": null
2336
+ }
2337
+ ],
2338
+ "meta": []
2339
+ },
2340
+ {
2341
+ "doc_id": "DLP-TICKET-04",
2342
+ "spans": [
2343
+ {
2344
+ "start": 56,
2345
+ "end": 68,
2346
+ "entity_type": "PERSON",
2347
+ "surface": "Omar Fixture",
2348
+ "language": "en",
2349
+ "field": null
2350
+ },
2351
+ {
2352
+ "start": 71,
2353
+ "end": 84,
2354
+ "entity_type": "PERSON",
2355
+ "surface": "سارا نمونه‌ای",
2356
+ "language": "fa",
2357
+ "field": null
2358
+ },
2359
+ {
2360
+ "start": 85,
2361
+ "end": 113,
2362
+ "entity_type": "EMAIL",
2363
+ "surface": "Omar.Fixture04@aria-lab.test",
2364
+ "language": "en",
2365
+ "field": null
2366
+ },
2367
+ {
2368
+ "start": 114,
2369
+ "end": 125,
2370
+ "entity_type": "PHONE",
2371
+ "surface": "09124265160",
2372
+ "language": "en",
2373
+ "field": null
2374
+ },
2375
+ {
2376
+ "start": 126,
2377
+ "end": 136,
2378
+ "entity_type": "IR_NATIONAL_ID",
2379
+ "surface": "0843397276",
2380
+ "language": "fa",
2381
+ "field": null
2382
+ }
2383
+ ],
2384
+ "meta": []
2385
+ },
2386
+ {
2387
+ "doc_id": "DLP-TICKET-05",
2388
+ "spans": [
2389
+ {
2390
+ "start": 56,
2391
+ "end": 67,
2392
+ "entity_type": "PERSON",
2393
+ "surface": "Ava Fixture",
2394
+ "language": "en",
2395
+ "field": null
2396
+ },
2397
+ {
2398
+ "start": 70,
2399
+ "end": 83,
2400
+ "entity_type": "PERSON",
2401
+ "surface": "امید نمونه‌ای",
2402
+ "language": "fa",
2403
+ "field": null
2404
+ },
2405
+ {
2406
+ "start": 84,
2407
+ "end": 111,
2408
+ "entity_type": "EMAIL",
2409
+ "surface": "Ava.Fixture05@aria-lab.test",
2410
+ "language": "en",
2411
+ "field": null
2412
+ },
2413
+ {
2414
+ "start": 112,
2415
+ "end": 123,
2416
+ "entity_type": "PHONE",
2417
+ "surface": "09127385597",
2418
+ "language": "en",
2419
+ "field": null
2420
+ },
2421
+ {
2422
+ "start": 124,
2423
+ "end": 134,
2424
+ "entity_type": "IR_NATIONAL_ID",
2425
+ "surface": "6714749521",
2426
+ "language": "fa",
2427
+ "field": null
2428
+ }
2429
+ ],
2430
+ "meta": []
2431
+ },
2432
+ {
2433
+ "doc_id": "DLP-TICKET-06",
2434
+ "spans": [
2435
+ {
2436
+ "start": 56,
2437
+ "end": 68,
2438
+ "entity_type": "PERSON",
2439
+ "surface": "Omar Fixture",
2440
+ "language": "en",
2441
+ "field": null
2442
+ },
2443
+ {
2444
+ "start": 71,
2445
+ "end": 83,
2446
+ "entity_type": "PERSON",
2447
+ "surface": "لیلا آزمایشی",
2448
+ "language": "fa",
2449
+ "field": null
2450
+ },
2451
+ {
2452
+ "start": 84,
2453
+ "end": 112,
2454
+ "entity_type": "EMAIL",
2455
+ "surface": "Omar.Fixture06@aria-lab.test",
2456
+ "language": "en",
2457
+ "field": null
2458
+ },
2459
+ {
2460
+ "start": 113,
2461
+ "end": 124,
2462
+ "entity_type": "PHONE",
2463
+ "surface": "09126454625",
2464
+ "language": "en",
2465
+ "field": null
2466
+ },
2467
+ {
2468
+ "start": 125,
2469
+ "end": 135,
2470
+ "entity_type": "IR_NATIONAL_ID",
2471
+ "surface": "7904846901",
2472
+ "language": "fa",
2473
+ "field": null
2474
+ }
2475
+ ],
2476
+ "meta": []
2477
+ },
2478
+ {
2479
+ "doc_id": "DLP-TICKET-07",
2480
+ "spans": [
2481
+ {
2482
+ "start": 56,
2483
+ "end": 69,
2484
+ "entity_type": "PERSON",
2485
+ "surface": "Helen Fixture",
2486
+ "language": "en",
2487
+ "field": null
2488
+ },
2489
+ {
2490
+ "start": 72,
2491
+ "end": 84,
2492
+ "entity_type": "PERSON",
2493
+ "surface": "سارا آزمایشی",
2494
+ "language": "fa",
2495
+ "field": null
2496
+ },
2497
+ {
2498
+ "start": 85,
2499
+ "end": 114,
2500
+ "entity_type": "EMAIL",
2501
+ "surface": "Helen.Fixture07@aria-lab.test",
2502
+ "language": "en",
2503
+ "field": null
2504
+ },
2505
+ {
2506
+ "start": 115,
2507
+ "end": 126,
2508
+ "entity_type": "PHONE",
2509
+ "surface": "09127938081",
2510
+ "language": "en",
2511
+ "field": null
2512
+ },
2513
+ {
2514
+ "start": 127,
2515
+ "end": 137,
2516
+ "entity_type": "IR_NATIONAL_ID",
2517
+ "surface": "6114411269",
2518
+ "language": "fa",
2519
+ "field": null
2520
+ }
2521
+ ],
2522
+ "meta": []
2523
+ },
2524
+ {
2525
+ "doc_id": "DLP-TICKET-08",
2526
+ "spans": [
2527
+ {
2528
+ "start": 56,
2529
+ "end": 67,
2530
+ "entity_type": "PERSON",
2531
+ "surface": "Ava Fixture",
2532
+ "language": "en",
2533
+ "field": null
2534
+ },
2535
+ {
2536
+ "start": 70,
2537
+ "end": 83,
2538
+ "entity_type": "PERSON",
2539
+ "surface": "لیلا نمونه‌ای",
2540
+ "language": "fa",
2541
+ "field": null
2542
+ },
2543
+ {
2544
+ "start": 84,
2545
+ "end": 111,
2546
+ "entity_type": "EMAIL",
2547
+ "surface": "Ava.Fixture08@aria-lab.test",
2548
+ "language": "en",
2549
+ "field": null
2550
+ },
2551
+ {
2552
+ "start": 112,
2553
+ "end": 123,
2554
+ "entity_type": "PHONE",
2555
+ "surface": "09125317904",
2556
+ "language": "en",
2557
+ "field": null
2558
+ },
2559
+ {
2560
+ "start": 124,
2561
+ "end": 134,
2562
+ "entity_type": "IR_NATIONAL_ID",
2563
+ "surface": "2799826970",
2564
+ "language": "fa",
2565
+ "field": null
2566
+ }
2567
+ ],
2568
+ "meta": []
2569
+ },
2570
+ {
2571
+ "doc_id": "DLP-TICKET-09",
2572
+ "spans": [
2573
+ {
2574
+ "start": 56,
2575
+ "end": 68,
2576
+ "entity_type": "PERSON",
2577
+ "surface": "Omar Fixture",
2578
+ "language": "en",
2579
+ "field": null
2580
+ },
2581
+ {
2582
+ "start": 71,
2583
+ "end": 84,
2584
+ "entity_type": "PERSON",
2585
+ "surface": "نیما نمونه‌ای",
2586
+ "language": "fa",
2587
+ "field": null
2588
+ },
2589
+ {
2590
+ "start": 85,
2591
+ "end": 113,
2592
+ "entity_type": "EMAIL",
2593
+ "surface": "Omar.Fixture09@aria-lab.test",
2594
+ "language": "en",
2595
+ "field": null
2596
+ },
2597
+ {
2598
+ "start": 114,
2599
+ "end": 125,
2600
+ "entity_type": "PHONE",
2601
+ "surface": "09125714005",
2602
+ "language": "en",
2603
+ "field": null
2604
+ },
2605
+ {
2606
+ "start": 126,
2607
+ "end": 136,
2608
+ "entity_type": "IR_NATIONAL_ID",
2609
+ "surface": "0468470395",
2610
+ "language": "fa",
2611
+ "field": null
2612
+ }
2613
+ ],
2614
+ "meta": []
2615
+ },
2616
+ {
2617
+ "doc_id": "DLP-TICKET-10",
2618
+ "spans": [
2619
+ {
2620
+ "start": 56,
2621
+ "end": 68,
2622
+ "entity_type": "PERSON",
2623
+ "surface": "Helen Labian",
2624
+ "language": "en",
2625
+ "field": null
2626
+ },
2627
+ {
2628
+ "start": 71,
2629
+ "end": 84,
2630
+ "entity_type": "PERSON",
2631
+ "surface": "سارا نمونه‌ای",
2632
+ "language": "fa",
2633
+ "field": null
2634
+ },
2635
+ {
2636
+ "start": 85,
2637
+ "end": 113,
2638
+ "entity_type": "EMAIL",
2639
+ "surface": "Helen.Labian10@aria-lab.test",
2640
+ "language": "en",
2641
+ "field": null
2642
+ },
2643
+ {
2644
+ "start": 114,
2645
+ "end": 125,
2646
+ "entity_type": "PHONE",
2647
+ "surface": "09126818844",
2648
+ "language": "en",
2649
+ "field": null
2650
+ },
2651
+ {
2652
+ "start": 126,
2653
+ "end": 136,
2654
+ "entity_type": "IR_NATIONAL_ID",
2655
+ "surface": "5049585570",
2656
+ "language": "fa",
2657
+ "field": null
2658
+ }
2659
+ ],
2660
+ "meta": []
2661
+ }
2662
+ ]
data/protocol.md ADDED
@@ -0,0 +1,23 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # AriaDLP laboratory protocol
2
+
3
+ Synthetic bilingual DLP corpus. Seed 24. Split by template.
4
+
5
+ ## What is labeled
6
+
7
+ Person names from a closed laboratory lexicon, mailbox addresses on `aria-lab.test`,
8
+ Iranian mobile numbers, checksum-valid national IDs, Luhn-valid test PANs, IBANs,
9
+ and `ARIA-24xxx` employee ids.
10
+
11
+ ## Hard negatives
12
+
13
+ Invoice / SKU tokens, ISBN, semantic versions, `Mr. Protocol`, dates, and numbers
14
+ that fail national-ID / Luhn / IBAN checks.
15
+
16
+ ## PDF
17
+
18
+ `en_pdf_memo` is a generated text PDF. Gold includes body spans and metadata fields.
19
+ Acceptance is content removal plus metadata wipe, not drawing a rectangle.
20
+
21
+ ## Limits
22
+
23
+ This corpus does not prove production recall. Unknown items stay on the review list.
data/sample_preview.json ADDED
@@ -0,0 +1,88 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "run_id": "ariadlp-6a3c5f54da644ce9",
3
+ "status": "scoped_hold",
4
+ "status_label": "Scoped hold \u2014 labeled gold was removed from this export; unknown types stay listed",
5
+ "kpis": {
6
+ "n_docs": 20,
7
+ "n_spans": 70,
8
+ "n_review": 0,
9
+ "n_redacted": 70,
10
+ "policy_precision": 1.0,
11
+ "policy_recall": 1.0,
12
+ "baseline_precision": 0.8333333333333334,
13
+ "baseline_recall": 0.7142857142857143,
14
+ "residual_leak": 0.0,
15
+ "over_redaction": 0.0,
16
+ "leftover_exports": 0,
17
+ "metadata_leaks": 0,
18
+ "elapsed_ms": 52,
19
+ "engine_profile": "presidio-style-checksum-policy",
20
+ "policy_action": "mask",
21
+ "pages": 0,
22
+ "chars": 2897
23
+ },
24
+ "kpi_html": "<div class='kpi-grid'><article class='kpi-card'><p class='kpi-label'>Documents in this run</p><p class='kpi-value'>20</p><p class='kpi-sub'>presidio-style-checksum-policy</p></article><article class='kpi-card'><p class='kpi-label'>Policy spans</p><p class='kpi-value'>70</p><p class='kpi-sub'>checksum policy</p></article><article class='kpi-card'><p class='kpi-label'>Policy precision</p><p class='kpi-value'>100%</p><p class='kpi-sub'>baseline 83%</p></article><article class='kpi-card'><p class='kpi-label'>Policy recall</p><p class='kpi-value'>100%</p><p class='kpi-sub'>baseline 71%</p></article><article class='kpi-card'><p class='kpi-label'>Exports with leftover gold</p><p class='kpi-value'>0</p><p class='kpi-sub'>52 ms</p></article></div>",
25
+ "catalog": [
26
+ {
27
+ "scenario_id": "S01",
28
+ "method": "checksum_policy",
29
+ "precision": 1.0,
30
+ "recall": 1.0,
31
+ "over_redaction": 0.0,
32
+ "residual_leak": 0.0,
33
+ "leftover_exports": 0,
34
+ "metadata_leaks": 0,
35
+ "hardneg_hits": 0,
36
+ "n_docs": 20
37
+ },
38
+ {
39
+ "scenario_id": "S02",
40
+ "method": "checksum_policy",
41
+ "precision": 1.0,
42
+ "recall": 1.0,
43
+ "over_redaction": 0.0,
44
+ "residual_leak": 0.0,
45
+ "leftover_exports": 0,
46
+ "metadata_leaks": 0,
47
+ "hardneg_hits": 0,
48
+ "n_docs": 20
49
+ },
50
+ {
51
+ "scenario_id": "S03",
52
+ "method": "checksum_policy",
53
+ "precision": 1.0,
54
+ "recall": 1.0,
55
+ "over_redaction": 0.0,
56
+ "residual_leak": 0.0,
57
+ "leftover_exports": 0,
58
+ "metadata_leaks": 0,
59
+ "hardneg_hits": 0,
60
+ "n_docs": 10
61
+ },
62
+ {
63
+ "scenario_id": "S04",
64
+ "method": "checksum_policy",
65
+ "precision": 1.0,
66
+ "recall": 1.0,
67
+ "over_redaction": 0.0,
68
+ "residual_leak": 0.0,
69
+ "leftover_exports": 0,
70
+ "metadata_leaks": 0,
71
+ "hardneg_hits": 10,
72
+ "n_docs": 20
73
+ },
74
+ {
75
+ "scenario_id": "S05",
76
+ "method": "checksum_policy",
77
+ "precision": 1.0,
78
+ "recall": 1.0,
79
+ "over_redaction": 0.0,
80
+ "residual_leak": 0.0,
81
+ "leftover_exports": 0,
82
+ "metadata_leaks": 0,
83
+ "hardneg_hits": 10,
84
+ "n_docs": 80
85
+ }
86
+ ],
87
+ "precomputed": true
88
+ }
data/scenarios.jsonl ADDED
@@ -0,0 +1,5 @@
 
 
 
 
 
 
1
+ {"scenario_id":"S01","title":"English mailbox and invoice CSV","description":"Development templates: support email plus invoice CSV.","description_fa":"قالب‌های توسعه: ایمیل پشتیبانی و سی‌اس‌وی فاکتور.","doc_ids":["DLP-EN-EMAIL-01","DLP-EN-EMAIL-02","DLP-EN-EMAIL-03","DLP-EN-EMAIL-04","DLP-EN-EMAIL-05","DLP-EN-EMAIL-06","DLP-EN-EMAIL-07","DLP-EN-EMAIL-08","DLP-EN-EMAIL-09","DLP-EN-EMAIL-10","DLP-CSV-INV-01","DLP-CSV-INV-02","DLP-CSV-INV-03","DLP-CSV-INV-04","DLP-CSV-INV-05","DLP-CSV-INV-06","DLP-CSV-INV-07","DLP-CSV-INV-08","DLP-CSV-INV-09","DLP-CSV-INV-10"],"split":"development"}
2
+ {"scenario_id":"S02","title":"Persian HR and onboarding","description":"Persian digits / ZWNJ HR letter plus English onboarding with test PAN and IBAN.","description_fa":"نامه منابع انسانی با رقم فارسی و نیم‌فاصله به‌همراه برگه ورود.","doc_ids":["DLP-FA-HR-01","DLP-FA-HR-02","DLP-FA-HR-03","DLP-FA-HR-04","DLP-FA-HR-05","DLP-FA-HR-06","DLP-FA-HR-07","DLP-FA-HR-08","DLP-FA-HR-09","DLP-FA-HR-10","DLP-EN-ONB-01","DLP-EN-ONB-02","DLP-EN-ONB-03","DLP-EN-ONB-04","DLP-EN-ONB-05","DLP-EN-ONB-06","DLP-EN-ONB-07","DLP-EN-ONB-08","DLP-EN-ONB-09","DLP-EN-ONB-10"],"split":"development"}
3
+ {"scenario_id":"S03","title":"Hard negatives","description":"Failed checksums, ISBN, Mr. Protocol, versions. Policy should not over-redact.","description_fa":"checksum نامعتبر، شابک، آقای پروتکل و نسخه. سیاست نباید بیش‌پوشانی کند.","doc_ids":["DLP-HARDNEG-01","DLP-HARDNEG-02","DLP-HARDNEG-03","DLP-HARDNEG-04","DLP-HARDNEG-05","DLP-HARDNEG-06","DLP-HARDNEG-07","DLP-HARDNEG-08","DLP-HARDNEG-09","DLP-HARDNEG-10"],"split":"test"}
4
+ {"scenario_id":"S04","title":"PDF memo and Persian contract","description":"Text PDF with metadata PII plus a Persian contract that keeps invalid national IDs for review.","description_fa":"پی‌دی‌اف متنی با فراداده حساس و قرارداد فارسی با کد ملی نامعتبر برای بازبینی.","doc_ids":["DLP-FA-CTR-01","DLP-FA-CTR-02","DLP-FA-CTR-03","DLP-FA-CTR-04","DLP-FA-CTR-05","DLP-FA-CTR-06","DLP-FA-CTR-07","DLP-FA-CTR-08","DLP-FA-CTR-09","DLP-FA-CTR-10","DLP-PDF-MEMO-01","DLP-PDF-MEMO-02","DLP-PDF-MEMO-03","DLP-PDF-MEMO-04","DLP-PDF-MEMO-05","DLP-PDF-MEMO-06","DLP-PDF-MEMO-07","DLP-PDF-MEMO-08","DLP-PDF-MEMO-09","DLP-PDF-MEMO-10"],"split":"test"}
5
+ {"scenario_id":"S05","title":"Full 80-document catalog","description":"Template-disjoint development/test split, seed 24.","description_fa":"تفکیک قالب توسعه/آزمون، بذر ۲۴.","doc_ids":["DLP-EN-EMAIL-01","DLP-EN-EMAIL-02","DLP-EN-EMAIL-03","DLP-EN-EMAIL-04","DLP-EN-EMAIL-05","DLP-EN-EMAIL-06","DLP-EN-EMAIL-07","DLP-EN-EMAIL-08","DLP-EN-EMAIL-09","DLP-EN-EMAIL-10","DLP-FA-HR-01","DLP-FA-HR-02","DLP-FA-HR-03","DLP-FA-HR-04","DLP-FA-HR-05","DLP-FA-HR-06","DLP-FA-HR-07","DLP-FA-HR-08","DLP-FA-HR-09","DLP-FA-HR-10","DLP-CSV-INV-01","DLP-CSV-INV-02","DLP-CSV-INV-03","DLP-CSV-INV-04","DLP-CSV-INV-05","DLP-CSV-INV-06","DLP-CSV-INV-07","DLP-CSV-INV-08","DLP-CSV-INV-09","DLP-CSV-INV-10","DLP-EN-ONB-01","DLP-EN-ONB-02","DLP-EN-ONB-03","DLP-EN-ONB-04","DLP-EN-ONB-05","DLP-EN-ONB-06","DLP-EN-ONB-07","DLP-EN-ONB-08","DLP-EN-ONB-09","DLP-EN-ONB-10","DLP-FA-CTR-01","DLP-FA-CTR-02","DLP-FA-CTR-03","DLP-FA-CTR-04","DLP-FA-CTR-05","DLP-FA-CTR-06","DLP-FA-CTR-07","DLP-FA-CTR-08","DLP-FA-CTR-09","DLP-FA-CTR-10","DLP-PDF-MEMO-01","DLP-PDF-MEMO-02","DLP-PDF-MEMO-03","DLP-PDF-MEMO-04","DLP-PDF-MEMO-05","DLP-PDF-MEMO-06","DLP-PDF-MEMO-07","DLP-PDF-MEMO-08","DLP-PDF-MEMO-09","DLP-PDF-MEMO-10","DLP-HARDNEG-01","DLP-HARDNEG-02","DLP-HARDNEG-03","DLP-HARDNEG-04","DLP-HARDNEG-05","DLP-HARDNEG-06","DLP-HARDNEG-07","DLP-HARDNEG-08","DLP-HARDNEG-09","DLP-HARDNEG-10","DLP-TICKET-01","DLP-TICKET-02","DLP-TICKET-03","DLP-TICKET-04","DLP-TICKET-05","DLP-TICKET-06","DLP-TICKET-07","DLP-TICKET-08","DLP-TICKET-09","DLP-TICKET-10"],"split":"all"}
data/schema.json ADDED
@@ -0,0 +1,25 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_id": "aria.dlp.lab_document.v1",
3
+ "fields": [
4
+ "doc_id",
5
+ "template",
6
+ "split",
7
+ "language",
8
+ "format",
9
+ "body",
10
+ "gold_spans",
11
+ "gold_meta",
12
+ "hard_negatives"
13
+ ],
14
+ "entity_types": [
15
+ "PERSON",
16
+ "EMAIL",
17
+ "PHONE",
18
+ "IR_NATIONAL_ID",
19
+ "CREDIT_CARD",
20
+ "IBAN",
21
+ "EMPLOYEE_ID",
22
+ "UNKNOWN"
23
+ ],
24
+ "notes": "Synthetic only. Addresses use aria-lab.test. PANs are Luhn-valid test numbers."
25
+ }
data/splits.json ADDED
@@ -0,0 +1,17 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "development_templates": [
3
+ "en_onboarding",
4
+ "en_support_email",
5
+ "fa_hr_letter",
6
+ "mixed_invoice_csv"
7
+ ],
8
+ "test_templates": [
9
+ "en_pdf_memo",
10
+ "fa_contract_notice",
11
+ "hardneg_catalog",
12
+ "mixed_ticket"
13
+ ],
14
+ "rule": "Split is by document template, not random rows.",
15
+ "n_documents": 80,
16
+ "seed": 24
17
+ }
demo/.gitattributes ADDED
@@ -0,0 +1,11 @@
 
 
 
 
 
 
 
 
 
 
 
 
1
+ * text=auto eol=lf
2
+ *.py text eol=lf
3
+ *.md text eol=lf
4
+ *.csv text eol=lf
5
+ *.json text eol=lf
6
+ *.jsonl text eol=lf
7
+ *.yml text eol=lf
8
+ *.yaml text eol=lf
9
+ *.html text eol=lf
10
+ *.txt text eol=lf
11
+ *.log text eol=lf
demo/.gitignore ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ __pycache__/
2
+ *.py[cod]
3
+ .pytest_cache/
4
+ .venv/
5
+ .venv-dlp/
6
+ *.egg-info/
7
+ .DS_Store
8
+ reports/
demo/LICENSE ADDED
@@ -0,0 +1,17 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ Apache License
2
+ Version 2.0, January 2004
3
+ http://www.apache.org/licenses/
4
+
5
+ Copyright 2026 Aria AI
6
+
7
+ Licensed under the Apache License, Version 2.0 (the "License");
8
+ you may not use this file except in compliance with the License.
9
+ You may obtain a copy of the License at
10
+
11
+ http://www.apache.org/licenses/LICENSE-2.0
12
+
13
+ Unless required by applicable law or agreed to in writing, software
14
+ distributed under the License is distributed on an "AS IS" BASIS,
15
+ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
16
+ See the License for the specific language governing permissions and
17
+ limitations under the License.
demo/README.md ADDED
@@ -0,0 +1,55 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ title: AriaDLP Persian English Redaction
3
+ emoji: 🔏
4
+ colorFrom: blue
5
+ colorTo: indigo
6
+ sdk: gradio
7
+ sdk_version: 6.15.1
8
+ app_file: app.py
9
+ fullWidth: true
10
+ python_version: "3.12"
11
+ short_description: Find and redact lab PII in text, CSV, and PDF
12
+ pinned: true
13
+ license: apache-2.0
14
+ tags:
15
+ - cybersecurity
16
+ - pii-detection
17
+ - redaction
18
+ - document-analysis
19
+ - text-analysis
20
+ datasets:
21
+ - AriaAICompany/dlp-lab
22
+ ---
23
+
24
+ # AriaDLP — Persian / English Redaction Lab
25
+
26
+ **Aria AI** workstation for privacy and document-sharing reviews. Find **laboratory personal data** in text, CSV, and text-layer PDFs, then preview a `mask` or `remove` export before download.
27
+
28
+ The MVP is CPU-only. Regex-only is the baseline. The policy engine is a **Presidio-style recognizer registry** with Iranian national-ID, Luhn, and IBAN checksums, Persian/Latin digit folding, ZWNJ-aware context, and a closed name lexicon. Official Presidio Persian coverage is **not** assumed. A clean lab suite is **not** a claim that every sensitive value was removed.
29
+
30
+ ## What you can do
31
+
32
+ 1. Run a focused scenario or the full 80-document pack (seed 24).
33
+ 2. Compare regex-only vs checksum policy on the same documents.
34
+ 3. Open spans, suppress a false positive, and preview the export.
35
+ 4. Load S00 and paste text, or upload TXT / CSV / text PDF (10 pages).
36
+ 5. Download a hashed CSV/JSON bundle, plus a redacted PDF when the source was PDF.
37
+
38
+ Scanned PDFs, production identity documents, and live LLM calls are out of scope. An optional later model would only see already-redacted text.
39
+
40
+ ## Hosting
41
+
42
+ - Dataset: [AriaAICompany/dlp-lab](https://huggingface.co/datasets/AriaAICompany/dlp-lab)
43
+ - Collection: [Aria AI — Cybersecurity](https://huggingface.co/collections/AriaAICompany/aria-ai-cybersecurity-6ab01faa3995a3bb737b58a7)
44
+ - Live app: `alirezaaminzadeh/aria-dlp` after the Hub daily Space-creation cap resets (`python scripts/publish.py`)
45
+ - Organization card: `AriaAICompany/aria-dlp` (static; org Gradio needs Team/Enterprise)
46
+
47
+ ## Persian
48
+
49
+ آریادی‌ال‌پی داده شخصی آزمایشگاهی را در متن، CSV و PDF متن‌دار پیدا و پیش‌نمایش می‌کند. مبنا فقط regex است. سیاست checksum و رقم فارسی دارد. پوشش فارسی آماده Presidio فرض نشده. سوئیت تمیز به‌معنی حذف همه داده حساس نیست. PDF اسکن‌شده خارج از محدوده است.
50
+
51
+ ## Links
52
+
53
+ - [AriaAICompany](https://huggingface.co/AriaAICompany)
54
+ - [aria-ai.ir](https://aria-ai.ir/)
55
+ - Roadmap project 24 of 25
demo/app.py ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ import gradio as gr
2
+
3
+ from src.ui import CSS, build_demo
4
+
5
+ demo = build_demo()
6
+
7
+ if __name__ == "__main__":
8
+ demo.launch(theme=gr.themes.Soft(primary_hue="indigo", secondary_hue="blue"), css=CSS)
demo/evaluations/evaluation.json ADDED
@@ -0,0 +1,88 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "seed": 24,
3
+ "n_documents": 80,
4
+ "n_scenarios": 5,
5
+ "engine_profile": "presidio-style-checksum-policy",
6
+ "rows": [
7
+ {
8
+ "scenario_id": "S01",
9
+ "method": "checksum_policy",
10
+ "precision": 1.0,
11
+ "recall": 1.0,
12
+ "over_redaction": 0.0,
13
+ "residual_leak": 0.0,
14
+ "leftover_exports": 0,
15
+ "metadata_leaks": 0,
16
+ "hardneg_hits": 0,
17
+ "n_docs": 20
18
+ },
19
+ {
20
+ "scenario_id": "S02",
21
+ "method": "checksum_policy",
22
+ "precision": 1.0,
23
+ "recall": 1.0,
24
+ "over_redaction": 0.0,
25
+ "residual_leak": 0.0,
26
+ "leftover_exports": 0,
27
+ "metadata_leaks": 0,
28
+ "hardneg_hits": 0,
29
+ "n_docs": 20
30
+ },
31
+ {
32
+ "scenario_id": "S03",
33
+ "method": "checksum_policy",
34
+ "precision": 1.0,
35
+ "recall": 1.0,
36
+ "over_redaction": 0.0,
37
+ "residual_leak": 0.0,
38
+ "leftover_exports": 0,
39
+ "metadata_leaks": 0,
40
+ "hardneg_hits": 0,
41
+ "n_docs": 10
42
+ },
43
+ {
44
+ "scenario_id": "S04",
45
+ "method": "checksum_policy",
46
+ "precision": 1.0,
47
+ "recall": 1.0,
48
+ "over_redaction": 0.0,
49
+ "residual_leak": 0.0,
50
+ "leftover_exports": 0,
51
+ "metadata_leaks": 0,
52
+ "hardneg_hits": 10,
53
+ "n_docs": 20
54
+ },
55
+ {
56
+ "scenario_id": "S05",
57
+ "method": "checksum_policy",
58
+ "precision": 1.0,
59
+ "recall": 1.0,
60
+ "over_redaction": 0.0,
61
+ "residual_leak": 0.0,
62
+ "leftover_exports": 0,
63
+ "metadata_leaks": 0,
64
+ "hardneg_hits": 10,
65
+ "n_docs": 80
66
+ }
67
+ ],
68
+ "sample_kpis": {
69
+ "n_docs": 20,
70
+ "n_spans": 70,
71
+ "n_review": 0,
72
+ "n_redacted": 70,
73
+ "policy_precision": 1.0,
74
+ "policy_recall": 1.0,
75
+ "baseline_precision": 0.8333333333333334,
76
+ "baseline_recall": 0.7142857142857143,
77
+ "residual_leak": 0.0,
78
+ "over_redaction": 0.0,
79
+ "leftover_exports": 0,
80
+ "metadata_leaks": 0,
81
+ "elapsed_ms": 52,
82
+ "engine_profile": "presidio-style-checksum-policy",
83
+ "policy_action": "mask",
84
+ "pages": 0,
85
+ "chars": 2897
86
+ },
87
+ "note": "Synthetic laboratory corpus. A clean suite is not complete production redaction."
88
+ }
demo/examples/documents.jsonl ADDED
@@ -0,0 +1,80 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {"doc_id":"DLP-EN-EMAIL-01","template":"en_support_email","family":"en_support_email","split":"development","language":"en","format":"text","title":"English support mailbox","body":"From: Nima Fixture <Nima.Fixture01@aria-lab.test>\nPhone: 09124063383\nEmployee: ARIA-24001\nPlease reset the laboratory mailbox for Nima. Ticket ORD-2024-0001. Version 2.4.1.\n","metadata":{},"gold_spans":[{"start":6,"end":18,"entity_type":"PERSON","surface":"Nima Fixture","language":"en","field":null},{"start":20,"end":48,"entity_type":"EMAIL","surface":"Nima.Fixture01@aria-lab.test","language":"en","field":null},{"start":57,"end":68,"entity_type":"PHONE","surface":"09124063383","language":"en","field":null},{"start":79,"end":89,"entity_type":"EMPLOYEE_ID","surface":"ARIA-24001","language":"en","field":null}],"gold_meta":[],"hard_negatives":["ORD-2024-0001","2.4.1"],"description":"English support email with name, mailbox, phone, and employee id.","description_fa":"ایمیل پشتیبانی انگلیسی با نام، صندوق، تلفن و شناسه کارمند."}
2
+ {"doc_id":"DLP-EN-EMAIL-02","template":"en_support_email","family":"en_support_email","split":"development","language":"en","format":"text","title":"English support mailbox","body":"From: Noah Labian <Noah.Labian02@aria-lab.test>\nPhone: 09124256381\nEmployee: ARIA-24002\nPlease reset the laboratory mailbox for Noah. Ticket ORD-2024-0002. Version 2.4.1.\n","metadata":{},"gold_spans":[{"start":6,"end":17,"entity_type":"PERSON","surface":"Noah Labian","language":"en","field":null},{"start":19,"end":46,"entity_type":"EMAIL","surface":"Noah.Labian02@aria-lab.test","language":"en","field":null},{"start":55,"end":66,"entity_type":"PHONE","surface":"09124256381","language":"en","field":null},{"start":77,"end":87,"entity_type":"EMPLOYEE_ID","surface":"ARIA-24002","language":"en","field":null}],"gold_meta":[],"hard_negatives":["ORD-2024-0002","2.4.1"],"description":"English support email with name, mailbox, phone, and employee id.","description_fa":"ایمیل پشتیبانی انگلیسی با نام، صندوق، تلفن و شناسه کارمند."}
3
+ {"doc_id":"DLP-EN-EMAIL-03","template":"en_support_email","family":"en_support_email","split":"development","language":"en","format":"text","title":"English support mailbox","body":"From: Noah Labian <Noah.Labian03@aria-lab.test>\nPhone: 09123538397\nEmployee: ARIA-24003\nPlease reset the laboratory mailbox for Noah. Ticket ORD-2024-0003. Version 2.4.1.\n","metadata":{},"gold_spans":[{"start":6,"end":17,"entity_type":"PERSON","surface":"Noah Labian","language":"en","field":null},{"start":19,"end":46,"entity_type":"EMAIL","surface":"Noah.Labian03@aria-lab.test","language":"en","field":null},{"start":55,"end":66,"entity_type":"PHONE","surface":"09123538397","language":"en","field":null},{"start":77,"end":87,"entity_type":"EMPLOYEE_ID","surface":"ARIA-24003","language":"en","field":null}],"gold_meta":[],"hard_negatives":["ORD-2024-0003","2.4.1"],"description":"English support email with name, mailbox, phone, and employee id.","description_fa":"ایمیل پشتیبانی انگلیسی با نام، صندوق، تلفن و شناسه کارمند."}
4
+ {"doc_id":"DLP-EN-EMAIL-04","template":"en_support_email","family":"en_support_email","split":"development","language":"en","format":"text","title":"English support mailbox","body":"From: Nima Fixture <Nima.Fixture04@aria-lab.test>\nPhone: 09121213266\nEmployee: ARIA-24004\nPlease reset the laboratory mailbox for Nima. Ticket ORD-2024-0004. Version 2.4.1.\n","metadata":{},"gold_spans":[{"start":6,"end":18,"entity_type":"PERSON","surface":"Nima Fixture","language":"en","field":null},{"start":20,"end":48,"entity_type":"EMAIL","surface":"Nima.Fixture04@aria-lab.test","language":"en","field":null},{"start":57,"end":68,"entity_type":"PHONE","surface":"09121213266","language":"en","field":null},{"start":79,"end":89,"entity_type":"EMPLOYEE_ID","surface":"ARIA-24004","language":"en","field":null}],"gold_meta":[],"hard_negatives":["ORD-2024-0004","2.4.1"],"description":"English support email with name, mailbox, phone, and employee id.","description_fa":"ایمیل پشتیبانی انگلیسی با نام، صندوق، تلفن و شناسه کارمند."}
5
+ {"doc_id":"DLP-EN-EMAIL-05","template":"en_support_email","family":"en_support_email","split":"development","language":"en","format":"text","title":"English support mailbox","body":"From: Omar Fixture <Omar.Fixture05@aria-lab.test>\nPhone: 09122939341\nEmployee: ARIA-24005\nPlease reset the laboratory mailbox for Omar. Ticket ORD-2024-0005. Version 2.4.1.\n","metadata":{},"gold_spans":[{"start":6,"end":18,"entity_type":"PERSON","surface":"Omar Fixture","language":"en","field":null},{"start":20,"end":48,"entity_type":"EMAIL","surface":"Omar.Fixture05@aria-lab.test","language":"en","field":null},{"start":57,"end":68,"entity_type":"PHONE","surface":"09122939341","language":"en","field":null},{"start":79,"end":89,"entity_type":"EMPLOYEE_ID","surface":"ARIA-24005","language":"en","field":null}],"gold_meta":[],"hard_negatives":["ORD-2024-0005","2.4.1"],"description":"English support email with name, mailbox, phone, and employee id.","description_fa":"ایمیل پشتیبانی انگلیسی با نام، صندوق، تلفن و شناسه کارمند."}
6
+ {"doc_id":"DLP-EN-EMAIL-06","template":"en_support_email","family":"en_support_email","split":"development","language":"en","format":"text","title":"English support mailbox","body":"From: Ava Labian <Ava.Labian06@aria-lab.test>\nPhone: 09129242814\nEmployee: ARIA-24006\nPlease reset the laboratory mailbox for Ava. Ticket ORD-2024-0006. Version 2.4.1.\n","metadata":{},"gold_spans":[{"start":6,"end":16,"entity_type":"PERSON","surface":"Ava Labian","language":"en","field":null},{"start":18,"end":44,"entity_type":"EMAIL","surface":"Ava.Labian06@aria-lab.test","language":"en","field":null},{"start":53,"end":64,"entity_type":"PHONE","surface":"09129242814","language":"en","field":null},{"start":75,"end":85,"entity_type":"EMPLOYEE_ID","surface":"ARIA-24006","language":"en","field":null}],"gold_meta":[],"hard_negatives":["ORD-2024-0006","2.4.1"],"description":"English support email with name, mailbox, phone, and employee id.","description_fa":"ایمیل پشتیبانی انگلیسی با نام، صندوق، تلفن و شناسه کارمند."}
7
+ {"doc_id":"DLP-EN-EMAIL-07","template":"en_support_email","family":"en_support_email","split":"development","language":"en","format":"text","title":"English support mailbox","body":"From: Nima Fixture <Nima.Fixture07@aria-lab.test>\nPhone: 09126110356\nEmployee: ARIA-24007\nPlease reset the laboratory mailbox for Nima. Ticket ORD-2024-0007. Version 2.4.1.\n","metadata":{},"gold_spans":[{"start":6,"end":18,"entity_type":"PERSON","surface":"Nima Fixture","language":"en","field":null},{"start":20,"end":48,"entity_type":"EMAIL","surface":"Nima.Fixture07@aria-lab.test","language":"en","field":null},{"start":57,"end":68,"entity_type":"PHONE","surface":"09126110356","language":"en","field":null},{"start":79,"end":89,"entity_type":"EMPLOYEE_ID","surface":"ARIA-24007","language":"en","field":null}],"gold_meta":[],"hard_negatives":["ORD-2024-0007","2.4.1"],"description":"English support email with name, mailbox, phone, and employee id.","description_fa":"ایمیل پشتیبانی انگلیسی با نام، صندوق، تلفن و شناسه کارمند."}
8
+ {"doc_id":"DLP-EN-EMAIL-08","template":"en_support_email","family":"en_support_email","split":"development","language":"en","format":"text","title":"English support mailbox","body":"From: Omar Labian <Omar.Labian08@aria-lab.test>\nPhone: 09125259879\nEmployee: ARIA-24008\nPlease reset the laboratory mailbox for Omar. Ticket ORD-2024-0008. Version 2.4.1.\n","metadata":{},"gold_spans":[{"start":6,"end":17,"entity_type":"PERSON","surface":"Omar Labian","language":"en","field":null},{"start":19,"end":46,"entity_type":"EMAIL","surface":"Omar.Labian08@aria-lab.test","language":"en","field":null},{"start":55,"end":66,"entity_type":"PHONE","surface":"09125259879","language":"en","field":null},{"start":77,"end":87,"entity_type":"EMPLOYEE_ID","surface":"ARIA-24008","language":"en","field":null}],"gold_meta":[],"hard_negatives":["ORD-2024-0008","2.4.1"],"description":"English support email with name, mailbox, phone, and employee id.","description_fa":"ایمیل پشتیبانی انگلیسی با نام، صندوق، تلفن و شناسه کارمند."}
9
+ {"doc_id":"DLP-EN-EMAIL-09","template":"en_support_email","family":"en_support_email","split":"development","language":"en","format":"text","title":"English support mailbox","body":"From: Helen Labian <Helen.Labian09@aria-lab.test>\nPhone: 09126483978\nEmployee: ARIA-24009\nPlease reset the laboratory mailbox for Helen. Ticket ORD-2024-0009. Version 2.4.1.\n","metadata":{},"gold_spans":[{"start":6,"end":18,"entity_type":"PERSON","surface":"Helen Labian","language":"en","field":null},{"start":20,"end":48,"entity_type":"EMAIL","surface":"Helen.Labian09@aria-lab.test","language":"en","field":null},{"start":57,"end":68,"entity_type":"PHONE","surface":"09126483978","language":"en","field":null},{"start":79,"end":89,"entity_type":"EMPLOYEE_ID","surface":"ARIA-24009","language":"en","field":null}],"gold_meta":[],"hard_negatives":["ORD-2024-0009","2.4.1"],"description":"English support email with name, mailbox, phone, and employee id.","description_fa":"ایمیل پشتیبانی انگلیسی با نام، صندوق، تلفن و شناسه کارمند."}
10
+ {"doc_id":"DLP-EN-EMAIL-10","template":"en_support_email","family":"en_support_email","split":"development","language":"en","format":"text","title":"English support mailbox","body":"From: Leila Labian <Leila.Labian10@aria-lab.test>\nPhone: 09127157194\nEmployee: ARIA-24010\nPlease reset the laboratory mailbox for Leila. Ticket ORD-2024-0010. Version 2.4.1.\n","metadata":{},"gold_spans":[{"start":6,"end":18,"entity_type":"PERSON","surface":"Leila Labian","language":"en","field":null},{"start":20,"end":48,"entity_type":"EMAIL","surface":"Leila.Labian10@aria-lab.test","language":"en","field":null},{"start":57,"end":68,"entity_type":"PHONE","surface":"09127157194","language":"en","field":null},{"start":79,"end":89,"entity_type":"EMPLOYEE_ID","surface":"ARIA-24010","language":"en","field":null}],"gold_meta":[],"hard_negatives":["ORD-2024-0010","2.4.1"],"description":"English support email with name, mailbox, phone, and employee id.","description_fa":"ایمیل پشتیبانی انگلیسی با نام، صندوق، تلفن و شناسه کارمند."}
11
+ {"doc_id":"DLP-FA-HR-01","template":"fa_hr_letter","family":"fa_hr_letter","split":"development","language":"fa","format":"text","title":"Persian HR letter","body":"خانم سارا نمونه‌ای،\nکد‌ملی ۵۵۱۴۱۳۹۲۷۱\nتلفن همراه ۰۹۱۲۵۷۶۷۰۹۶\nایمیل user.01@aria-lab.test\nشماره پیگیری INV-1403-0001 و نسخه ۳٫۲٫۱ فقط ارجاع داخلی‌اند.\n","metadata":{},"gold_spans":[{"start":5,"end":18,"entity_type":"PERSON","surface":"سارا نمونه‌ای","language":"fa","field":null},{"start":27,"end":37,"entity_type":"IR_NATIONAL_ID","surface":"۵۵۱۴۱۳۹۲۷۱","language":"fa","field":null},{"start":49,"end":60,"entity_type":"PHONE","surface":"۰۹۱۲۵۷۶۷۰۹۶","language":"fa","field":null},{"start":67,"end":88,"entity_type":"EMAIL","surface":"user.01@aria-lab.test","language":"en","field":null}],"gold_meta":[],"hard_negatives":["INV-1403-0001"],"description":"Persian HR note with ZWNJ, Persian digits, national id, and phone.","description_fa":"نامه منابع انسانی با نیم‌فاصله، رقم فارسی، کد ملی و تلفن."}
12
+ {"doc_id":"DLP-FA-HR-02","template":"fa_hr_letter","family":"fa_hr_letter","split":"development","language":"fa","format":"text","title":"Persian HR letter","body":"خانم لیلا آزمایشی،\nکد‌ملی ۳۱۹۷۲۲۵۷۳۲\nتلفن همراه ۰۹۱۲۳۲۰۳۶۶۹\nایمیل user.02@aria-lab.test\nشماره پیگیری INV-1403-0002 و نسخه ۳٫۲٫۱ فقط ارجاع داخلی‌اند.\n","metadata":{},"gold_spans":[{"start":5,"end":17,"entity_type":"PERSON","surface":"لیلا آزمایشی","language":"fa","field":null},{"start":26,"end":36,"entity_type":"IR_NATIONAL_ID","surface":"۳۱۹۷۲۲۵۷۳۲","language":"fa","field":null},{"start":48,"end":59,"entity_type":"PHONE","surface":"۰۹۱۲۳۲۰۳۶۶۹","language":"fa","field":null},{"start":66,"end":87,"entity_type":"EMAIL","surface":"user.02@aria-lab.test","language":"en","field":null}],"gold_meta":[],"hard_negatives":["INV-1403-0002"],"description":"Persian HR note with ZWNJ, Persian digits, national id, and phone.","description_fa":"نامه منابع انسانی با نیم‌فاصله، رقم فارسی، کد ملی و تلفن."}
13
+ {"doc_id":"DLP-FA-HR-03","template":"fa_hr_letter","family":"fa_hr_letter","split":"development","language":"fa","format":"text","title":"Persian HR letter","body":"خانم امید آزمایشی،\nکد‌ملی ۸۳۸۳۱۲۴۵۸۹\nتلفن همراه ۰۹۱۲۴۸۶۵۴۴۱\nایمیل user.03@aria-lab.test\nشماره پیگیری INV-1403-0003 و نسخه ۳٫۲٫۱ فقط ارجاع داخلی‌اند.\n","metadata":{},"gold_spans":[{"start":5,"end":17,"entity_type":"PERSON","surface":"امید آزمایشی","language":"fa","field":null},{"start":26,"end":36,"entity_type":"IR_NATIONAL_ID","surface":"۸۳۸۳۱۲۴۵۸۹","language":"fa","field":null},{"start":48,"end":59,"entity_type":"PHONE","surface":"۰۹۱۲۴۸۶۵۴۴۱","language":"fa","field":null},{"start":66,"end":87,"entity_type":"EMAIL","surface":"user.03@aria-lab.test","language":"en","field":null}],"gold_meta":[],"hard_negatives":["INV-1403-0003"],"description":"Persian HR note with ZWNJ, Persian digits, national id, and phone.","description_fa":"نامه منابع انسانی با نیم‌فاصله، رقم فارسی، کد ملی و تلفن."}
14
+ {"doc_id":"DLP-FA-HR-04","template":"fa_hr_letter","family":"fa_hr_letter","split":"development","language":"fa","format":"text","title":"Persian HR letter","body":"خانم امید نمونه‌ای،\nکد‌ملی ۴۶۱۷۵۵۴۲۹۷\nتلفن همراه ۰۹۱۲۸۲۹۱۰۰۲\nایمیل user.04@aria-lab.test\nشماره پیگیری INV-1403-0004 و نسخه ۳٫۲٫۱ فقط ارجاع داخلی‌اند.\n","metadata":{},"gold_spans":[{"start":5,"end":18,"entity_type":"PERSON","surface":"امید نمونه‌ای","language":"fa","field":null},{"start":27,"end":37,"entity_type":"IR_NATIONAL_ID","surface":"۴۶۱۷۵۵۴۲۹۷","language":"fa","field":null},{"start":49,"end":60,"entity_type":"PHONE","surface":"۰۹۱۲۸۲۹۱۰۰۲","language":"fa","field":null},{"start":67,"end":88,"entity_type":"EMAIL","surface":"user.04@aria-lab.test","language":"en","field":null}],"gold_meta":[],"hard_negatives":["INV-1403-0004"],"description":"Persian HR note with ZWNJ, Persian digits, national id, and phone.","description_fa":"نامه منابع انسانی با نیم‌فاصله، رقم فارسی، کد ملی و تلفن."}
15
+ {"doc_id":"DLP-FA-HR-05","template":"fa_hr_letter","family":"fa_hr_letter","split":"development","language":"fa","format":"text","title":"Persian HR letter","body":"خانم لیلا نمونه‌ای،\nکد‌ملی ۵۰۰۷۵۷۷۱۰۳\nتلفن همراه ۰۹۱۲۲۳۶۷۲۱۴\nایمیل user.05@aria-lab.test\nشماره پیگیری INV-1403-0005 و نسخه ۳٫۲٫۱ فقط ارجاع داخلی‌اند.\n","metadata":{},"gold_spans":[{"start":5,"end":18,"entity_type":"PERSON","surface":"لیلا نمونه‌ای","language":"fa","field":null},{"start":27,"end":37,"entity_type":"IR_NATIONAL_ID","surface":"۵۰۰۷۵۷۷۱۰۳","language":"fa","field":null},{"start":49,"end":60,"entity_type":"PHONE","surface":"۰۹۱۲۲۳۶۷۲۱۴","language":"fa","field":null},{"start":67,"end":88,"entity_type":"EMAIL","surface":"user.05@aria-lab.test","language":"en","field":null}],"gold_meta":[],"hard_negatives":["INV-1403-0005"],"description":"Persian HR note with ZWNJ, Persian digits, national id, and phone.","description_fa":"نامه منابع انسانی با نیم‌فاصله، رقم فارسی، کد ملی و تلفن."}
16
+ {"doc_id":"DLP-FA-HR-06","template":"fa_hr_letter","family":"fa_hr_letter","split":"development","language":"fa","format":"text","title":"Persian HR letter","body":"خانم لیلا آزمایشی،\nکد‌ملی ۶۸۷۹۸۶۶۳۴۴\nتلفن همراه ۰۹۱۲۶۰۸۳۲۸۳\nایمیل user.06@aria-lab.test\nشماره پیگیری INV-1403-0006 و نسخه ۳٫۲٫۱ فقط ارجاع داخلی‌اند.\n","metadata":{},"gold_spans":[{"start":5,"end":17,"entity_type":"PERSON","surface":"لیلا آزمایشی","language":"fa","field":null},{"start":26,"end":36,"entity_type":"IR_NATIONAL_ID","surface":"۶۸۷۹۸۶۶۳۴۴","language":"fa","field":null},{"start":48,"end":59,"entity_type":"PHONE","surface":"۰۹۱۲۶۰۸۳۲۸۳","language":"fa","field":null},{"start":66,"end":87,"entity_type":"EMAIL","surface":"user.06@aria-lab.test","language":"en","field":null}],"gold_meta":[],"hard_negatives":["INV-1403-0006"],"description":"Persian HR note with ZWNJ, Persian digits, national id, and phone.","description_fa":"نامه منابع انسانی با نیم‌فاصله، رقم فارسی، کد ملی و تلفن."}
17
+ {"doc_id":"DLP-FA-HR-07","template":"fa_hr_letter","family":"fa_hr_letter","split":"development","language":"fa","format":"text","title":"Persian HR letter","body":"خانم لیلا آزمایشی،\nکد‌ملی ۲۵۱۹۰۴۸۰۶۹\nتلفن همراه ۰۹۱۲۷۷۱۴۶۰۱\nایمیل user.07@aria-lab.test\nشماره پیگیری INV-1403-0007 و نسخه ۳٫۲٫۱ فقط ارجاع داخلی‌اند.\n","metadata":{},"gold_spans":[{"start":5,"end":17,"entity_type":"PERSON","surface":"لیلا آزمایشی","language":"fa","field":null},{"start":26,"end":36,"entity_type":"IR_NATIONAL_ID","surface":"۲۵۱۹۰۴۸۰۶۹","language":"fa","field":null},{"start":48,"end":59,"entity_type":"PHONE","surface":"۰۹۱۲۷۷۱۴۶۰۱","language":"fa","field":null},{"start":66,"end":87,"entity_type":"EMAIL","surface":"user.07@aria-lab.test","language":"en","field":null}],"gold_meta":[],"hard_negatives":["INV-1403-0007"],"description":"Persian HR note with ZWNJ, Persian digits, national id, and phone.","description_fa":"نامه منابع انسانی با نیم‌فاصله، رقم فارسی، کد ملی و تلفن."}
18
+ {"doc_id":"DLP-FA-HR-08","template":"fa_hr_letter","family":"fa_hr_letter","split":"development","language":"fa","format":"text","title":"Persian HR letter","body":"خانم امید آزمایشی،\nکد‌ملی ۴۷۶۵۵۰۶۰۷۱\nتلفن همراه ۰۹۱۲۸۷۶۰۹۰۱\nایمیل user.08@aria-lab.test\nشماره پیگیری INV-1403-0008 و نسخه ۳٫۲٫۱ فقط ارجاع داخلی‌اند.\n","metadata":{},"gold_spans":[{"start":5,"end":17,"entity_type":"PERSON","surface":"امید آزمایشی","language":"fa","field":null},{"start":26,"end":36,"entity_type":"IR_NATIONAL_ID","surface":"۴۷۶۵۵۰۶۰۷۱","language":"fa","field":null},{"start":48,"end":59,"entity_type":"PHONE","surface":"۰۹۱۲۸۷۶۰۹۰۱","language":"fa","field":null},{"start":66,"end":87,"entity_type":"EMAIL","surface":"user.08@aria-lab.test","language":"en","field":null}],"gold_meta":[],"hard_negatives":["INV-1403-0008"],"description":"Persian HR note with ZWNJ, Persian digits, national id, and phone.","description_fa":"نامه منابع انسانی با نیم‌فاصله، رقم فارسی، کد ملی و تلفن."}
19
+ {"doc_id":"DLP-FA-HR-09","template":"fa_hr_letter","family":"fa_hr_letter","split":"development","language":"fa","format":"text","title":"Persian HR letter","body":"خانم سارا نمونه‌ای،\nکد‌ملی ۳۲۸۹۳۷۲۲۰۰\nتلفن همراه ۰۹۱۲۲۸۹۳۲۸۹\nایمیل user.09@aria-lab.test\nشماره پیگیری INV-1403-0009 و نسخه ۳٫۲٫۱ فقط ارجاع داخلی‌اند.\n","metadata":{},"gold_spans":[{"start":5,"end":18,"entity_type":"PERSON","surface":"سارا نمونه‌ای","language":"fa","field":null},{"start":27,"end":37,"entity_type":"IR_NATIONAL_ID","surface":"۳۲۸۹۳۷۲۲۰۰","language":"fa","field":null},{"start":49,"end":60,"entity_type":"PHONE","surface":"۰۹۱۲۲۸۹۳۲۸۹","language":"fa","field":null},{"start":67,"end":88,"entity_type":"EMAIL","surface":"user.09@aria-lab.test","language":"en","field":null}],"gold_meta":[],"hard_negatives":["INV-1403-0009"],"description":"Persian HR note with ZWNJ, Persian digits, national id, and phone.","description_fa":"نامه منابع انسانی با نیم‌فاصله، رقم فارسی، کد ملی و تلفن."}
20
+ {"doc_id":"DLP-FA-HR-10","template":"fa_hr_letter","family":"fa_hr_letter","split":"development","language":"fa","format":"text","title":"Persian HR letter","body":"خانم سارا نمونه‌ای،\nکد‌ملی ۳۰۸۷۰۲۵۴۱۰\nتلفن همراه ۰۹۱۲۴۰۶۷۵۴۸\nایمیل user.10@aria-lab.test\nشماره پیگیری INV-1403-0010 و نسخه ۳٫۲٫۱ فقط ارجاع داخلی‌اند.\n","metadata":{},"gold_spans":[{"start":5,"end":18,"entity_type":"PERSON","surface":"سارا نمونه‌ای","language":"fa","field":null},{"start":27,"end":37,"entity_type":"IR_NATIONAL_ID","surface":"۳۰۸۷۰۲۵۴۱۰","language":"fa","field":null},{"start":49,"end":60,"entity_type":"PHONE","surface":"۰۹۱۲۴۰۶۷۵۴۸","language":"fa","field":null},{"start":67,"end":88,"entity_type":"EMAIL","surface":"user.10@aria-lab.test","language":"en","field":null}],"gold_meta":[],"hard_negatives":["INV-1403-0010"],"description":"Persian HR note with ZWNJ, Persian digits, national id, and phone.","description_fa":"نامه منابع انسانی با نیم‌فاصله، رقم فارسی، کد ملی و تلفن."}
21
+ {"doc_id":"DLP-CSV-INV-01","template":"mixed_invoice_csv","family":"mixed_invoice_csv","split":"development","language":"mixed","format":"csv","title":"Invoice CSV","body":"invoice_id,customer,email,phone,lookup\nINV-2024-1001,Ava Labian,Ava.Labian01@aria-lab.test,09129456762,1987844778\n","metadata":{},"gold_spans":[{"start":53,"end":63,"entity_type":"PERSON","surface":"Ava Labian","language":"en","field":null},{"start":64,"end":90,"entity_type":"EMAIL","surface":"Ava.Labian01@aria-lab.test","language":"en","field":null},{"start":91,"end":102,"entity_type":"PHONE","surface":"09129456762","language":"en","field":null}],"gold_meta":[],"hard_negatives":["INV-2024-1001","1987844778"],"description":"CSV with a customer row and a checksum-failing lookup number.","description_fa":"سی‌اس‌وی مشتری با شماره جست‌وجوی بدون checksum معتبر."}
22
+ {"doc_id":"DLP-CSV-INV-02","template":"mixed_invoice_csv","family":"mixed_invoice_csv","split":"development","language":"mixed","format":"csv","title":"Invoice CSV","body":"invoice_id,customer,email,phone,lookup\nINV-2024-1002,Nima Fixture,Nima.Fixture02@aria-lab.test,09127891210,2513822278\n","metadata":{},"gold_spans":[{"start":53,"end":65,"entity_type":"PERSON","surface":"Nima Fixture","language":"en","field":null},{"start":66,"end":94,"entity_type":"EMAIL","surface":"Nima.Fixture02@aria-lab.test","language":"en","field":null},{"start":95,"end":106,"entity_type":"PHONE","surface":"09127891210","language":"en","field":null}],"gold_meta":[],"hard_negatives":["INV-2024-1002","2513822278"],"description":"CSV with a customer row and a checksum-failing lookup number.","description_fa":"سی‌اس‌وی مشتری با شماره جست‌وجوی بدون checksum معتبر."}
23
+ {"doc_id":"DLP-CSV-INV-03","template":"mixed_invoice_csv","family":"mixed_invoice_csv","split":"development","language":"mixed","format":"csv","title":"Invoice CSV","body":"invoice_id,customer,email,phone,lookup\nINV-2024-1003,Leila Fixture,Leila.Fixture03@aria-lab.test,09124895575,9231499344\n","metadata":{},"gold_spans":[{"start":53,"end":66,"entity_type":"PERSON","surface":"Leila Fixture","language":"en","field":null},{"start":67,"end":96,"entity_type":"EMAIL","surface":"Leila.Fixture03@aria-lab.test","language":"en","field":null},{"start":97,"end":108,"entity_type":"PHONE","surface":"09124895575","language":"en","field":null}],"gold_meta":[],"hard_negatives":["INV-2024-1003","9231499344"],"description":"CSV with a customer row and a checksum-failing lookup number.","description_fa":"سی‌اس‌وی مشتری با شماره جست‌وجوی بدون checksum معتبر."}
24
+ {"doc_id":"DLP-CSV-INV-04","template":"mixed_invoice_csv","family":"mixed_invoice_csv","split":"development","language":"mixed","format":"csv","title":"Invoice CSV","body":"invoice_id,customer,email,phone,lookup\nINV-2024-1004,Omar Labian,Omar.Labian04@aria-lab.test,09124151655,3699795722\n","metadata":{},"gold_spans":[{"start":53,"end":64,"entity_type":"PERSON","surface":"Omar Labian","language":"en","field":null},{"start":65,"end":92,"entity_type":"EMAIL","surface":"Omar.Labian04@aria-lab.test","language":"en","field":null},{"start":93,"end":104,"entity_type":"PHONE","surface":"09124151655","language":"en","field":null}],"gold_meta":[],"hard_negatives":["INV-2024-1004","3699795722"],"description":"CSV with a customer row and a checksum-failing lookup number.","description_fa":"سی‌اس‌وی مشتری با شماره جست‌وجوی بدون checksum معتبر."}
25
+ {"doc_id":"DLP-CSV-INV-05","template":"mixed_invoice_csv","family":"mixed_invoice_csv","split":"development","language":"mixed","format":"csv","title":"Invoice CSV","body":"invoice_id,customer,email,phone,lookup\nINV-2024-1005,Ava Labian,Ava.Labian05@aria-lab.test,09123831675,3303447994\n","metadata":{},"gold_spans":[{"start":53,"end":63,"entity_type":"PERSON","surface":"Ava Labian","language":"en","field":null},{"start":64,"end":90,"entity_type":"EMAIL","surface":"Ava.Labian05@aria-lab.test","language":"en","field":null},{"start":91,"end":102,"entity_type":"PHONE","surface":"09123831675","language":"en","field":null}],"gold_meta":[],"hard_negatives":["INV-2024-1005","3303447994"],"description":"CSV with a customer row and a checksum-failing lookup number.","description_fa":"سی‌اس‌وی مشتری با شماره جست‌وجوی بدون checksum معتبر."}
26
+ {"doc_id":"DLP-CSV-INV-06","template":"mixed_invoice_csv","family":"mixed_invoice_csv","split":"development","language":"mixed","format":"csv","title":"Invoice CSV","body":"invoice_id,customer,email,phone,lookup\nINV-2024-1006,Noah Fixture,Noah.Fixture06@aria-lab.test,09125428133,3478403597\n","metadata":{},"gold_spans":[{"start":53,"end":65,"entity_type":"PERSON","surface":"Noah Fixture","language":"en","field":null},{"start":66,"end":94,"entity_type":"EMAIL","surface":"Noah.Fixture06@aria-lab.test","language":"en","field":null},{"start":95,"end":106,"entity_type":"PHONE","surface":"09125428133","language":"en","field":null}],"gold_meta":[],"hard_negatives":["INV-2024-1006","3478403597"],"description":"CSV with a customer row and a checksum-failing lookup number.","description_fa":"سی‌اس‌وی مشتری با شماره جست‌وجوی بدون checksum معتبر."}
27
+ {"doc_id":"DLP-CSV-INV-07","template":"mixed_invoice_csv","family":"mixed_invoice_csv","split":"development","language":"mixed","format":"csv","title":"Invoice CSV","body":"invoice_id,customer,email,phone,lookup\nINV-2024-1007,Helen Fixture,Helen.Fixture07@aria-lab.test,09124672166,5386511573\n","metadata":{},"gold_spans":[{"start":53,"end":66,"entity_type":"PERSON","surface":"Helen Fixture","language":"en","field":null},{"start":67,"end":96,"entity_type":"EMAIL","surface":"Helen.Fixture07@aria-lab.test","language":"en","field":null},{"start":97,"end":108,"entity_type":"PHONE","surface":"09124672166","language":"en","field":null}],"gold_meta":[],"hard_negatives":["INV-2024-1007","5386511573"],"description":"CSV with a customer row and a checksum-failing lookup number.","description_fa":"سی‌اس‌وی مشتری با شماره جست‌وجوی بدون checksum معتبر."}
28
+ {"doc_id":"DLP-CSV-INV-08","template":"mixed_invoice_csv","family":"mixed_invoice_csv","split":"development","language":"mixed","format":"csv","title":"Invoice CSV","body":"invoice_id,customer,email,phone,lookup\nINV-2024-1008,Omar Fixture,Omar.Fixture08@aria-lab.test,09128527564,7963012808\n","metadata":{},"gold_spans":[{"start":53,"end":65,"entity_type":"PERSON","surface":"Omar Fixture","language":"en","field":null},{"start":66,"end":94,"entity_type":"EMAIL","surface":"Omar.Fixture08@aria-lab.test","language":"en","field":null},{"start":95,"end":106,"entity_type":"PHONE","surface":"09128527564","language":"en","field":null}],"gold_meta":[],"hard_negatives":["INV-2024-1008","7963012808"],"description":"CSV with a customer row and a checksum-failing lookup number.","description_fa":"سی‌اس‌وی مشتری با شماره جست‌وجوی بدون checksum معتبر."}
29
+ {"doc_id":"DLP-CSV-INV-09","template":"mixed_invoice_csv","family":"mixed_invoice_csv","split":"development","language":"mixed","format":"csv","title":"Invoice CSV","body":"invoice_id,customer,email,phone,lookup\nINV-2024-1009,Leila Labian,Leila.Labian09@aria-lab.test,09127031590,3935218380\n","metadata":{},"gold_spans":[{"start":53,"end":65,"entity_type":"PERSON","surface":"Leila Labian","language":"en","field":null},{"start":66,"end":94,"entity_type":"EMAIL","surface":"Leila.Labian09@aria-lab.test","language":"en","field":null},{"start":95,"end":106,"entity_type":"PHONE","surface":"09127031590","language":"en","field":null}],"gold_meta":[],"hard_negatives":["INV-2024-1009","3935218380"],"description":"CSV with a customer row and a checksum-failing lookup number.","description_fa":"سی‌اس‌وی مشتری با شماره جست‌وجوی بدون checksum معتبر."}
30
+ {"doc_id":"DLP-CSV-INV-10","template":"mixed_invoice_csv","family":"mixed_invoice_csv","split":"development","language":"mixed","format":"csv","title":"Invoice CSV","body":"invoice_id,customer,email,phone,lookup\nINV-2024-1010,Leila Fixture,Leila.Fixture10@aria-lab.test,09123631859,4874249265\n","metadata":{},"gold_spans":[{"start":53,"end":66,"entity_type":"PERSON","surface":"Leila Fixture","language":"en","field":null},{"start":67,"end":96,"entity_type":"EMAIL","surface":"Leila.Fixture10@aria-lab.test","language":"en","field":null},{"start":97,"end":108,"entity_type":"PHONE","surface":"09123631859","language":"en","field":null}],"gold_meta":[],"hard_negatives":["INV-2024-1010","4874249265"],"description":"CSV with a customer row and a checksum-failing lookup number.","description_fa":"سی‌اس‌وی مشتری با شماره جست‌وجوی بدون checksum معتبر."}
31
+ {"doc_id":"DLP-EN-ONB-01","template":"en_onboarding","family":"en_onboarding","split":"development","language":"en","format":"text","title":"Onboarding worksheet","body":"Name: Helen Labian\nWork email: Helen.Labian01@aria-lab.test\nTravel advance test PAN: 4111137038772834\nPayroll IBAN: IR308660110372211081968590\nISBN 978-3-16-148410-0 is a catalog example, not an account.\n","metadata":{},"gold_spans":[{"start":6,"end":18,"entity_type":"PERSON","surface":"Helen Labian","language":"en","field":null},{"start":31,"end":59,"entity_type":"EMAIL","surface":"Helen.Labian01@aria-lab.test","language":"en","field":null},{"start":85,"end":101,"entity_type":"CREDIT_CARD","surface":"4111137038772834","language":"en","field":null},{"start":116,"end":142,"entity_type":"IBAN","surface":"IR308660110372211081968590","language":"en","field":null}],"gold_meta":[],"hard_negatives":["978-3-16-148410-0","9783161484100"],"description":"Onboarding sheet with test PAN and IBAN plus an ISBN hard negative.","description_fa":"برگه ورود با کارت و شبا آزمایشی و ISBN به‌عنوان منفی سخت."}
32
+ {"doc_id":"DLP-EN-ONB-02","template":"en_onboarding","family":"en_onboarding","split":"development","language":"en","format":"text","title":"Onboarding worksheet","body":"Name: Helen Labian\nWork email: Helen.Labian02@aria-lab.test\nTravel advance test PAN: 4111529393529637\nPayroll IBAN: IR044280439609406102122129\nISBN 978-3-16-148410-0 is a catalog example, not an account.\n","metadata":{},"gold_spans":[{"start":6,"end":18,"entity_type":"PERSON","surface":"Helen Labian","language":"en","field":null},{"start":31,"end":59,"entity_type":"EMAIL","surface":"Helen.Labian02@aria-lab.test","language":"en","field":null},{"start":85,"end":101,"entity_type":"CREDIT_CARD","surface":"4111529393529637","language":"en","field":null},{"start":116,"end":142,"entity_type":"IBAN","surface":"IR044280439609406102122129","language":"en","field":null}],"gold_meta":[],"hard_negatives":["978-3-16-148410-0","9783161484100"],"description":"Onboarding sheet with test PAN and IBAN plus an ISBN hard negative.","description_fa":"برگه ورود با کارت و شبا آزمایشی و ISBN به‌عنوان منفی سخت."}
33
+ {"doc_id":"DLP-EN-ONB-03","template":"en_onboarding","family":"en_onboarding","split":"development","language":"en","format":"text","title":"Onboarding worksheet","body":"Name: Helen Labian\nWork email: Helen.Labian03@aria-lab.test\nTravel advance test PAN: 4111785437328044\nPayroll IBAN: IR019630083854359765553642\nISBN 978-3-16-148410-0 is a catalog example, not an account.\n","metadata":{},"gold_spans":[{"start":6,"end":18,"entity_type":"PERSON","surface":"Helen Labian","language":"en","field":null},{"start":31,"end":59,"entity_type":"EMAIL","surface":"Helen.Labian03@aria-lab.test","language":"en","field":null},{"start":85,"end":101,"entity_type":"CREDIT_CARD","surface":"4111785437328044","language":"en","field":null},{"start":116,"end":142,"entity_type":"IBAN","surface":"IR019630083854359765553642","language":"en","field":null}],"gold_meta":[],"hard_negatives":["978-3-16-148410-0","9783161484100"],"description":"Onboarding sheet with test PAN and IBAN plus an ISBN hard negative.","description_fa":"برگه ورود با کارت و شبا آزمایشی و ISBN به‌عنوان منفی سخت."}
34
+ {"doc_id":"DLP-EN-ONB-04","template":"en_onboarding","family":"en_onboarding","split":"development","language":"en","format":"text","title":"Onboarding worksheet","body":"Name: Helen Labian\nWork email: Helen.Labian04@aria-lab.test\nTravel advance test PAN: 4111759320597908\nPayroll IBAN: IR472630756281276293569333\nISBN 978-3-16-148410-0 is a catalog example, not an account.\n","metadata":{},"gold_spans":[{"start":6,"end":18,"entity_type":"PERSON","surface":"Helen Labian","language":"en","field":null},{"start":31,"end":59,"entity_type":"EMAIL","surface":"Helen.Labian04@aria-lab.test","language":"en","field":null},{"start":85,"end":101,"entity_type":"CREDIT_CARD","surface":"4111759320597908","language":"en","field":null},{"start":116,"end":142,"entity_type":"IBAN","surface":"IR472630756281276293569333","language":"en","field":null}],"gold_meta":[],"hard_negatives":["978-3-16-148410-0","9783161484100"],"description":"Onboarding sheet with test PAN and IBAN plus an ISBN hard negative.","description_fa":"برگه ورود با کارت و شبا آزمایشی و ISBN به‌عنوان منفی سخت."}
35
+ {"doc_id":"DLP-EN-ONB-05","template":"en_onboarding","family":"en_onboarding","split":"development","language":"en","format":"text","title":"Onboarding worksheet","body":"Name: Leila Fixture\nWork email: Leila.Fixture05@aria-lab.test\nTravel advance test PAN: 4111313694373346\nPayroll IBAN: IR466530650673099236853917\nISBN 978-3-16-148410-0 is a catalog example, not an account.\n","metadata":{},"gold_spans":[{"start":6,"end":19,"entity_type":"PERSON","surface":"Leila Fixture","language":"en","field":null},{"start":32,"end":61,"entity_type":"EMAIL","surface":"Leila.Fixture05@aria-lab.test","language":"en","field":null},{"start":87,"end":103,"entity_type":"CREDIT_CARD","surface":"4111313694373346","language":"en","field":null},{"start":118,"end":144,"entity_type":"IBAN","surface":"IR466530650673099236853917","language":"en","field":null}],"gold_meta":[],"hard_negatives":["978-3-16-148410-0","9783161484100"],"description":"Onboarding sheet with test PAN and IBAN plus an ISBN hard negative.","description_fa":"برگه ورود با کارت و شبا آزمایشی و ISBN به‌عنوان منفی سخت."}
36
+ {"doc_id":"DLP-EN-ONB-06","template":"en_onboarding","family":"en_onboarding","split":"development","language":"en","format":"text","title":"Onboarding worksheet","body":"Name: Nima Labian\nWork email: Nima.Labian06@aria-lab.test\nTravel advance test PAN: 4111814102596287\nPayroll IBAN: IR581060722681444648227055\nISBN 978-3-16-148410-0 is a catalog example, not an account.\n","metadata":{},"gold_spans":[{"start":6,"end":17,"entity_type":"PERSON","surface":"Nima Labian","language":"en","field":null},{"start":30,"end":57,"entity_type":"EMAIL","surface":"Nima.Labian06@aria-lab.test","language":"en","field":null},{"start":83,"end":99,"entity_type":"CREDIT_CARD","surface":"4111814102596287","language":"en","field":null},{"start":114,"end":140,"entity_type":"IBAN","surface":"IR581060722681444648227055","language":"en","field":null}],"gold_meta":[],"hard_negatives":["978-3-16-148410-0","9783161484100"],"description":"Onboarding sheet with test PAN and IBAN plus an ISBN hard negative.","description_fa":"برگه ورود با کارت و شبا آزمایشی و ISBN به‌عنوان منفی سخت."}
37
+ {"doc_id":"DLP-EN-ONB-07","template":"en_onboarding","family":"en_onboarding","split":"development","language":"en","format":"text","title":"Onboarding worksheet","body":"Name: Noah Fixture\nWork email: Noah.Fixture07@aria-lab.test\nTravel advance test PAN: 4111842413164297\nPayroll IBAN: IR895120978818064078339048\nISBN 978-3-16-148410-0 is a catalog example, not an account.\n","metadata":{},"gold_spans":[{"start":6,"end":18,"entity_type":"PERSON","surface":"Noah Fixture","language":"en","field":null},{"start":31,"end":59,"entity_type":"EMAIL","surface":"Noah.Fixture07@aria-lab.test","language":"en","field":null},{"start":85,"end":101,"entity_type":"CREDIT_CARD","surface":"4111842413164297","language":"en","field":null},{"start":116,"end":142,"entity_type":"IBAN","surface":"IR895120978818064078339048","language":"en","field":null}],"gold_meta":[],"hard_negatives":["978-3-16-148410-0","9783161484100"],"description":"Onboarding sheet with test PAN and IBAN plus an ISBN hard negative.","description_fa":"برگه ورود با کارت و شبا آزمایشی و ISBN به‌عنوان منفی سخت."}
38
+ {"doc_id":"DLP-EN-ONB-08","template":"en_onboarding","family":"en_onboarding","split":"development","language":"en","format":"text","title":"Onboarding worksheet","body":"Name: Noah Fixture\nWork email: Noah.Fixture08@aria-lab.test\nTravel advance test PAN: 4111329710194664\nPayroll IBAN: IR237680114663394226281259\nISBN 978-3-16-148410-0 is a catalog example, not an account.\n","metadata":{},"gold_spans":[{"start":6,"end":18,"entity_type":"PERSON","surface":"Noah Fixture","language":"en","field":null},{"start":31,"end":59,"entity_type":"EMAIL","surface":"Noah.Fixture08@aria-lab.test","language":"en","field":null},{"start":85,"end":101,"entity_type":"CREDIT_CARD","surface":"4111329710194664","language":"en","field":null},{"start":116,"end":142,"entity_type":"IBAN","surface":"IR237680114663394226281259","language":"en","field":null}],"gold_meta":[],"hard_negatives":["978-3-16-148410-0","9783161484100"],"description":"Onboarding sheet with test PAN and IBAN plus an ISBN hard negative.","description_fa":"برگه ورود با کارت و شبا آزمایشی و ISBN به‌عنوان منفی سخت."}
39
+ {"doc_id":"DLP-EN-ONB-09","template":"en_onboarding","family":"en_onboarding","split":"development","language":"en","format":"text","title":"Onboarding worksheet","body":"Name: Leila Fixture\nWork email: Leila.Fixture09@aria-lab.test\nTravel advance test PAN: 4111996662241709\nPayroll IBAN: IR772810906854799104860878\nISBN 978-3-16-148410-0 is a catalog example, not an account.\n","metadata":{},"gold_spans":[{"start":6,"end":19,"entity_type":"PERSON","surface":"Leila Fixture","language":"en","field":null},{"start":32,"end":61,"entity_type":"EMAIL","surface":"Leila.Fixture09@aria-lab.test","language":"en","field":null},{"start":87,"end":103,"entity_type":"CREDIT_CARD","surface":"4111996662241709","language":"en","field":null},{"start":118,"end":144,"entity_type":"IBAN","surface":"IR772810906854799104860878","language":"en","field":null}],"gold_meta":[],"hard_negatives":["978-3-16-148410-0","9783161484100"],"description":"Onboarding sheet with test PAN and IBAN plus an ISBN hard negative.","description_fa":"برگه ورود با کارت و شبا آزمایشی و ISBN به‌عنوان منفی سخت."}
40
+ {"doc_id":"DLP-EN-ONB-10","template":"en_onboarding","family":"en_onboarding","split":"development","language":"en","format":"text","title":"Onboarding worksheet","body":"Name: Nima Fixture\nWork email: Nima.Fixture10@aria-lab.test\nTravel advance test PAN: 4111041294625191\nPayroll IBAN: IR338550745309562340114615\nISBN 978-3-16-148410-0 is a catalog example, not an account.\n","metadata":{},"gold_spans":[{"start":6,"end":18,"entity_type":"PERSON","surface":"Nima Fixture","language":"en","field":null},{"start":31,"end":59,"entity_type":"EMAIL","surface":"Nima.Fixture10@aria-lab.test","language":"en","field":null},{"start":85,"end":101,"entity_type":"CREDIT_CARD","surface":"4111041294625191","language":"en","field":null},{"start":116,"end":142,"entity_type":"IBAN","surface":"IR338550745309562340114615","language":"en","field":null}],"gold_meta":[],"hard_negatives":["978-3-16-148410-0","9783161484100"],"description":"Onboarding sheet with test PAN and IBAN plus an ISBN hard negative.","description_fa":"برگه ورود با کارت و شبا آزمایشی و ISBN به‌عنوان منفی سخت."}
41
+ {"doc_id":"DLP-FA-CTR-01","template":"fa_contract_notice","family":"fa_contract_notice","split":"test","language":"fa","format":"text","title":"Persian contract notice","body":"موضوع: اخطار قراردادی\nنام: لیلا آزمایشی\nکد ملی 4985528323\nتلفن ۰۹۱۲۳۴۲۱۵۱۴\nکد ملی 3666916466 در سامانه پذیرفته نشد و برای بازبینی مانده است.\nتاریخ ۱۴۰۳/۰۵/۱۲ و نسخه ۱.۰.۰ ارجاع سند هستند.\n","metadata":{},"gold_spans":[{"start":27,"end":39,"entity_type":"PERSON","surface":"لیلا آزمایشی","language":"fa","field":null},{"start":47,"end":57,"entity_type":"IR_NATIONAL_ID","surface":"4985528323","language":"fa","field":null},{"start":63,"end":74,"entity_type":"PHONE","surface":"۰۹۱۲۳۴۲۱۵۱۴","language":"fa","field":null}],"gold_meta":[],"hard_negatives":["3666916466","1403/05/12","۱۴۰۳/۰۵/۱۲","1.0.0"],"description":"Persian contract with a failed national-id left for review.","description_fa":"اخطار قراردادی با کد ملی نامعتبر برای بازبینی."}
42
+ {"doc_id":"DLP-FA-CTR-02","template":"fa_contract_notice","family":"fa_contract_notice","split":"test","language":"fa","format":"text","title":"Persian contract notice","body":"موضوع: اخطار قراردادی\nنام: سارا آزمایشی\nکد ملی 6361423751\nتلفن ۰۹۱۲۳۶۳۰۹۶۹\nکد ملی 9032951212 در سامانه پذیرفته نشد و برای بازبینی مانده است.\nتاریخ ۱۴۰۳/۰۵/۱۲ و نسخه ۱.۰.۰ ارجاع سند هستند.\n","metadata":{},"gold_spans":[{"start":27,"end":39,"entity_type":"PERSON","surface":"سارا آزمایشی","language":"fa","field":null},{"start":47,"end":57,"entity_type":"IR_NATIONAL_ID","surface":"6361423751","language":"fa","field":null},{"start":63,"end":74,"entity_type":"PHONE","surface":"۰۹۱۲۳۶۳۰۹۶۹","language":"fa","field":null}],"gold_meta":[],"hard_negatives":["9032951212","1403/05/12","۱۴۰۳/۰۵/۱۲","1.0.0"],"description":"Persian contract with a failed national-id left for review.","description_fa":"اخطار قراردادی با کد ملی نامعتبر برای بازبینی."}
43
+ {"doc_id":"DLP-FA-CTR-03","template":"fa_contract_notice","family":"fa_contract_notice","split":"test","language":"fa","format":"text","title":"Persian contract notice","body":"موضوع: اخطار قراردادی\nنام: نیما نمونه‌ای\nکد ملی 3589737379\nتلفن ۰۹۱۲۵۹۲۸۹۵۰\nکد ملی 9386005221 در سامانه پذیرفته نشد و برای بازبینی مانده است.\nتاریخ ۱۴۰۳/۰۵/۱۲ و نسخه ۱.۰.۰ ارجاع سند هستند.\n","metadata":{},"gold_spans":[{"start":27,"end":40,"entity_type":"PERSON","surface":"نیما نمونه‌ای","language":"fa","field":null},{"start":48,"end":58,"entity_type":"IR_NATIONAL_ID","surface":"3589737379","language":"fa","field":null},{"start":64,"end":75,"entity_type":"PHONE","surface":"۰۹۱۲۵۹۲۸۹۵۰","language":"fa","field":null}],"gold_meta":[],"hard_negatives":["9386005221","1403/05/12","۱۴۰۳/۰۵/۱۲","1.0.0"],"description":"Persian contract with a failed national-id left for review.","description_fa":"اخطار قراردادی با کد ملی نامعتبر برای بازبینی."}
44
+ {"doc_id":"DLP-FA-CTR-04","template":"fa_contract_notice","family":"fa_contract_notice","split":"test","language":"fa","format":"text","title":"Persian contract notice","body":"موضوع: اخطار قراردادی\nنام: نیما آزمایشی\nکد ملی 2341457452\nتلفن ۰۹۱۲۲۳۶۵۹۰۷\nکد ملی 2506786750 در سامانه پذیرفته نشد و برای بازبینی مانده است.\nتار��خ ۱۴۰۳/۰۵/۱۲ و نسخه ۱.۰.۰ ارجاع سند هستند.\n","metadata":{},"gold_spans":[{"start":27,"end":39,"entity_type":"PERSON","surface":"نیما آزمایشی","language":"fa","field":null},{"start":47,"end":57,"entity_type":"IR_NATIONAL_ID","surface":"2341457452","language":"fa","field":null},{"start":63,"end":74,"entity_type":"PHONE","surface":"۰۹۱۲۲۳۶۵۹۰۷","language":"fa","field":null}],"gold_meta":[],"hard_negatives":["2506786750","1403/05/12","۱۴۰۳/۰۵/۱۲","1.0.0"],"description":"Persian contract with a failed national-id left for review.","description_fa":"اخطار قراردادی با کد ملی نامعتبر برای بازبینی."}
45
+ {"doc_id":"DLP-FA-CTR-05","template":"fa_contract_notice","family":"fa_contract_notice","split":"test","language":"fa","format":"text","title":"Persian contract notice","body":"موضوع: اخطار قراردادی\nنام: نیما نمونه‌ای\nکد ملی 4061666126\nتلفن ۰۹۱۲۴۹۶۶۴۱۱\nکد ملی 9044212126 در سامانه پذیرفته نشد و برای بازبینی مانده است.\nتاریخ ۱۴۰۳/۰۵/۱۲ و نسخه ۱.۰.۰ ارجاع سند هستند.\n","metadata":{},"gold_spans":[{"start":27,"end":40,"entity_type":"PERSON","surface":"نیما نمونه‌ای","language":"fa","field":null},{"start":48,"end":58,"entity_type":"IR_NATIONAL_ID","surface":"4061666126","language":"fa","field":null},{"start":64,"end":75,"entity_type":"PHONE","surface":"۰۹۱۲۴۹۶۶۴۱۱","language":"fa","field":null}],"gold_meta":[],"hard_negatives":["9044212126","1403/05/12","۱۴۰۳/۰۵/۱۲","1.0.0"],"description":"Persian contract with a failed national-id left for review.","description_fa":"اخطار قراردادی با کد ملی نامعتبر برای بازبینی."}
46
+ {"doc_id":"DLP-FA-CTR-06","template":"fa_contract_notice","family":"fa_contract_notice","split":"test","language":"fa","format":"text","title":"Persian contract notice","body":"موضوع: اخطار قراردادی\nنام: لیلا نمونه‌ای\nکد ملی 3179136243\nتلفن ۰۹۱۲۸۱۳۵۴۲۶\nکد ملی 4511025429 در سامانه پذیرفته نشد و برای بازبینی مانده است.\nتاریخ ۱۴۰۳/۰۵/۱۲ و نسخه ۱.۰.۰ ارجاع سند هستند.\n","metadata":{},"gold_spans":[{"start":27,"end":40,"entity_type":"PERSON","surface":"لیلا نمونه‌ای","language":"fa","field":null},{"start":48,"end":58,"entity_type":"IR_NATIONAL_ID","surface":"3179136243","language":"fa","field":null},{"start":64,"end":75,"entity_type":"PHONE","surface":"۰۹۱۲۸۱۳۵۴۲۶","language":"fa","field":null}],"gold_meta":[],"hard_negatives":["4511025429","1403/05/12","۱۴۰۳/۰۵/۱۲","1.0.0"],"description":"Persian contract with a failed national-id left for review.","description_fa":"اخطار قراردادی با کد ملی نامعتبر برای بازبینی."}
47
+ {"doc_id":"DLP-FA-CTR-07","template":"fa_contract_notice","family":"fa_contract_notice","split":"test","language":"fa","format":"text","title":"Persian contract notice","body":"موضوع: اخطار قراردادی\nنام: امید آزمایشی\nکد ملی 2154052452\nتلفن ۰۹۱۲۸۸۱۷۱۸۵\nکد ملی 2976140040 در سامانه پذیرفته نشد و برای بازبینی مانده است.\nتاریخ ۱۴۰۳/۰۵/۱۲ و نسخه ۱.۰.۰ ارجاع سند هستند.\n","metadata":{},"gold_spans":[{"start":27,"end":39,"entity_type":"PERSON","surface":"امید آزمایشی","language":"fa","field":null},{"start":47,"end":57,"entity_type":"IR_NATIONAL_ID","surface":"2154052452","language":"fa","field":null},{"start":63,"end":74,"entity_type":"PHONE","surface":"۰۹۱۲۸۸۱۷۱۸۵","language":"fa","field":null}],"gold_meta":[],"hard_negatives":["2976140040","1403/05/12","۱۴۰۳/۰۵/۱۲","1.0.0"],"description":"Persian contract with a failed national-id left for review.","description_fa":"اخطار قراردادی با کد ملی نامعتبر برای بازبینی."}
48
+ {"doc_id":"DLP-FA-CTR-08","template":"fa_contract_notice","family":"fa_contract_notice","split":"test","language":"fa","format":"text","title":"Persian contract notice","body":"موضوع: اخطار قراردادی\nنام: سارا آزمایشی\nکد ملی 2313920291\nتلفن ۰۹۱۲۱۳۴۴۶۷۵\nکد ملی 7199959037 در سامانه پذیرفته نشد و برای بازبینی مانده است.\nتاریخ ۱۴۰۳/۰۵/۱۲ و نسخه ۱.۰.۰ ارجاع سند هستند.\n","metadata":{},"gold_spans":[{"start":27,"end":39,"entity_type":"PERSON","surface":"سارا آزمایشی","language":"fa","field":null},{"start":47,"end":57,"entity_type":"IR_NATIONAL_ID","surface":"2313920291","language":"fa","field":null},{"start":63,"end":74,"entity_type":"PHONE","surface":"۰۹۱۲۱۳۴۴۶۷۵","language":"fa","field":null}],"gold_meta":[],"hard_negatives":["7199959037","1403/05/12","۱۴۰۳/۰۵/۱۲","1.0.0"],"description":"Persian contract with a failed national-id left for review.","description_fa":"اخطار قراردادی با کد ملی نامعتبر برای بازبینی."}
49
+ {"doc_id":"DLP-FA-CTR-09","template":"fa_contract_notice","family":"fa_contract_notice","split":"test","language":"fa","format":"text","title":"Persian contract notice","body":"موضوع: اخطار قراردادی\nنام: نیما نمونه‌ای\nکد ملی 0190065516\nتلفن ۰۹۱۲۷۲۰۹۰۸۸\nکد ملی 2601360047 در سامانه پذیرفته نشد و برای بازبینی مانده است.\nتاریخ ۱۴۰۳/۰۵/۱۲ و نسخه ۱.۰.۰ ارجاع سند هستند.\n","metadata":{},"gold_spans":[{"start":27,"end":40,"entity_type":"PERSON","surface":"نیما نمونه‌ای","language":"fa","field":null},{"start":48,"end":58,"entity_type":"IR_NATIONAL_ID","surface":"0190065516","language":"fa","field":null},{"start":64,"end":75,"entity_type":"PHONE","surface":"۰۹۱۲۷۲۰۹۰۸۸","language":"fa","field":null}],"gold_meta":[],"hard_negatives":["2601360047","1403/05/12","۱۴۰۳/۰۵/۱۲","1.0.0"],"description":"Persian contract with a failed national-id left for review.","description_fa":"اخطار قراردادی با کد ملی نامعتبر برای بازبینی."}
50
+ {"doc_id":"DLP-FA-CTR-10","template":"fa_contract_notice","family":"fa_contract_notice","split":"test","language":"fa","format":"text","title":"Persian contract notice","body":"موضوع: اخطار قراردادی\nنام: لیلا آزمایشی\nکد ملی 8212633349\nتلفن ۰۹۱۲۷۴۸۴۳۶۷\nکد ملی 4063697593 در سامانه پذیرفته نشد و برای بازبینی مانده است.\nتاریخ ۱۴۰۳/۰۵/۱۲ و نسخه ۱.۰.۰ ارجاع سند هستند.\n","metadata":{},"gold_spans":[{"start":27,"end":39,"entity_type":"PERSON","surface":"لیلا آزمایشی","language":"fa","field":null},{"start":47,"end":57,"entity_type":"IR_NATIONAL_ID","surface":"8212633349","language":"fa","field":null},{"start":63,"end":74,"entity_type":"PHONE","surface":"۰۹۱۲۷۴۸۴۳۶۷","language":"fa","field":null}],"gold_meta":[],"hard_negatives":["4063697593","1403/05/12","۱۴۰۳/۰۵/۱۲","1.0.0"],"description":"Persian contract with a failed national-id left for review.","description_fa":"اخطار قراردادی با کد ملی نامعتبر برای بازبینی."}
51
+ {"doc_id":"DLP-PDF-MEMO-01","template":"en_pdf_memo","family":"en_pdf_memo","split":"test","language":"en","format":"pdf","title":"PDF memo","body":"Internal memo\nAuthor: Omar Fixture\nContact: Omar.Fixture01@aria-lab.test / 09127692383\nMr. Protocol is the software module, not a person.\n","metadata":{"author":"Omar Fixture","keywords":"Omar.Fixture01@aria-lab.test","title":"Memo 1"},"gold_spans":[{"start":22,"end":34,"entity_type":"PERSON","surface":"Omar Fixture","language":"en","field":null},{"start":44,"end":72,"entity_type":"EMAIL","surface":"Omar.Fixture01@aria-lab.test","language":"en","field":null},{"start":75,"end":86,"entity_type":"PHONE","surface":"09127692383","language":"en","field":null}],"gold_meta":[{"field":"author","entity_type":"PERSON","surface":"Omar Fixture"},{"field":"keywords","entity_type":"EMAIL","surface":"Omar.Fixture01@aria-lab.test"}],"hard_negatives":["Mr. Protocol"],"description":"Text PDF whose metadata also repeats the author mailbox.","description_fa":"پی‌دی‌اف متنی که فراداده هم نام و ایمیل را تکرار می‌کند."}
52
+ {"doc_id":"DLP-PDF-MEMO-02","template":"en_pdf_memo","family":"en_pdf_memo","split":"test","language":"en","format":"pdf","title":"PDF memo","body":"Internal memo\nAuthor: Noah Fixture\nContact: Noah.Fixture02@aria-lab.test / 09125328205\nMr. Protocol is the software module, not a person.\n","metadata":{"author":"Noah Fixture","keywords":"Noah.Fixture02@aria-lab.test","title":"Memo 2"},"gold_spans":[{"start":22,"end":34,"entity_type":"PERSON","surface":"Noah Fixture","language":"en","field":null},{"start":44,"end":72,"entity_type":"EMAIL","surface":"Noah.Fixture02@aria-lab.test","language":"en","field":null},{"start":75,"end":86,"entity_type":"PHONE","surface":"09125328205","language":"en","field":null}],"gold_meta":[{"field":"author","entity_type":"PERSON","surface":"Noah Fixture"},{"field":"keywords","entity_type":"EMAIL","surface":"Noah.Fixture02@aria-lab.test"}],"hard_negatives":["Mr. Protocol"],"description":"Text PDF whose metadata also repeats the author mailbox.","description_fa":"پی‌دی‌اف متنی که فراداده هم نام و ایمیل را تکرار می‌کند."}
53
+ {"doc_id":"DLP-PDF-MEMO-03","template":"en_pdf_memo","family":"en_pdf_memo","split":"test","language":"en","format":"pdf","title":"PDF memo","body":"Internal memo\nAuthor: Omar Labian\nContact: Omar.Labian03@aria-lab.test / 09125603855\nMr. Protocol is the software module, not a person.\n","metadata":{"author":"Omar Labian","keywords":"Omar.Labian03@aria-lab.test","title":"Memo 3"},"gold_spans":[{"start":22,"end":33,"entity_type":"PERSON","surface":"Omar Labian","language":"en","field":null},{"start":43,"end":70,"entity_type":"EMAIL","surface":"Omar.Labian03@aria-lab.test","language":"en","field":null},{"start":73,"end":84,"entity_type":"PHONE","surface":"09125603855","language":"en","field":null}],"gold_meta":[{"field":"author","entity_type":"PERSON","surface":"Omar Labian"},{"field":"keywords","entity_type":"EMAIL","surface":"Omar.Labian03@aria-lab.test"}],"hard_negatives":["Mr. Protocol"],"description":"Text PDF whose metadata also repeats the author mailbox.","description_fa":"پی‌دی‌اف متنی که فراداده هم نام و ایمیل را تکرار می‌کند."}
54
+ {"doc_id":"DLP-PDF-MEMO-04","template":"en_pdf_memo","family":"en_pdf_memo","split":"test","language":"en","format":"pdf","title":"PDF memo","body":"Internal memo\nAuthor: Noah Labian\nContact: Noah.Labian04@aria-lab.test / 09127591902\nMr. Protocol is the software module, not a person.\n","metadata":{"author":"Noah Labian","keywords":"Noah.Labian04@aria-lab.test","title":"Memo 4"},"gold_spans":[{"start":22,"end":33,"entity_type":"PERSON","surface":"Noah Labian","language":"en","field":null},{"start":43,"end":70,"entity_type":"EMAIL","surface":"Noah.Labian04@aria-lab.test","language":"en","field":null},{"start":73,"end":84,"entity_type":"PHONE","surface":"09127591902","language":"en","field":null}],"gold_meta":[{"field":"author","entity_type":"PERSON","surface":"Noah Labian"},{"field":"keywords","entity_type":"EMAIL","surface":"Noah.Labian04@aria-lab.test"}],"hard_negatives":["Mr. Protocol"],"description":"Text PDF whose metadata also repeats the author mailbox.","description_fa":"پی‌دی‌اف متنی که فراداده هم نام و ایمیل را تکرار می‌کند."}
55
+ {"doc_id":"DLP-PDF-MEMO-05","template":"en_pdf_memo","family":"en_pdf_memo","split":"test","language":"en","format":"pdf","title":"PDF memo","body":"Internal memo\nAuthor: Nima Labian\nContact: Nima.Labian05@aria-lab.test / 09129473675\nMr. Protocol is the software module, not a person.\n","metadata":{"author":"Nima Labian","keywords":"Nima.Labian05@aria-lab.test","title":"Memo 5"},"gold_spans":[{"start":22,"end":33,"entity_type":"PERSON","surface":"Nima Labian","language":"en","field":null},{"start":43,"end":70,"entity_type":"EMAIL","surface":"Nima.Labian05@aria-lab.test","language":"en","field":null},{"start":73,"end":84,"entity_type":"PHONE","surface":"09129473675","language":"en","field":null}],"gold_meta":[{"field":"author","entity_type":"PERSON","surface":"Nima Labian"},{"field":"keywords","entity_type":"EMAIL","surface":"Nima.Labian05@aria-lab.test"}],"hard_negatives":["Mr. Protocol"],"description":"Text PDF whose metadata also repeats the author mailbox.","description_fa":"پی‌دی‌اف متنی که فراداده هم نام و ایمیل را تکرار می‌کند."}
56
+ {"doc_id":"DLP-PDF-MEMO-06","template":"en_pdf_memo","family":"en_pdf_memo","split":"test","language":"en","format":"pdf","title":"PDF memo","body":"Internal memo\nAuthor: Helen Fixture\nContact: Helen.Fixture06@aria-lab.test / 09126506157\nMr. Protocol is the software module, not a person.\n","metadata":{"author":"Helen Fixture","keywords":"Helen.Fixture06@aria-lab.test","title":"Memo 6"},"gold_spans":[{"start":22,"end":35,"entity_type":"PERSON","surface":"Helen Fixture","language":"en","field":null},{"start":45,"end":74,"entity_type":"EMAIL","surface":"Helen.Fixture06@aria-lab.test","language":"en","field":null},{"start":77,"end":88,"entity_type":"PHONE","surface":"09126506157","language":"en","field":null}],"gold_meta":[{"field":"author","entity_type":"PERSON","surface":"Helen Fixture"},{"field":"keywords","entity_type":"EMAIL","surface":"Helen.Fixture06@aria-lab.test"}],"hard_negatives":["Mr. Protocol"],"description":"Text PDF whose metadata also repeats the author mailbox.","description_fa":"پی‌دی‌اف متنی که فراداده هم نام و ایمیل را تکرار می‌کند."}
57
+ {"doc_id":"DLP-PDF-MEMO-07","template":"en_pdf_memo","family":"en_pdf_memo","split":"test","language":"en","format":"pdf","title":"PDF memo","body":"Internal memo\nAuthor: Ava Fixture\nContact: Ava.Fixture07@aria-lab.test / 09126633882\nMr. Protocol is the software module, not a person.\n","metadata":{"author":"Ava Fixture","keywords":"Ava.Fixture07@aria-lab.test","title":"Memo 7"},"gold_spans":[{"start":22,"end":33,"entity_type":"PERSON","surface":"Ava Fixture","language":"en","field":null},{"start":43,"end":70,"entity_type":"EMAIL","surface":"Ava.Fixture07@aria-lab.test","language":"en","field":null},{"start":73,"end":84,"entity_type":"PHONE","surface":"09126633882","language":"en","field":null}],"gold_meta":[{"field":"author","entity_type":"PERSON","surface":"Ava Fixture"},{"field":"keywords","entity_type":"EMAIL","surface":"Ava.Fixture07@aria-lab.test"}],"hard_negatives":["Mr. Protocol"],"description":"Text PDF whose metadata also repeats the author mailbox.","description_fa":"پی‌دی‌اف متنی که فراداده هم نام و ایمیل را تکرار می‌کند."}
58
+ {"doc_id":"DLP-PDF-MEMO-08","template":"en_pdf_memo","family":"en_pdf_memo","split":"test","language":"en","format":"pdf","title":"PDF memo","body":"Internal memo\nAuthor: Nima Labian\nContact: Nima.Labian08@aria-lab.test / 09128661790\nMr. Protocol is the software module, not a person.\n","metadata":{"author":"Nima Labian","keywords":"Nima.Labian08@aria-lab.test","title":"Memo 8"},"gold_spans":[{"start":22,"end":33,"entity_type":"PERSON","surface":"Nima Labian","language":"en","field":null},{"start":43,"end":70,"entity_type":"EMAIL","surface":"Nima.Labian08@aria-lab.test","language":"en","field":null},{"start":73,"end":84,"entity_type":"PHONE","surface":"09128661790","language":"en","field":null}],"gold_meta":[{"field":"author","entity_type":"PERSON","surface":"Nima Labian"},{"field":"keywords","entity_type":"EMAIL","surface":"Nima.Labian08@aria-lab.test"}],"hard_negatives":["Mr. Protocol"],"description":"Text PDF whose metadata also repeats the author mailbox.","description_fa":"پی‌دی‌اف متنی که فراداده هم نام و ایمیل را تکرار می‌کند."}
59
+ {"doc_id":"DLP-PDF-MEMO-09","template":"en_pdf_memo","family":"en_pdf_memo","split":"test","language":"en","format":"pdf","title":"PDF memo","body":"Internal memo\nAuthor: Ava Fixture\nContact: Ava.Fixture09@aria-lab.test / 09122509645\nMr. Protocol is the software module, not a person.\n","metadata":{"author":"Ava Fixture","keywords":"Ava.Fixture09@aria-lab.test","title":"Memo 9"},"gold_spans":[{"start":22,"end":33,"entity_type":"PERSON","surface":"Ava Fixture","language":"en","field":null},{"start":43,"end":70,"entity_type":"EMAIL","surface":"Ava.Fixture09@aria-lab.test","language":"en","field":null},{"start":73,"end":84,"entity_type":"PHONE","surface":"09122509645","language":"en","field":null}],"gold_meta":[{"field":"author","entity_type":"PERSON","surface":"Ava Fixture"},{"field":"keywords","entity_type":"EMAIL","surface":"Ava.Fixture09@aria-lab.test"}],"hard_negatives":["Mr. Protocol"],"description":"Text PDF whose metadata also repeats the author mailbox.","description_fa":"پی‌دی‌اف متنی که فراداده هم نام و ایمیل را تکرار می‌کند."}
60
+ {"doc_id":"DLP-PDF-MEMO-10","template":"en_pdf_memo","family":"en_pdf_memo","split":"test","language":"en","format":"pdf","title":"PDF memo","body":"Internal memo\nAuthor: Nima Fixture\nContact: Nima.Fixture10@aria-lab.test / 09122528399\nMr. Protocol is the software module, not a person.\n","metadata":{"author":"Nima Fixture","keywords":"Nima.Fixture10@aria-lab.test","title":"Memo 10"},"gold_spans":[{"start":22,"end":34,"entity_type":"PERSON","surface":"Nima Fixture","language":"en","field":null},{"start":44,"end":72,"entity_type":"EMAIL","surface":"Nima.Fixture10@aria-lab.test","language":"en","field":null},{"start":75,"end":86,"entity_type":"PHONE","surface":"09122528399","language":"en","field":null}],"gold_meta":[{"field":"author","entity_type":"PERSON","surface":"Nima Fixture"},{"field":"keywords","entity_type":"EMAIL","surface":"Nima.Fixture10@aria-lab.test"}],"hard_negatives":["Mr. Protocol"],"description":"Text PDF whose metadata also repeats the author mailbox.","description_fa":"پی‌دی‌اف متنی که فراداده هم نام و ایمیل را تکرار می‌کند."}
61
+ {"doc_id":"DLP-HARDNEG-01","template":"hardneg_catalog","family":"hardneg_catalog","split":"test","language":"en","format":"text","title":"Hard-negative catalog","body":"Catalog row 1\nISBN 978-0-306-40615-7\nInvoice 9882096604\nCard-like 4111799672516487\nIBAN-like IR969760340568087700433687\nMr. Protocol handler v2.4.1\nDate 2026/09/21 SKU-9001\n","metadata":{},"gold_spans":[],"gold_meta":[],"hard_negatives":["4111799672516487","IR969760340568087700433687","9882096604","Mr. Protocol","2.4.1","978-0-306-40615-7"],"description":"Lookalikes that must not be redacted: failed checksums, ISBN, protocol title.","description_fa":"شبیه‌سازی‌هایی که نباید پوشانده شوند: checksum نامعتبر، شابک، عنوان پروتکل."}
62
+ {"doc_id":"DLP-HARDNEG-02","template":"hardneg_catalog","family":"hardneg_catalog","split":"test","language":"en","format":"text","title":"Hard-negative catalog","body":"Catalog row 2\nISBN 978-0-306-40615-7\nInvoice 7209502842\nCard-like 4111134559932047\nIBAN-like IR859360346783598885359083\nMr. Protocol handler v2.4.1\nDate 2026/09/21 SKU-9002\n","metadata":{},"gold_spans":[],"gold_meta":[],"hard_negatives":["4111134559932047","IR859360346783598885359083","7209502842","Mr. Protocol","2.4.1","978-0-306-40615-7"],"description":"Lookalikes that must not be redacted: failed checksums, ISBN, protocol title.","description_fa":"شبیه‌سازی‌هایی که نباید پوشانده شوند: checksum نامعتبر، شابک، عنوان پروتکل."}
63
+ {"doc_id":"DLP-HARDNEG-03","template":"hardneg_catalog","family":"hardneg_catalog","split":"test","language":"en","format":"text","title":"Hard-negative catalog","body":"Catalog row 3\nISBN 978-0-306-40615-7\nInvoice 8163562111\nCard-like 4111842764173344\nIBAN-like IR189190290458319906920450\nMr. Protocol handler v2.4.1\nDate 2026/09/21 SKU-9003\n","metadata":{},"gold_spans":[],"gold_meta":[],"hard_negatives":["4111842764173344","IR189190290458319906920450","8163562111","Mr. Protocol","2.4.1","978-0-306-40615-7"],"description":"Lookalikes that must not be redacted: failed checksums, ISBN, protocol title.","description_fa":"شبیه‌سازی‌هایی که نباید پوشانده شوند: checksum نامعتبر، شابک، عنوان پروتکل."}
64
+ {"doc_id":"DLP-HARDNEG-04","template":"hardneg_catalog","family":"hardneg_catalog","split":"test","language":"en","format":"text","title":"Hard-negative catalog","body":"Catalog row 4\nISBN 978-0-306-40615-7\nInvoice 2665961572\nCard-like 4111391694332034\nIBAN-like IR289390013050842973101189\nMr. Protocol handler v2.4.1\nDate 2026/09/21 SKU-9004\n","metadata":{},"gold_spans":[],"gold_meta":[],"hard_negatives":["4111391694332034","IR289390013050842973101189","2665961572","Mr. Protocol","2.4.1","978-0-306-40615-7"],"description":"Lookalikes that must not be redacted: failed checksums, ISBN, protocol title.","description_fa":"شبیه‌سازی‌هایی که نباید پوشانده شوند: checksum نامعتبر، شابک، عنوان پروتکل."}
65
+ {"doc_id":"DLP-HARDNEG-05","template":"hardneg_catalog","family":"hardneg_catalog","split":"test","language":"en","format":"text","title":"Hard-negative catalog","body":"Catalog row 5\nISBN 978-0-306-40615-7\nInvoice 5490470241\nCard-like 4111874194171432\nIBAN-like IR549770672881619226418819\nMr. Protocol handler v2.4.1\nDate 2026/09/21 SKU-9005\n","metadata":{},"gold_spans":[],"gold_meta":[],"hard_negatives":["4111874194171432","IR549770672881619226418819","5490470241","Mr. Protocol","2.4.1","978-0-306-40615-7"],"description":"Lookalikes that must not be redacted: failed checksums, ISBN, protocol title.","description_fa":"شبیه‌سازی‌هایی که نباید پوشانده شوند: checksum نامعتبر، شابک، عنوان پروتکل."}
66
+ {"doc_id":"DLP-HARDNEG-06","template":"hardneg_catalog","family":"hardneg_catalog","split":"test","language":"en","format":"text","title":"Hard-negative catalog","body":"Catalog row 6\nISBN 978-0-306-40615-7\nInvoice 9997521600\nCard-like 4111260024722695\nIBAN-like IR299310569176970687717687\nMr. Protocol handler v2.4.1\nDate 2026/09/21 SKU-9006\n","metadata":{},"gold_spans":[],"gold_meta":[],"hard_negatives":["4111260024722695","IR299310569176970687717687","9997521600","Mr. Protocol","2.4.1","978-0-306-40615-7"],"description":"Lookalikes that must not be redacted: failed checksums, ISBN, protocol title.","description_fa":"شبیه‌سازی‌هایی که نباید پوشانده شوند: checksum نامعتبر، شابک، عنوان پروتکل."}
67
+ {"doc_id":"DLP-HARDNEG-07","template":"hardneg_catalog","family":"hardneg_catalog","split":"test","language":"en","format":"text","title":"Hard-negative catalog","body":"Catalog row 7\nISBN 978-0-306-40615-7\nInvoice 4413142056\nCard-like 4111156662872018\nIBAN-like IR849380980760628231287991\nMr. Protocol handler v2.4.1\nDate 2026/09/21 SKU-9007\n","metadata":{},"gold_spans":[],"gold_meta":[],"hard_negatives":["4111156662872018","IR849380980760628231287991","4413142056","Mr. Protocol","2.4.1","978-0-306-40615-7"],"description":"Lookalikes that must not be redacted: failed checksums, ISBN, protocol title.","description_fa":"شبیه‌سازی‌هایی که نباید پوشانده شوند: checksum نامعتبر، شابک، عنوان پروتکل."}
68
+ {"doc_id":"DLP-HARDNEG-08","template":"hardneg_catalog","family":"hardneg_catalog","split":"test","language":"en","format":"text","title":"Hard-negative catalog","body":"Catalog row 8\nISBN 978-0-306-40615-7\nInvoice 9844497419\nCard-like 4111658027974926\nIBAN-like IR749310432272162828796518\nMr. Protocol handler v2.4.1\nDate 2026/09/21 SKU-9008\n","metadata":{},"gold_spans":[],"gold_meta":[],"hard_negatives":["4111658027974926","IR749310432272162828796518","9844497419","Mr. Protocol","2.4.1","978-0-306-40615-7"],"description":"Lookalikes that must not be redacted: failed checksums, ISBN, protocol title.","description_fa":"شبیه‌سازی‌هایی که نباید پوشانده شوند: checksum نامعتبر، شابک، عنوان پروتکل."}
69
+ {"doc_id":"DLP-HARDNEG-09","template":"hardneg_catalog","family":"hardneg_catalog","split":"test","language":"en","format":"text","title":"Hard-negative catalog","body":"Catalog row 9\nISBN 978-0-306-40615-7\nInvoice 5948674852\nCard-like 4111784726661737\nIBAN-like IR529710074259549882292605\nMr. Protocol handler v2.4.1\nDate 2026/09/21 SKU-9009\n","metadata":{},"gold_spans":[],"gold_meta":[],"hard_negatives":["4111784726661737","IR529710074259549882292605","5948674852","Mr. Protocol","2.4.1","978-0-306-40615-7"],"description":"Lookalikes that must not be redacted: failed checksums, ISBN, protocol title.","description_fa":"شبیه‌سازی‌هایی که نباید پوشانده شوند: checksum نامعتبر، شابک، عنوان پروتکل."}
70
+ {"doc_id":"DLP-HARDNEG-10","template":"hardneg_catalog","family":"hardneg_catalog","split":"test","language":"en","format":"text","title":"Hard-negative catalog","body":"Catalog row 10\nISBN 978-0-306-40615-7\nInvoice 2672086444\nCard-like 4111733944863765\nIBAN-like IR739740106542912708354792\nMr. Protocol handler v2.4.1\nDate 2026/09/21 SKU-9010\n","metadata":{},"gold_spans":[],"gold_meta":[],"hard_negatives":["4111733944863765","IR739740106542912708354792","2672086444","Mr. Protocol","2.4.1","978-0-306-40615-7"],"description":"Lookalikes that must not be redacted: failed checksums, ISBN, protocol title.","description_fa":"شبیه‌سازی‌هایی که نباید پوشانده شوند: checksum نامعتبر، شابک، عنوان پروتکل."}
71
+ {"doc_id":"DLP-TICKET-01","template":"mixed_ticket","family":"mixed_ticket","split":"test","language":"mixed","format":"csv","title":"Mixed-language ticket CSV","body":"ticket_id,reporter,email,phone,national_id\nINV-MIX-0001,Leila Fixture / سارا آزمایشی,Leila.Fixture01@aria-lab.test,09126663672,8810500075\n","metadata":{},"gold_spans":[{"start":56,"end":69,"entity_type":"PERSON","surface":"Leila Fixture","language":"en","field":null},{"start":72,"end":84,"entity_type":"PERSON","surface":"سارا آزمایشی","language":"fa","field":null},{"start":85,"end":114,"entity_type":"EMAIL","surface":"Leila.Fixture01@aria-lab.test","language":"en","field":null},{"start":115,"end":126,"entity_type":"PHONE","surface":"09126663672","language":"en","field":null},{"start":127,"end":137,"entity_type":"IR_NATIONAL_ID","surface":"8810500075","language":"fa","field":null}],"gold_meta":[],"hard_negatives":["INV-MIX-0001"],"description":"Mixed English/Persian ticket row.","description_fa":"ردیف تیکت دوزبانه."}
72
+ {"doc_id":"DLP-TICKET-02","template":"mixed_ticket","family":"mixed_ticket","split":"test","language":"mixed","format":"csv","title":"Mixed-language ticket CSV","body":"ticket_id,reporter,email,phone,national_id\nINV-MIX-0002,Leila Fixture / امید نمونه‌ای,Leila.Fixture02@aria-lab.test,09124960362,7433177132\n","metadata":{},"gold_spans":[{"start":56,"end":69,"entity_type":"PERSON","surface":"Leila Fixture","language":"en","field":null},{"start":72,"end":85,"entity_type":"PERSON","surface":"امید نمونه‌ای","language":"fa","field":null},{"start":86,"end":115,"entity_type":"EMAIL","surface":"Leila.Fixture02@aria-lab.test","language":"en","field":null},{"start":116,"end":127,"entity_type":"PHONE","surface":"09124960362","language":"en","field":null},{"start":128,"end":138,"entity_type":"IR_NATIONAL_ID","surface":"7433177132","language":"fa","field":null}],"gold_meta":[],"hard_negatives":["INV-MIX-0002"],"description":"Mixed English/Persian ticket row.","description_fa":"ردیف تیکت دوزبانه."}
73
+ {"doc_id":"DLP-TICKET-03","template":"mixed_ticket","family":"mixed_ticket","split":"test","language":"mixed","format":"csv","title":"Mixed-language ticket CSV","body":"ticket_id,reporter,email,phone,national_id\nINV-MIX-0003,Ava Fixture / نیما نمونه‌ای,Ava.Fixture03@aria-lab.test,09129013939,4751024639\n","metadata":{},"gold_spans":[{"start":56,"end":67,"entity_type":"PERSON","surface":"Ava Fixture","language":"en","field":null},{"start":70,"end":83,"entity_type":"PERSON","surface":"نیما نمونه‌ای","language":"fa","field":null},{"start":84,"end":111,"entity_type":"EMAIL","surface":"Ava.Fixture03@aria-lab.test","language":"en","field":null},{"start":112,"end":123,"entity_type":"PHONE","surface":"09129013939","language":"en","field":null},{"start":124,"end":134,"entity_type":"IR_NATIONAL_ID","surface":"4751024639","language":"fa","field":null}],"gold_meta":[],"hard_negatives":["INV-MIX-0003"],"description":"Mixed English/Persian ticket row.","description_fa":"ردیف تیکت دوزبانه."}
74
+ {"doc_id":"DLP-TICKET-04","template":"mixed_ticket","family":"mixed_ticket","split":"test","language":"mixed","format":"csv","title":"Mixed-language ticket CSV","body":"ticket_id,reporter,email,phone,national_id\nINV-MIX-0004,Omar Fixture / سارا نمونه‌ای,Omar.Fixture04@aria-lab.test,09124265160,0843397276\n","metadata":{},"gold_spans":[{"start":56,"end":68,"entity_type":"PERSON","surface":"Omar Fixture","language":"en","field":null},{"start":71,"end":84,"entity_type":"PERSON","surface":"سارا نمونه‌ای","language":"fa","field":null},{"start":85,"end":113,"entity_type":"EMAIL","surface":"Omar.Fixture04@aria-lab.test","language":"en","field":null},{"start":114,"end":125,"entity_type":"PHONE","surface":"09124265160","language":"en","field":null},{"start":126,"end":136,"entity_type":"IR_NATIONAL_ID","surface":"0843397276","language":"fa","field":null}],"gold_meta":[],"hard_negatives":["INV-MIX-0004"],"description":"Mixed English/Persian ticket row.","description_fa":"ردیف تیکت دوزبانه."}
75
+ {"doc_id":"DLP-TICKET-05","template":"mixed_ticket","family":"mixed_ticket","split":"test","language":"mixed","format":"csv","title":"Mixed-language ticket CSV","body":"ticket_id,reporter,email,phone,national_id\nINV-MIX-0005,Ava Fixture / امید نمونه‌ای,Ava.Fixture05@aria-lab.test,09127385597,6714749521\n","metadata":{},"gold_spans":[{"start":56,"end":67,"entity_type":"PERSON","surface":"Ava Fixture","language":"en","field":null},{"start":70,"end":83,"entity_type":"PERSON","surface":"امید نمونه‌ای","language":"fa","field":null},{"start":84,"end":111,"entity_type":"EMAIL","surface":"Ava.Fixture05@aria-lab.test","language":"en","field":null},{"start":112,"end":123,"entity_type":"PHONE","surface":"09127385597","language":"en","field":null},{"start":124,"end":134,"entity_type":"IR_NATIONAL_ID","surface":"6714749521","language":"fa","field":null}],"gold_meta":[],"hard_negatives":["INV-MIX-0005"],"description":"Mixed English/Persian ticket row.","description_fa":"ردیف تیکت دوزبانه."}
76
+ {"doc_id":"DLP-TICKET-06","template":"mixed_ticket","family":"mixed_ticket","split":"test","language":"mixed","format":"csv","title":"Mixed-language ticket CSV","body":"ticket_id,reporter,email,phone,national_id\nINV-MIX-0006,Omar Fixture / لیلا آزمایشی,Omar.Fixture06@aria-lab.test,09126454625,7904846901\n","metadata":{},"gold_spans":[{"start":56,"end":68,"entity_type":"PERSON","surface":"Omar Fixture","language":"en","field":null},{"start":71,"end":83,"entity_type":"PERSON","surface":"لیلا آزمایشی","language":"fa","field":null},{"start":84,"end":112,"entity_type":"EMAIL","surface":"Omar.Fixture06@aria-lab.test","language":"en","field":null},{"start":113,"end":124,"entity_type":"PHONE","surface":"09126454625","language":"en","field":null},{"start":125,"end":135,"entity_type":"IR_NATIONAL_ID","surface":"7904846901","language":"fa","field":null}],"gold_meta":[],"hard_negatives":["INV-MIX-0006"],"description":"Mixed English/Persian ticket row.","description_fa":"ردیف تیکت دوزبانه."}
77
+ {"doc_id":"DLP-TICKET-07","template":"mixed_ticket","family":"mixed_ticket","split":"test","language":"mixed","format":"csv","title":"Mixed-language ticket CSV","body":"ticket_id,reporter,email,phone,national_id\nINV-MIX-0007,Helen Fixture / سارا آزمایشی,Helen.Fixture07@aria-lab.test,09127938081,6114411269\n","metadata":{},"gold_spans":[{"start":56,"end":69,"entity_type":"PERSON","surface":"Helen Fixture","language":"en","field":null},{"start":72,"end":84,"entity_type":"PERSON","surface":"سارا آزمایشی","language":"fa","field":null},{"start":85,"end":114,"entity_type":"EMAIL","surface":"Helen.Fixture07@aria-lab.test","language":"en","field":null},{"start":115,"end":126,"entity_type":"PHONE","surface":"09127938081","language":"en","field":null},{"start":127,"end":137,"entity_type":"IR_NATIONAL_ID","surface":"6114411269","language":"fa","field":null}],"gold_meta":[],"hard_negatives":["INV-MIX-0007"],"description":"Mixed English/Persian ticket row.","description_fa":"ردیف تیکت دوزبانه."}
78
+ {"doc_id":"DLP-TICKET-08","template":"mixed_ticket","family":"mixed_ticket","split":"test","language":"mixed","format":"csv","title":"Mixed-language ticket CSV","body":"ticket_id,reporter,email,phone,national_id\nINV-MIX-0008,Ava Fixture / لیلا نمونه‌ای,Ava.Fixture08@aria-lab.test,09125317904,2799826970\n","metadata":{},"gold_spans":[{"start":56,"end":67,"entity_type":"PERSON","surface":"Ava Fixture","language":"en","field":null},{"start":70,"end":83,"entity_type":"PERSON","surface":"لیلا نمونه‌ای","language":"fa","field":null},{"start":84,"end":111,"entity_type":"EMAIL","surface":"Ava.Fixture08@aria-lab.test","language":"en","field":null},{"start":112,"end":123,"entity_type":"PHONE","surface":"09125317904","language":"en","field":null},{"start":124,"end":134,"entity_type":"IR_NATIONAL_ID","surface":"2799826970","language":"fa","field":null}],"gold_meta":[],"hard_negatives":["INV-MIX-0008"],"description":"Mixed English/Persian ticket row.","description_fa":"ردیف تیکت دوزبانه."}
79
+ {"doc_id":"DLP-TICKET-09","template":"mixed_ticket","family":"mixed_ticket","split":"test","language":"mixed","format":"csv","title":"Mixed-language ticket CSV","body":"ticket_id,reporter,email,phone,national_id\nINV-MIX-0009,Omar Fixture / نیما نمونه‌ای,Omar.Fixture09@aria-lab.test,09125714005,0468470395\n","metadata":{},"gold_spans":[{"start":56,"end":68,"entity_type":"PERSON","surface":"Omar Fixture","language":"en","field":null},{"start":71,"end":84,"entity_type":"PERSON","surface":"نیما نمونه‌ای","language":"fa","field":null},{"start":85,"end":113,"entity_type":"EMAIL","surface":"Omar.Fixture09@aria-lab.test","language":"en","field":null},{"start":114,"end":125,"entity_type":"PHONE","surface":"09125714005","language":"en","field":null},{"start":126,"end":136,"entity_type":"IR_NATIONAL_ID","surface":"0468470395","language":"fa","field":null}],"gold_meta":[],"hard_negatives":["INV-MIX-0009"],"description":"Mixed English/Persian ticket row.","description_fa":"ردیف تیکت دوزبانه."}
80
+ {"doc_id":"DLP-TICKET-10","template":"mixed_ticket","family":"mixed_ticket","split":"test","language":"mixed","format":"csv","title":"Mixed-language ticket CSV","body":"ticket_id,reporter,email,phone,national_id\nINV-MIX-0010,Helen Labian / سارا نمونه‌ای,Helen.Labian10@aria-lab.test,09126818844,5049585570\n","metadata":{},"gold_spans":[{"start":56,"end":68,"entity_type":"PERSON","surface":"Helen Labian","language":"en","field":null},{"start":71,"end":84,"entity_type":"PERSON","surface":"سارا نمونه‌ای","language":"fa","field":null},{"start":85,"end":113,"entity_type":"EMAIL","surface":"Helen.Labian10@aria-lab.test","language":"en","field":null},{"start":114,"end":125,"entity_type":"PHONE","surface":"09126818844","language":"en","field":null},{"start":126,"end":136,"entity_type":"IR_NATIONAL_ID","surface":"5049585570","language":"fa","field":null}],"gold_meta":[],"hard_negatives":["INV-MIX-0010"],"description":"Mixed English/Persian ticket row.","description_fa":"ردیف تیکت دوزبانه."}
demo/examples/evaluation.json ADDED
@@ -0,0 +1,88 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "seed": 24,
3
+ "n_documents": 80,
4
+ "n_scenarios": 5,
5
+ "engine_profile": "presidio-style-checksum-policy",
6
+ "rows": [
7
+ {
8
+ "scenario_id": "S01",
9
+ "method": "checksum_policy",
10
+ "precision": 1.0,
11
+ "recall": 1.0,
12
+ "over_redaction": 0.0,
13
+ "residual_leak": 0.0,
14
+ "leftover_exports": 0,
15
+ "metadata_leaks": 0,
16
+ "hardneg_hits": 0,
17
+ "n_docs": 20
18
+ },
19
+ {
20
+ "scenario_id": "S02",
21
+ "method": "checksum_policy",
22
+ "precision": 1.0,
23
+ "recall": 1.0,
24
+ "over_redaction": 0.0,
25
+ "residual_leak": 0.0,
26
+ "leftover_exports": 0,
27
+ "metadata_leaks": 0,
28
+ "hardneg_hits": 0,
29
+ "n_docs": 20
30
+ },
31
+ {
32
+ "scenario_id": "S03",
33
+ "method": "checksum_policy",
34
+ "precision": 1.0,
35
+ "recall": 1.0,
36
+ "over_redaction": 0.0,
37
+ "residual_leak": 0.0,
38
+ "leftover_exports": 0,
39
+ "metadata_leaks": 0,
40
+ "hardneg_hits": 0,
41
+ "n_docs": 10
42
+ },
43
+ {
44
+ "scenario_id": "S04",
45
+ "method": "checksum_policy",
46
+ "precision": 1.0,
47
+ "recall": 1.0,
48
+ "over_redaction": 0.0,
49
+ "residual_leak": 0.0,
50
+ "leftover_exports": 0,
51
+ "metadata_leaks": 0,
52
+ "hardneg_hits": 10,
53
+ "n_docs": 20
54
+ },
55
+ {
56
+ "scenario_id": "S05",
57
+ "method": "checksum_policy",
58
+ "precision": 1.0,
59
+ "recall": 1.0,
60
+ "over_redaction": 0.0,
61
+ "residual_leak": 0.0,
62
+ "leftover_exports": 0,
63
+ "metadata_leaks": 0,
64
+ "hardneg_hits": 10,
65
+ "n_docs": 80
66
+ }
67
+ ],
68
+ "sample_kpis": {
69
+ "n_docs": 20,
70
+ "n_spans": 70,
71
+ "n_review": 0,
72
+ "n_redacted": 70,
73
+ "policy_precision": 1.0,
74
+ "policy_recall": 1.0,
75
+ "baseline_precision": 0.8333333333333334,
76
+ "baseline_recall": 0.7142857142857143,
77
+ "residual_leak": 0.0,
78
+ "over_redaction": 0.0,
79
+ "leftover_exports": 0,
80
+ "metadata_leaks": 0,
81
+ "elapsed_ms": 52,
82
+ "engine_profile": "presidio-style-checksum-policy",
83
+ "policy_action": "mask",
84
+ "pages": 0,
85
+ "chars": 2897
86
+ },
87
+ "note": "Synthetic laboratory corpus. A clean suite is not complete production redaction."
88
+ }
demo/examples/feature_card.md ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ # Feature card
2
+
3
+ - Task: span detection and reversible laboratory redaction
4
+ - Languages: English, Persian (digits, ZWNJ)
5
+ - Formats: text, CSV, text PDF (≤10 pages)
6
+ - Baseline: ASCII regex without checksum
7
+ - Policy: Presidio-style recognizer registry + checksum + lexicon
8
+ - Split: template-disjoint
demo/examples/gold_spans.json ADDED
@@ -0,0 +1,2662 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ [
2
+ {
3
+ "doc_id": "DLP-EN-EMAIL-01",
4
+ "spans": [
5
+ {
6
+ "start": 6,
7
+ "end": 18,
8
+ "entity_type": "PERSON",
9
+ "surface": "Nima Fixture",
10
+ "language": "en",
11
+ "field": null
12
+ },
13
+ {
14
+ "start": 20,
15
+ "end": 48,
16
+ "entity_type": "EMAIL",
17
+ "surface": "Nima.Fixture01@aria-lab.test",
18
+ "language": "en",
19
+ "field": null
20
+ },
21
+ {
22
+ "start": 57,
23
+ "end": 68,
24
+ "entity_type": "PHONE",
25
+ "surface": "09124063383",
26
+ "language": "en",
27
+ "field": null
28
+ },
29
+ {
30
+ "start": 79,
31
+ "end": 89,
32
+ "entity_type": "EMPLOYEE_ID",
33
+ "surface": "ARIA-24001",
34
+ "language": "en",
35
+ "field": null
36
+ }
37
+ ],
38
+ "meta": []
39
+ },
40
+ {
41
+ "doc_id": "DLP-EN-EMAIL-02",
42
+ "spans": [
43
+ {
44
+ "start": 6,
45
+ "end": 17,
46
+ "entity_type": "PERSON",
47
+ "surface": "Noah Labian",
48
+ "language": "en",
49
+ "field": null
50
+ },
51
+ {
52
+ "start": 19,
53
+ "end": 46,
54
+ "entity_type": "EMAIL",
55
+ "surface": "Noah.Labian02@aria-lab.test",
56
+ "language": "en",
57
+ "field": null
58
+ },
59
+ {
60
+ "start": 55,
61
+ "end": 66,
62
+ "entity_type": "PHONE",
63
+ "surface": "09124256381",
64
+ "language": "en",
65
+ "field": null
66
+ },
67
+ {
68
+ "start": 77,
69
+ "end": 87,
70
+ "entity_type": "EMPLOYEE_ID",
71
+ "surface": "ARIA-24002",
72
+ "language": "en",
73
+ "field": null
74
+ }
75
+ ],
76
+ "meta": []
77
+ },
78
+ {
79
+ "doc_id": "DLP-EN-EMAIL-03",
80
+ "spans": [
81
+ {
82
+ "start": 6,
83
+ "end": 17,
84
+ "entity_type": "PERSON",
85
+ "surface": "Noah Labian",
86
+ "language": "en",
87
+ "field": null
88
+ },
89
+ {
90
+ "start": 19,
91
+ "end": 46,
92
+ "entity_type": "EMAIL",
93
+ "surface": "Noah.Labian03@aria-lab.test",
94
+ "language": "en",
95
+ "field": null
96
+ },
97
+ {
98
+ "start": 55,
99
+ "end": 66,
100
+ "entity_type": "PHONE",
101
+ "surface": "09123538397",
102
+ "language": "en",
103
+ "field": null
104
+ },
105
+ {
106
+ "start": 77,
107
+ "end": 87,
108
+ "entity_type": "EMPLOYEE_ID",
109
+ "surface": "ARIA-24003",
110
+ "language": "en",
111
+ "field": null
112
+ }
113
+ ],
114
+ "meta": []
115
+ },
116
+ {
117
+ "doc_id": "DLP-EN-EMAIL-04",
118
+ "spans": [
119
+ {
120
+ "start": 6,
121
+ "end": 18,
122
+ "entity_type": "PERSON",
123
+ "surface": "Nima Fixture",
124
+ "language": "en",
125
+ "field": null
126
+ },
127
+ {
128
+ "start": 20,
129
+ "end": 48,
130
+ "entity_type": "EMAIL",
131
+ "surface": "Nima.Fixture04@aria-lab.test",
132
+ "language": "en",
133
+ "field": null
134
+ },
135
+ {
136
+ "start": 57,
137
+ "end": 68,
138
+ "entity_type": "PHONE",
139
+ "surface": "09121213266",
140
+ "language": "en",
141
+ "field": null
142
+ },
143
+ {
144
+ "start": 79,
145
+ "end": 89,
146
+ "entity_type": "EMPLOYEE_ID",
147
+ "surface": "ARIA-24004",
148
+ "language": "en",
149
+ "field": null
150
+ }
151
+ ],
152
+ "meta": []
153
+ },
154
+ {
155
+ "doc_id": "DLP-EN-EMAIL-05",
156
+ "spans": [
157
+ {
158
+ "start": 6,
159
+ "end": 18,
160
+ "entity_type": "PERSON",
161
+ "surface": "Omar Fixture",
162
+ "language": "en",
163
+ "field": null
164
+ },
165
+ {
166
+ "start": 20,
167
+ "end": 48,
168
+ "entity_type": "EMAIL",
169
+ "surface": "Omar.Fixture05@aria-lab.test",
170
+ "language": "en",
171
+ "field": null
172
+ },
173
+ {
174
+ "start": 57,
175
+ "end": 68,
176
+ "entity_type": "PHONE",
177
+ "surface": "09122939341",
178
+ "language": "en",
179
+ "field": null
180
+ },
181
+ {
182
+ "start": 79,
183
+ "end": 89,
184
+ "entity_type": "EMPLOYEE_ID",
185
+ "surface": "ARIA-24005",
186
+ "language": "en",
187
+ "field": null
188
+ }
189
+ ],
190
+ "meta": []
191
+ },
192
+ {
193
+ "doc_id": "DLP-EN-EMAIL-06",
194
+ "spans": [
195
+ {
196
+ "start": 6,
197
+ "end": 16,
198
+ "entity_type": "PERSON",
199
+ "surface": "Ava Labian",
200
+ "language": "en",
201
+ "field": null
202
+ },
203
+ {
204
+ "start": 18,
205
+ "end": 44,
206
+ "entity_type": "EMAIL",
207
+ "surface": "Ava.Labian06@aria-lab.test",
208
+ "language": "en",
209
+ "field": null
210
+ },
211
+ {
212
+ "start": 53,
213
+ "end": 64,
214
+ "entity_type": "PHONE",
215
+ "surface": "09129242814",
216
+ "language": "en",
217
+ "field": null
218
+ },
219
+ {
220
+ "start": 75,
221
+ "end": 85,
222
+ "entity_type": "EMPLOYEE_ID",
223
+ "surface": "ARIA-24006",
224
+ "language": "en",
225
+ "field": null
226
+ }
227
+ ],
228
+ "meta": []
229
+ },
230
+ {
231
+ "doc_id": "DLP-EN-EMAIL-07",
232
+ "spans": [
233
+ {
234
+ "start": 6,
235
+ "end": 18,
236
+ "entity_type": "PERSON",
237
+ "surface": "Nima Fixture",
238
+ "language": "en",
239
+ "field": null
240
+ },
241
+ {
242
+ "start": 20,
243
+ "end": 48,
244
+ "entity_type": "EMAIL",
245
+ "surface": "Nima.Fixture07@aria-lab.test",
246
+ "language": "en",
247
+ "field": null
248
+ },
249
+ {
250
+ "start": 57,
251
+ "end": 68,
252
+ "entity_type": "PHONE",
253
+ "surface": "09126110356",
254
+ "language": "en",
255
+ "field": null
256
+ },
257
+ {
258
+ "start": 79,
259
+ "end": 89,
260
+ "entity_type": "EMPLOYEE_ID",
261
+ "surface": "ARIA-24007",
262
+ "language": "en",
263
+ "field": null
264
+ }
265
+ ],
266
+ "meta": []
267
+ },
268
+ {
269
+ "doc_id": "DLP-EN-EMAIL-08",
270
+ "spans": [
271
+ {
272
+ "start": 6,
273
+ "end": 17,
274
+ "entity_type": "PERSON",
275
+ "surface": "Omar Labian",
276
+ "language": "en",
277
+ "field": null
278
+ },
279
+ {
280
+ "start": 19,
281
+ "end": 46,
282
+ "entity_type": "EMAIL",
283
+ "surface": "Omar.Labian08@aria-lab.test",
284
+ "language": "en",
285
+ "field": null
286
+ },
287
+ {
288
+ "start": 55,
289
+ "end": 66,
290
+ "entity_type": "PHONE",
291
+ "surface": "09125259879",
292
+ "language": "en",
293
+ "field": null
294
+ },
295
+ {
296
+ "start": 77,
297
+ "end": 87,
298
+ "entity_type": "EMPLOYEE_ID",
299
+ "surface": "ARIA-24008",
300
+ "language": "en",
301
+ "field": null
302
+ }
303
+ ],
304
+ "meta": []
305
+ },
306
+ {
307
+ "doc_id": "DLP-EN-EMAIL-09",
308
+ "spans": [
309
+ {
310
+ "start": 6,
311
+ "end": 18,
312
+ "entity_type": "PERSON",
313
+ "surface": "Helen Labian",
314
+ "language": "en",
315
+ "field": null
316
+ },
317
+ {
318
+ "start": 20,
319
+ "end": 48,
320
+ "entity_type": "EMAIL",
321
+ "surface": "Helen.Labian09@aria-lab.test",
322
+ "language": "en",
323
+ "field": null
324
+ },
325
+ {
326
+ "start": 57,
327
+ "end": 68,
328
+ "entity_type": "PHONE",
329
+ "surface": "09126483978",
330
+ "language": "en",
331
+ "field": null
332
+ },
333
+ {
334
+ "start": 79,
335
+ "end": 89,
336
+ "entity_type": "EMPLOYEE_ID",
337
+ "surface": "ARIA-24009",
338
+ "language": "en",
339
+ "field": null
340
+ }
341
+ ],
342
+ "meta": []
343
+ },
344
+ {
345
+ "doc_id": "DLP-EN-EMAIL-10",
346
+ "spans": [
347
+ {
348
+ "start": 6,
349
+ "end": 18,
350
+ "entity_type": "PERSON",
351
+ "surface": "Leila Labian",
352
+ "language": "en",
353
+ "field": null
354
+ },
355
+ {
356
+ "start": 20,
357
+ "end": 48,
358
+ "entity_type": "EMAIL",
359
+ "surface": "Leila.Labian10@aria-lab.test",
360
+ "language": "en",
361
+ "field": null
362
+ },
363
+ {
364
+ "start": 57,
365
+ "end": 68,
366
+ "entity_type": "PHONE",
367
+ "surface": "09127157194",
368
+ "language": "en",
369
+ "field": null
370
+ },
371
+ {
372
+ "start": 79,
373
+ "end": 89,
374
+ "entity_type": "EMPLOYEE_ID",
375
+ "surface": "ARIA-24010",
376
+ "language": "en",
377
+ "field": null
378
+ }
379
+ ],
380
+ "meta": []
381
+ },
382
+ {
383
+ "doc_id": "DLP-FA-HR-01",
384
+ "spans": [
385
+ {
386
+ "start": 5,
387
+ "end": 18,
388
+ "entity_type": "PERSON",
389
+ "surface": "سارا نمونه‌ای",
390
+ "language": "fa",
391
+ "field": null
392
+ },
393
+ {
394
+ "start": 27,
395
+ "end": 37,
396
+ "entity_type": "IR_NATIONAL_ID",
397
+ "surface": "۵۵۱۴۱۳۹۲۷۱",
398
+ "language": "fa",
399
+ "field": null
400
+ },
401
+ {
402
+ "start": 49,
403
+ "end": 60,
404
+ "entity_type": "PHONE",
405
+ "surface": "۰۹۱۲۵۷۶۷۰۹۶",
406
+ "language": "fa",
407
+ "field": null
408
+ },
409
+ {
410
+ "start": 67,
411
+ "end": 88,
412
+ "entity_type": "EMAIL",
413
+ "surface": "user.01@aria-lab.test",
414
+ "language": "en",
415
+ "field": null
416
+ }
417
+ ],
418
+ "meta": []
419
+ },
420
+ {
421
+ "doc_id": "DLP-FA-HR-02",
422
+ "spans": [
423
+ {
424
+ "start": 5,
425
+ "end": 17,
426
+ "entity_type": "PERSON",
427
+ "surface": "لیلا آزمایشی",
428
+ "language": "fa",
429
+ "field": null
430
+ },
431
+ {
432
+ "start": 26,
433
+ "end": 36,
434
+ "entity_type": "IR_NATIONAL_ID",
435
+ "surface": "۳۱۹۷۲۲۵۷۳۲",
436
+ "language": "fa",
437
+ "field": null
438
+ },
439
+ {
440
+ "start": 48,
441
+ "end": 59,
442
+ "entity_type": "PHONE",
443
+ "surface": "۰۹۱۲۳۲۰۳۶۶۹",
444
+ "language": "fa",
445
+ "field": null
446
+ },
447
+ {
448
+ "start": 66,
449
+ "end": 87,
450
+ "entity_type": "EMAIL",
451
+ "surface": "user.02@aria-lab.test",
452
+ "language": "en",
453
+ "field": null
454
+ }
455
+ ],
456
+ "meta": []
457
+ },
458
+ {
459
+ "doc_id": "DLP-FA-HR-03",
460
+ "spans": [
461
+ {
462
+ "start": 5,
463
+ "end": 17,
464
+ "entity_type": "PERSON",
465
+ "surface": "امید آزمایشی",
466
+ "language": "fa",
467
+ "field": null
468
+ },
469
+ {
470
+ "start": 26,
471
+ "end": 36,
472
+ "entity_type": "IR_NATIONAL_ID",
473
+ "surface": "۸۳۸۳۱۲۴۵۸۹",
474
+ "language": "fa",
475
+ "field": null
476
+ },
477
+ {
478
+ "start": 48,
479
+ "end": 59,
480
+ "entity_type": "PHONE",
481
+ "surface": "۰۹۱۲۴۸۶۵۴۴۱",
482
+ "language": "fa",
483
+ "field": null
484
+ },
485
+ {
486
+ "start": 66,
487
+ "end": 87,
488
+ "entity_type": "EMAIL",
489
+ "surface": "user.03@aria-lab.test",
490
+ "language": "en",
491
+ "field": null
492
+ }
493
+ ],
494
+ "meta": []
495
+ },
496
+ {
497
+ "doc_id": "DLP-FA-HR-04",
498
+ "spans": [
499
+ {
500
+ "start": 5,
501
+ "end": 18,
502
+ "entity_type": "PERSON",
503
+ "surface": "امید نمونه‌ای",
504
+ "language": "fa",
505
+ "field": null
506
+ },
507
+ {
508
+ "start": 27,
509
+ "end": 37,
510
+ "entity_type": "IR_NATIONAL_ID",
511
+ "surface": "۴۶۱۷۵۵۴۲۹۷",
512
+ "language": "fa",
513
+ "field": null
514
+ },
515
+ {
516
+ "start": 49,
517
+ "end": 60,
518
+ "entity_type": "PHONE",
519
+ "surface": "۰۹۱۲۸۲۹۱۰۰۲",
520
+ "language": "fa",
521
+ "field": null
522
+ },
523
+ {
524
+ "start": 67,
525
+ "end": 88,
526
+ "entity_type": "EMAIL",
527
+ "surface": "user.04@aria-lab.test",
528
+ "language": "en",
529
+ "field": null
530
+ }
531
+ ],
532
+ "meta": []
533
+ },
534
+ {
535
+ "doc_id": "DLP-FA-HR-05",
536
+ "spans": [
537
+ {
538
+ "start": 5,
539
+ "end": 18,
540
+ "entity_type": "PERSON",
541
+ "surface": "لیلا نمونه‌ای",
542
+ "language": "fa",
543
+ "field": null
544
+ },
545
+ {
546
+ "start": 27,
547
+ "end": 37,
548
+ "entity_type": "IR_NATIONAL_ID",
549
+ "surface": "۵۰۰۷۵۷۷۱۰۳",
550
+ "language": "fa",
551
+ "field": null
552
+ },
553
+ {
554
+ "start": 49,
555
+ "end": 60,
556
+ "entity_type": "PHONE",
557
+ "surface": "۰۹۱۲۲۳۶۷۲۱۴",
558
+ "language": "fa",
559
+ "field": null
560
+ },
561
+ {
562
+ "start": 67,
563
+ "end": 88,
564
+ "entity_type": "EMAIL",
565
+ "surface": "user.05@aria-lab.test",
566
+ "language": "en",
567
+ "field": null
568
+ }
569
+ ],
570
+ "meta": []
571
+ },
572
+ {
573
+ "doc_id": "DLP-FA-HR-06",
574
+ "spans": [
575
+ {
576
+ "start": 5,
577
+ "end": 17,
578
+ "entity_type": "PERSON",
579
+ "surface": "لیلا آزمایشی",
580
+ "language": "fa",
581
+ "field": null
582
+ },
583
+ {
584
+ "start": 26,
585
+ "end": 36,
586
+ "entity_type": "IR_NATIONAL_ID",
587
+ "surface": "۶۸۷۹۸۶۶۳۴۴",
588
+ "language": "fa",
589
+ "field": null
590
+ },
591
+ {
592
+ "start": 48,
593
+ "end": 59,
594
+ "entity_type": "PHONE",
595
+ "surface": "۰۹۱۲۶۰۸۳۲۸۳",
596
+ "language": "fa",
597
+ "field": null
598
+ },
599
+ {
600
+ "start": 66,
601
+ "end": 87,
602
+ "entity_type": "EMAIL",
603
+ "surface": "user.06@aria-lab.test",
604
+ "language": "en",
605
+ "field": null
606
+ }
607
+ ],
608
+ "meta": []
609
+ },
610
+ {
611
+ "doc_id": "DLP-FA-HR-07",
612
+ "spans": [
613
+ {
614
+ "start": 5,
615
+ "end": 17,
616
+ "entity_type": "PERSON",
617
+ "surface": "لیلا آزمایشی",
618
+ "language": "fa",
619
+ "field": null
620
+ },
621
+ {
622
+ "start": 26,
623
+ "end": 36,
624
+ "entity_type": "IR_NATIONAL_ID",
625
+ "surface": "۲۵۱۹۰۴۸۰۶۹",
626
+ "language": "fa",
627
+ "field": null
628
+ },
629
+ {
630
+ "start": 48,
631
+ "end": 59,
632
+ "entity_type": "PHONE",
633
+ "surface": "۰۹۱۲۷۷۱۴۶۰۱",
634
+ "language": "fa",
635
+ "field": null
636
+ },
637
+ {
638
+ "start": 66,
639
+ "end": 87,
640
+ "entity_type": "EMAIL",
641
+ "surface": "user.07@aria-lab.test",
642
+ "language": "en",
643
+ "field": null
644
+ }
645
+ ],
646
+ "meta": []
647
+ },
648
+ {
649
+ "doc_id": "DLP-FA-HR-08",
650
+ "spans": [
651
+ {
652
+ "start": 5,
653
+ "end": 17,
654
+ "entity_type": "PERSON",
655
+ "surface": "امید آزمایشی",
656
+ "language": "fa",
657
+ "field": null
658
+ },
659
+ {
660
+ "start": 26,
661
+ "end": 36,
662
+ "entity_type": "IR_NATIONAL_ID",
663
+ "surface": "۴۷۶۵۵۰۶۰۷۱",
664
+ "language": "fa",
665
+ "field": null
666
+ },
667
+ {
668
+ "start": 48,
669
+ "end": 59,
670
+ "entity_type": "PHONE",
671
+ "surface": "۰۹۱۲۸۷۶۰۹۰۱",
672
+ "language": "fa",
673
+ "field": null
674
+ },
675
+ {
676
+ "start": 66,
677
+ "end": 87,
678
+ "entity_type": "EMAIL",
679
+ "surface": "user.08@aria-lab.test",
680
+ "language": "en",
681
+ "field": null
682
+ }
683
+ ],
684
+ "meta": []
685
+ },
686
+ {
687
+ "doc_id": "DLP-FA-HR-09",
688
+ "spans": [
689
+ {
690
+ "start": 5,
691
+ "end": 18,
692
+ "entity_type": "PERSON",
693
+ "surface": "سارا نمونه‌ای",
694
+ "language": "fa",
695
+ "field": null
696
+ },
697
+ {
698
+ "start": 27,
699
+ "end": 37,
700
+ "entity_type": "IR_NATIONAL_ID",
701
+ "surface": "۳۲۸۹۳۷۲۲۰۰",
702
+ "language": "fa",
703
+ "field": null
704
+ },
705
+ {
706
+ "start": 49,
707
+ "end": 60,
708
+ "entity_type": "PHONE",
709
+ "surface": "۰۹۱۲۲۸۹۳۲۸۹",
710
+ "language": "fa",
711
+ "field": null
712
+ },
713
+ {
714
+ "start": 67,
715
+ "end": 88,
716
+ "entity_type": "EMAIL",
717
+ "surface": "user.09@aria-lab.test",
718
+ "language": "en",
719
+ "field": null
720
+ }
721
+ ],
722
+ "meta": []
723
+ },
724
+ {
725
+ "doc_id": "DLP-FA-HR-10",
726
+ "spans": [
727
+ {
728
+ "start": 5,
729
+ "end": 18,
730
+ "entity_type": "PERSON",
731
+ "surface": "سارا نمونه‌ای",
732
+ "language": "fa",
733
+ "field": null
734
+ },
735
+ {
736
+ "start": 27,
737
+ "end": 37,
738
+ "entity_type": "IR_NATIONAL_ID",
739
+ "surface": "۳۰۸۷۰۲۵۴۱۰",
740
+ "language": "fa",
741
+ "field": null
742
+ },
743
+ {
744
+ "start": 49,
745
+ "end": 60,
746
+ "entity_type": "PHONE",
747
+ "surface": "۰۹۱۲۴۰۶۷۵۴۸",
748
+ "language": "fa",
749
+ "field": null
750
+ },
751
+ {
752
+ "start": 67,
753
+ "end": 88,
754
+ "entity_type": "EMAIL",
755
+ "surface": "user.10@aria-lab.test",
756
+ "language": "en",
757
+ "field": null
758
+ }
759
+ ],
760
+ "meta": []
761
+ },
762
+ {
763
+ "doc_id": "DLP-CSV-INV-01",
764
+ "spans": [
765
+ {
766
+ "start": 53,
767
+ "end": 63,
768
+ "entity_type": "PERSON",
769
+ "surface": "Ava Labian",
770
+ "language": "en",
771
+ "field": null
772
+ },
773
+ {
774
+ "start": 64,
775
+ "end": 90,
776
+ "entity_type": "EMAIL",
777
+ "surface": "Ava.Labian01@aria-lab.test",
778
+ "language": "en",
779
+ "field": null
780
+ },
781
+ {
782
+ "start": 91,
783
+ "end": 102,
784
+ "entity_type": "PHONE",
785
+ "surface": "09129456762",
786
+ "language": "en",
787
+ "field": null
788
+ }
789
+ ],
790
+ "meta": []
791
+ },
792
+ {
793
+ "doc_id": "DLP-CSV-INV-02",
794
+ "spans": [
795
+ {
796
+ "start": 53,
797
+ "end": 65,
798
+ "entity_type": "PERSON",
799
+ "surface": "Nima Fixture",
800
+ "language": "en",
801
+ "field": null
802
+ },
803
+ {
804
+ "start": 66,
805
+ "end": 94,
806
+ "entity_type": "EMAIL",
807
+ "surface": "Nima.Fixture02@aria-lab.test",
808
+ "language": "en",
809
+ "field": null
810
+ },
811
+ {
812
+ "start": 95,
813
+ "end": 106,
814
+ "entity_type": "PHONE",
815
+ "surface": "09127891210",
816
+ "language": "en",
817
+ "field": null
818
+ }
819
+ ],
820
+ "meta": []
821
+ },
822
+ {
823
+ "doc_id": "DLP-CSV-INV-03",
824
+ "spans": [
825
+ {
826
+ "start": 53,
827
+ "end": 66,
828
+ "entity_type": "PERSON",
829
+ "surface": "Leila Fixture",
830
+ "language": "en",
831
+ "field": null
832
+ },
833
+ {
834
+ "start": 67,
835
+ "end": 96,
836
+ "entity_type": "EMAIL",
837
+ "surface": "Leila.Fixture03@aria-lab.test",
838
+ "language": "en",
839
+ "field": null
840
+ },
841
+ {
842
+ "start": 97,
843
+ "end": 108,
844
+ "entity_type": "PHONE",
845
+ "surface": "09124895575",
846
+ "language": "en",
847
+ "field": null
848
+ }
849
+ ],
850
+ "meta": []
851
+ },
852
+ {
853
+ "doc_id": "DLP-CSV-INV-04",
854
+ "spans": [
855
+ {
856
+ "start": 53,
857
+ "end": 64,
858
+ "entity_type": "PERSON",
859
+ "surface": "Omar Labian",
860
+ "language": "en",
861
+ "field": null
862
+ },
863
+ {
864
+ "start": 65,
865
+ "end": 92,
866
+ "entity_type": "EMAIL",
867
+ "surface": "Omar.Labian04@aria-lab.test",
868
+ "language": "en",
869
+ "field": null
870
+ },
871
+ {
872
+ "start": 93,
873
+ "end": 104,
874
+ "entity_type": "PHONE",
875
+ "surface": "09124151655",
876
+ "language": "en",
877
+ "field": null
878
+ }
879
+ ],
880
+ "meta": []
881
+ },
882
+ {
883
+ "doc_id": "DLP-CSV-INV-05",
884
+ "spans": [
885
+ {
886
+ "start": 53,
887
+ "end": 63,
888
+ "entity_type": "PERSON",
889
+ "surface": "Ava Labian",
890
+ "language": "en",
891
+ "field": null
892
+ },
893
+ {
894
+ "start": 64,
895
+ "end": 90,
896
+ "entity_type": "EMAIL",
897
+ "surface": "Ava.Labian05@aria-lab.test",
898
+ "language": "en",
899
+ "field": null
900
+ },
901
+ {
902
+ "start": 91,
903
+ "end": 102,
904
+ "entity_type": "PHONE",
905
+ "surface": "09123831675",
906
+ "language": "en",
907
+ "field": null
908
+ }
909
+ ],
910
+ "meta": []
911
+ },
912
+ {
913
+ "doc_id": "DLP-CSV-INV-06",
914
+ "spans": [
915
+ {
916
+ "start": 53,
917
+ "end": 65,
918
+ "entity_type": "PERSON",
919
+ "surface": "Noah Fixture",
920
+ "language": "en",
921
+ "field": null
922
+ },
923
+ {
924
+ "start": 66,
925
+ "end": 94,
926
+ "entity_type": "EMAIL",
927
+ "surface": "Noah.Fixture06@aria-lab.test",
928
+ "language": "en",
929
+ "field": null
930
+ },
931
+ {
932
+ "start": 95,
933
+ "end": 106,
934
+ "entity_type": "PHONE",
935
+ "surface": "09125428133",
936
+ "language": "en",
937
+ "field": null
938
+ }
939
+ ],
940
+ "meta": []
941
+ },
942
+ {
943
+ "doc_id": "DLP-CSV-INV-07",
944
+ "spans": [
945
+ {
946
+ "start": 53,
947
+ "end": 66,
948
+ "entity_type": "PERSON",
949
+ "surface": "Helen Fixture",
950
+ "language": "en",
951
+ "field": null
952
+ },
953
+ {
954
+ "start": 67,
955
+ "end": 96,
956
+ "entity_type": "EMAIL",
957
+ "surface": "Helen.Fixture07@aria-lab.test",
958
+ "language": "en",
959
+ "field": null
960
+ },
961
+ {
962
+ "start": 97,
963
+ "end": 108,
964
+ "entity_type": "PHONE",
965
+ "surface": "09124672166",
966
+ "language": "en",
967
+ "field": null
968
+ }
969
+ ],
970
+ "meta": []
971
+ },
972
+ {
973
+ "doc_id": "DLP-CSV-INV-08",
974
+ "spans": [
975
+ {
976
+ "start": 53,
977
+ "end": 65,
978
+ "entity_type": "PERSON",
979
+ "surface": "Omar Fixture",
980
+ "language": "en",
981
+ "field": null
982
+ },
983
+ {
984
+ "start": 66,
985
+ "end": 94,
986
+ "entity_type": "EMAIL",
987
+ "surface": "Omar.Fixture08@aria-lab.test",
988
+ "language": "en",
989
+ "field": null
990
+ },
991
+ {
992
+ "start": 95,
993
+ "end": 106,
994
+ "entity_type": "PHONE",
995
+ "surface": "09128527564",
996
+ "language": "en",
997
+ "field": null
998
+ }
999
+ ],
1000
+ "meta": []
1001
+ },
1002
+ {
1003
+ "doc_id": "DLP-CSV-INV-09",
1004
+ "spans": [
1005
+ {
1006
+ "start": 53,
1007
+ "end": 65,
1008
+ "entity_type": "PERSON",
1009
+ "surface": "Leila Labian",
1010
+ "language": "en",
1011
+ "field": null
1012
+ },
1013
+ {
1014
+ "start": 66,
1015
+ "end": 94,
1016
+ "entity_type": "EMAIL",
1017
+ "surface": "Leila.Labian09@aria-lab.test",
1018
+ "language": "en",
1019
+ "field": null
1020
+ },
1021
+ {
1022
+ "start": 95,
1023
+ "end": 106,
1024
+ "entity_type": "PHONE",
1025
+ "surface": "09127031590",
1026
+ "language": "en",
1027
+ "field": null
1028
+ }
1029
+ ],
1030
+ "meta": []
1031
+ },
1032
+ {
1033
+ "doc_id": "DLP-CSV-INV-10",
1034
+ "spans": [
1035
+ {
1036
+ "start": 53,
1037
+ "end": 66,
1038
+ "entity_type": "PERSON",
1039
+ "surface": "Leila Fixture",
1040
+ "language": "en",
1041
+ "field": null
1042
+ },
1043
+ {
1044
+ "start": 67,
1045
+ "end": 96,
1046
+ "entity_type": "EMAIL",
1047
+ "surface": "Leila.Fixture10@aria-lab.test",
1048
+ "language": "en",
1049
+ "field": null
1050
+ },
1051
+ {
1052
+ "start": 97,
1053
+ "end": 108,
1054
+ "entity_type": "PHONE",
1055
+ "surface": "09123631859",
1056
+ "language": "en",
1057
+ "field": null
1058
+ }
1059
+ ],
1060
+ "meta": []
1061
+ },
1062
+ {
1063
+ "doc_id": "DLP-EN-ONB-01",
1064
+ "spans": [
1065
+ {
1066
+ "start": 6,
1067
+ "end": 18,
1068
+ "entity_type": "PERSON",
1069
+ "surface": "Helen Labian",
1070
+ "language": "en",
1071
+ "field": null
1072
+ },
1073
+ {
1074
+ "start": 31,
1075
+ "end": 59,
1076
+ "entity_type": "EMAIL",
1077
+ "surface": "Helen.Labian01@aria-lab.test",
1078
+ "language": "en",
1079
+ "field": null
1080
+ },
1081
+ {
1082
+ "start": 85,
1083
+ "end": 101,
1084
+ "entity_type": "CREDIT_CARD",
1085
+ "surface": "4111137038772834",
1086
+ "language": "en",
1087
+ "field": null
1088
+ },
1089
+ {
1090
+ "start": 116,
1091
+ "end": 142,
1092
+ "entity_type": "IBAN",
1093
+ "surface": "IR308660110372211081968590",
1094
+ "language": "en",
1095
+ "field": null
1096
+ }
1097
+ ],
1098
+ "meta": []
1099
+ },
1100
+ {
1101
+ "doc_id": "DLP-EN-ONB-02",
1102
+ "spans": [
1103
+ {
1104
+ "start": 6,
1105
+ "end": 18,
1106
+ "entity_type": "PERSON",
1107
+ "surface": "Helen Labian",
1108
+ "language": "en",
1109
+ "field": null
1110
+ },
1111
+ {
1112
+ "start": 31,
1113
+ "end": 59,
1114
+ "entity_type": "EMAIL",
1115
+ "surface": "Helen.Labian02@aria-lab.test",
1116
+ "language": "en",
1117
+ "field": null
1118
+ },
1119
+ {
1120
+ "start": 85,
1121
+ "end": 101,
1122
+ "entity_type": "CREDIT_CARD",
1123
+ "surface": "4111529393529637",
1124
+ "language": "en",
1125
+ "field": null
1126
+ },
1127
+ {
1128
+ "start": 116,
1129
+ "end": 142,
1130
+ "entity_type": "IBAN",
1131
+ "surface": "IR044280439609406102122129",
1132
+ "language": "en",
1133
+ "field": null
1134
+ }
1135
+ ],
1136
+ "meta": []
1137
+ },
1138
+ {
1139
+ "doc_id": "DLP-EN-ONB-03",
1140
+ "spans": [
1141
+ {
1142
+ "start": 6,
1143
+ "end": 18,
1144
+ "entity_type": "PERSON",
1145
+ "surface": "Helen Labian",
1146
+ "language": "en",
1147
+ "field": null
1148
+ },
1149
+ {
1150
+ "start": 31,
1151
+ "end": 59,
1152
+ "entity_type": "EMAIL",
1153
+ "surface": "Helen.Labian03@aria-lab.test",
1154
+ "language": "en",
1155
+ "field": null
1156
+ },
1157
+ {
1158
+ "start": 85,
1159
+ "end": 101,
1160
+ "entity_type": "CREDIT_CARD",
1161
+ "surface": "4111785437328044",
1162
+ "language": "en",
1163
+ "field": null
1164
+ },
1165
+ {
1166
+ "start": 116,
1167
+ "end": 142,
1168
+ "entity_type": "IBAN",
1169
+ "surface": "IR019630083854359765553642",
1170
+ "language": "en",
1171
+ "field": null
1172
+ }
1173
+ ],
1174
+ "meta": []
1175
+ },
1176
+ {
1177
+ "doc_id": "DLP-EN-ONB-04",
1178
+ "spans": [
1179
+ {
1180
+ "start": 6,
1181
+ "end": 18,
1182
+ "entity_type": "PERSON",
1183
+ "surface": "Helen Labian",
1184
+ "language": "en",
1185
+ "field": null
1186
+ },
1187
+ {
1188
+ "start": 31,
1189
+ "end": 59,
1190
+ "entity_type": "EMAIL",
1191
+ "surface": "Helen.Labian04@aria-lab.test",
1192
+ "language": "en",
1193
+ "field": null
1194
+ },
1195
+ {
1196
+ "start": 85,
1197
+ "end": 101,
1198
+ "entity_type": "CREDIT_CARD",
1199
+ "surface": "4111759320597908",
1200
+ "language": "en",
1201
+ "field": null
1202
+ },
1203
+ {
1204
+ "start": 116,
1205
+ "end": 142,
1206
+ "entity_type": "IBAN",
1207
+ "surface": "IR472630756281276293569333",
1208
+ "language": "en",
1209
+ "field": null
1210
+ }
1211
+ ],
1212
+ "meta": []
1213
+ },
1214
+ {
1215
+ "doc_id": "DLP-EN-ONB-05",
1216
+ "spans": [
1217
+ {
1218
+ "start": 6,
1219
+ "end": 19,
1220
+ "entity_type": "PERSON",
1221
+ "surface": "Leila Fixture",
1222
+ "language": "en",
1223
+ "field": null
1224
+ },
1225
+ {
1226
+ "start": 32,
1227
+ "end": 61,
1228
+ "entity_type": "EMAIL",
1229
+ "surface": "Leila.Fixture05@aria-lab.test",
1230
+ "language": "en",
1231
+ "field": null
1232
+ },
1233
+ {
1234
+ "start": 87,
1235
+ "end": 103,
1236
+ "entity_type": "CREDIT_CARD",
1237
+ "surface": "4111313694373346",
1238
+ "language": "en",
1239
+ "field": null
1240
+ },
1241
+ {
1242
+ "start": 118,
1243
+ "end": 144,
1244
+ "entity_type": "IBAN",
1245
+ "surface": "IR466530650673099236853917",
1246
+ "language": "en",
1247
+ "field": null
1248
+ }
1249
+ ],
1250
+ "meta": []
1251
+ },
1252
+ {
1253
+ "doc_id": "DLP-EN-ONB-06",
1254
+ "spans": [
1255
+ {
1256
+ "start": 6,
1257
+ "end": 17,
1258
+ "entity_type": "PERSON",
1259
+ "surface": "Nima Labian",
1260
+ "language": "en",
1261
+ "field": null
1262
+ },
1263
+ {
1264
+ "start": 30,
1265
+ "end": 57,
1266
+ "entity_type": "EMAIL",
1267
+ "surface": "Nima.Labian06@aria-lab.test",
1268
+ "language": "en",
1269
+ "field": null
1270
+ },
1271
+ {
1272
+ "start": 83,
1273
+ "end": 99,
1274
+ "entity_type": "CREDIT_CARD",
1275
+ "surface": "4111814102596287",
1276
+ "language": "en",
1277
+ "field": null
1278
+ },
1279
+ {
1280
+ "start": 114,
1281
+ "end": 140,
1282
+ "entity_type": "IBAN",
1283
+ "surface": "IR581060722681444648227055",
1284
+ "language": "en",
1285
+ "field": null
1286
+ }
1287
+ ],
1288
+ "meta": []
1289
+ },
1290
+ {
1291
+ "doc_id": "DLP-EN-ONB-07",
1292
+ "spans": [
1293
+ {
1294
+ "start": 6,
1295
+ "end": 18,
1296
+ "entity_type": "PERSON",
1297
+ "surface": "Noah Fixture",
1298
+ "language": "en",
1299
+ "field": null
1300
+ },
1301
+ {
1302
+ "start": 31,
1303
+ "end": 59,
1304
+ "entity_type": "EMAIL",
1305
+ "surface": "Noah.Fixture07@aria-lab.test",
1306
+ "language": "en",
1307
+ "field": null
1308
+ },
1309
+ {
1310
+ "start": 85,
1311
+ "end": 101,
1312
+ "entity_type": "CREDIT_CARD",
1313
+ "surface": "4111842413164297",
1314
+ "language": "en",
1315
+ "field": null
1316
+ },
1317
+ {
1318
+ "start": 116,
1319
+ "end": 142,
1320
+ "entity_type": "IBAN",
1321
+ "surface": "IR895120978818064078339048",
1322
+ "language": "en",
1323
+ "field": null
1324
+ }
1325
+ ],
1326
+ "meta": []
1327
+ },
1328
+ {
1329
+ "doc_id": "DLP-EN-ONB-08",
1330
+ "spans": [
1331
+ {
1332
+ "start": 6,
1333
+ "end": 18,
1334
+ "entity_type": "PERSON",
1335
+ "surface": "Noah Fixture",
1336
+ "language": "en",
1337
+ "field": null
1338
+ },
1339
+ {
1340
+ "start": 31,
1341
+ "end": 59,
1342
+ "entity_type": "EMAIL",
1343
+ "surface": "Noah.Fixture08@aria-lab.test",
1344
+ "language": "en",
1345
+ "field": null
1346
+ },
1347
+ {
1348
+ "start": 85,
1349
+ "end": 101,
1350
+ "entity_type": "CREDIT_CARD",
1351
+ "surface": "4111329710194664",
1352
+ "language": "en",
1353
+ "field": null
1354
+ },
1355
+ {
1356
+ "start": 116,
1357
+ "end": 142,
1358
+ "entity_type": "IBAN",
1359
+ "surface": "IR237680114663394226281259",
1360
+ "language": "en",
1361
+ "field": null
1362
+ }
1363
+ ],
1364
+ "meta": []
1365
+ },
1366
+ {
1367
+ "doc_id": "DLP-EN-ONB-09",
1368
+ "spans": [
1369
+ {
1370
+ "start": 6,
1371
+ "end": 19,
1372
+ "entity_type": "PERSON",
1373
+ "surface": "Leila Fixture",
1374
+ "language": "en",
1375
+ "field": null
1376
+ },
1377
+ {
1378
+ "start": 32,
1379
+ "end": 61,
1380
+ "entity_type": "EMAIL",
1381
+ "surface": "Leila.Fixture09@aria-lab.test",
1382
+ "language": "en",
1383
+ "field": null
1384
+ },
1385
+ {
1386
+ "start": 87,
1387
+ "end": 103,
1388
+ "entity_type": "CREDIT_CARD",
1389
+ "surface": "4111996662241709",
1390
+ "language": "en",
1391
+ "field": null
1392
+ },
1393
+ {
1394
+ "start": 118,
1395
+ "end": 144,
1396
+ "entity_type": "IBAN",
1397
+ "surface": "IR772810906854799104860878",
1398
+ "language": "en",
1399
+ "field": null
1400
+ }
1401
+ ],
1402
+ "meta": []
1403
+ },
1404
+ {
1405
+ "doc_id": "DLP-EN-ONB-10",
1406
+ "spans": [
1407
+ {
1408
+ "start": 6,
1409
+ "end": 18,
1410
+ "entity_type": "PERSON",
1411
+ "surface": "Nima Fixture",
1412
+ "language": "en",
1413
+ "field": null
1414
+ },
1415
+ {
1416
+ "start": 31,
1417
+ "end": 59,
1418
+ "entity_type": "EMAIL",
1419
+ "surface": "Nima.Fixture10@aria-lab.test",
1420
+ "language": "en",
1421
+ "field": null
1422
+ },
1423
+ {
1424
+ "start": 85,
1425
+ "end": 101,
1426
+ "entity_type": "CREDIT_CARD",
1427
+ "surface": "4111041294625191",
1428
+ "language": "en",
1429
+ "field": null
1430
+ },
1431
+ {
1432
+ "start": 116,
1433
+ "end": 142,
1434
+ "entity_type": "IBAN",
1435
+ "surface": "IR338550745309562340114615",
1436
+ "language": "en",
1437
+ "field": null
1438
+ }
1439
+ ],
1440
+ "meta": []
1441
+ },
1442
+ {
1443
+ "doc_id": "DLP-FA-CTR-01",
1444
+ "spans": [
1445
+ {
1446
+ "start": 27,
1447
+ "end": 39,
1448
+ "entity_type": "PERSON",
1449
+ "surface": "لیلا آزمایشی",
1450
+ "language": "fa",
1451
+ "field": null
1452
+ },
1453
+ {
1454
+ "start": 47,
1455
+ "end": 57,
1456
+ "entity_type": "IR_NATIONAL_ID",
1457
+ "surface": "4985528323",
1458
+ "language": "fa",
1459
+ "field": null
1460
+ },
1461
+ {
1462
+ "start": 63,
1463
+ "end": 74,
1464
+ "entity_type": "PHONE",
1465
+ "surface": "۰۹۱۲۳۴۲۱۵۱۴",
1466
+ "language": "fa",
1467
+ "field": null
1468
+ }
1469
+ ],
1470
+ "meta": []
1471
+ },
1472
+ {
1473
+ "doc_id": "DLP-FA-CTR-02",
1474
+ "spans": [
1475
+ {
1476
+ "start": 27,
1477
+ "end": 39,
1478
+ "entity_type": "PERSON",
1479
+ "surface": "سارا آزمایشی",
1480
+ "language": "fa",
1481
+ "field": null
1482
+ },
1483
+ {
1484
+ "start": 47,
1485
+ "end": 57,
1486
+ "entity_type": "IR_NATIONAL_ID",
1487
+ "surface": "6361423751",
1488
+ "language": "fa",
1489
+ "field": null
1490
+ },
1491
+ {
1492
+ "start": 63,
1493
+ "end": 74,
1494
+ "entity_type": "PHONE",
1495
+ "surface": "۰۹۱۲۳۶۳۰۹۶۹",
1496
+ "language": "fa",
1497
+ "field": null
1498
+ }
1499
+ ],
1500
+ "meta": []
1501
+ },
1502
+ {
1503
+ "doc_id": "DLP-FA-CTR-03",
1504
+ "spans": [
1505
+ {
1506
+ "start": 27,
1507
+ "end": 40,
1508
+ "entity_type": "PERSON",
1509
+ "surface": "نیما نمونه‌ای",
1510
+ "language": "fa",
1511
+ "field": null
1512
+ },
1513
+ {
1514
+ "start": 48,
1515
+ "end": 58,
1516
+ "entity_type": "IR_NATIONAL_ID",
1517
+ "surface": "3589737379",
1518
+ "language": "fa",
1519
+ "field": null
1520
+ },
1521
+ {
1522
+ "start": 64,
1523
+ "end": 75,
1524
+ "entity_type": "PHONE",
1525
+ "surface": "۰۹۱۲۵۹۲۸۹۵۰",
1526
+ "language": "fa",
1527
+ "field": null
1528
+ }
1529
+ ],
1530
+ "meta": []
1531
+ },
1532
+ {
1533
+ "doc_id": "DLP-FA-CTR-04",
1534
+ "spans": [
1535
+ {
1536
+ "start": 27,
1537
+ "end": 39,
1538
+ "entity_type": "PERSON",
1539
+ "surface": "نیما آزمایشی",
1540
+ "language": "fa",
1541
+ "field": null
1542
+ },
1543
+ {
1544
+ "start": 47,
1545
+ "end": 57,
1546
+ "entity_type": "IR_NATIONAL_ID",
1547
+ "surface": "2341457452",
1548
+ "language": "fa",
1549
+ "field": null
1550
+ },
1551
+ {
1552
+ "start": 63,
1553
+ "end": 74,
1554
+ "entity_type": "PHONE",
1555
+ "surface": "۰۹۱۲۲۳۶۵۹۰۷",
1556
+ "language": "fa",
1557
+ "field": null
1558
+ }
1559
+ ],
1560
+ "meta": []
1561
+ },
1562
+ {
1563
+ "doc_id": "DLP-FA-CTR-05",
1564
+ "spans": [
1565
+ {
1566
+ "start": 27,
1567
+ "end": 40,
1568
+ "entity_type": "PERSON",
1569
+ "surface": "نیما نمونه‌ای",
1570
+ "language": "fa",
1571
+ "field": null
1572
+ },
1573
+ {
1574
+ "start": 48,
1575
+ "end": 58,
1576
+ "entity_type": "IR_NATIONAL_ID",
1577
+ "surface": "4061666126",
1578
+ "language": "fa",
1579
+ "field": null
1580
+ },
1581
+ {
1582
+ "start": 64,
1583
+ "end": 75,
1584
+ "entity_type": "PHONE",
1585
+ "surface": "۰۹۱۲۴۹۶۶۴۱۱",
1586
+ "language": "fa",
1587
+ "field": null
1588
+ }
1589
+ ],
1590
+ "meta": []
1591
+ },
1592
+ {
1593
+ "doc_id": "DLP-FA-CTR-06",
1594
+ "spans": [
1595
+ {
1596
+ "start": 27,
1597
+ "end": 40,
1598
+ "entity_type": "PERSON",
1599
+ "surface": "لیلا نمونه‌ای",
1600
+ "language": "fa",
1601
+ "field": null
1602
+ },
1603
+ {
1604
+ "start": 48,
1605
+ "end": 58,
1606
+ "entity_type": "IR_NATIONAL_ID",
1607
+ "surface": "3179136243",
1608
+ "language": "fa",
1609
+ "field": null
1610
+ },
1611
+ {
1612
+ "start": 64,
1613
+ "end": 75,
1614
+ "entity_type": "PHONE",
1615
+ "surface": "۰۹۱۲۸۱۳۵۴۲۶",
1616
+ "language": "fa",
1617
+ "field": null
1618
+ }
1619
+ ],
1620
+ "meta": []
1621
+ },
1622
+ {
1623
+ "doc_id": "DLP-FA-CTR-07",
1624
+ "spans": [
1625
+ {
1626
+ "start": 27,
1627
+ "end": 39,
1628
+ "entity_type": "PERSON",
1629
+ "surface": "امید آزمایشی",
1630
+ "language": "fa",
1631
+ "field": null
1632
+ },
1633
+ {
1634
+ "start": 47,
1635
+ "end": 57,
1636
+ "entity_type": "IR_NATIONAL_ID",
1637
+ "surface": "2154052452",
1638
+ "language": "fa",
1639
+ "field": null
1640
+ },
1641
+ {
1642
+ "start": 63,
1643
+ "end": 74,
1644
+ "entity_type": "PHONE",
1645
+ "surface": "۰۹۱۲۸۸۱۷۱۸۵",
1646
+ "language": "fa",
1647
+ "field": null
1648
+ }
1649
+ ],
1650
+ "meta": []
1651
+ },
1652
+ {
1653
+ "doc_id": "DLP-FA-CTR-08",
1654
+ "spans": [
1655
+ {
1656
+ "start": 27,
1657
+ "end": 39,
1658
+ "entity_type": "PERSON",
1659
+ "surface": "سارا آزمایشی",
1660
+ "language": "fa",
1661
+ "field": null
1662
+ },
1663
+ {
1664
+ "start": 47,
1665
+ "end": 57,
1666
+ "entity_type": "IR_NATIONAL_ID",
1667
+ "surface": "2313920291",
1668
+ "language": "fa",
1669
+ "field": null
1670
+ },
1671
+ {
1672
+ "start": 63,
1673
+ "end": 74,
1674
+ "entity_type": "PHONE",
1675
+ "surface": "۰۹۱۲۱۳۴۴۶۷۵",
1676
+ "language": "fa",
1677
+ "field": null
1678
+ }
1679
+ ],
1680
+ "meta": []
1681
+ },
1682
+ {
1683
+ "doc_id": "DLP-FA-CTR-09",
1684
+ "spans": [
1685
+ {
1686
+ "start": 27,
1687
+ "end": 40,
1688
+ "entity_type": "PERSON",
1689
+ "surface": "نیما نمونه‌ای",
1690
+ "language": "fa",
1691
+ "field": null
1692
+ },
1693
+ {
1694
+ "start": 48,
1695
+ "end": 58,
1696
+ "entity_type": "IR_NATIONAL_ID",
1697
+ "surface": "0190065516",
1698
+ "language": "fa",
1699
+ "field": null
1700
+ },
1701
+ {
1702
+ "start": 64,
1703
+ "end": 75,
1704
+ "entity_type": "PHONE",
1705
+ "surface": "۰۹۱۲۷۲۰۹۰۸۸",
1706
+ "language": "fa",
1707
+ "field": null
1708
+ }
1709
+ ],
1710
+ "meta": []
1711
+ },
1712
+ {
1713
+ "doc_id": "DLP-FA-CTR-10",
1714
+ "spans": [
1715
+ {
1716
+ "start": 27,
1717
+ "end": 39,
1718
+ "entity_type": "PERSON",
1719
+ "surface": "لیلا آزمایشی",
1720
+ "language": "fa",
1721
+ "field": null
1722
+ },
1723
+ {
1724
+ "start": 47,
1725
+ "end": 57,
1726
+ "entity_type": "IR_NATIONAL_ID",
1727
+ "surface": "8212633349",
1728
+ "language": "fa",
1729
+ "field": null
1730
+ },
1731
+ {
1732
+ "start": 63,
1733
+ "end": 74,
1734
+ "entity_type": "PHONE",
1735
+ "surface": "۰۹۱۲۷۴۸۴۳۶۷",
1736
+ "language": "fa",
1737
+ "field": null
1738
+ }
1739
+ ],
1740
+ "meta": []
1741
+ },
1742
+ {
1743
+ "doc_id": "DLP-PDF-MEMO-01",
1744
+ "spans": [
1745
+ {
1746
+ "start": 22,
1747
+ "end": 34,
1748
+ "entity_type": "PERSON",
1749
+ "surface": "Omar Fixture",
1750
+ "language": "en",
1751
+ "field": null
1752
+ },
1753
+ {
1754
+ "start": 44,
1755
+ "end": 72,
1756
+ "entity_type": "EMAIL",
1757
+ "surface": "Omar.Fixture01@aria-lab.test",
1758
+ "language": "en",
1759
+ "field": null
1760
+ },
1761
+ {
1762
+ "start": 75,
1763
+ "end": 86,
1764
+ "entity_type": "PHONE",
1765
+ "surface": "09127692383",
1766
+ "language": "en",
1767
+ "field": null
1768
+ }
1769
+ ],
1770
+ "meta": [
1771
+ {
1772
+ "field": "author",
1773
+ "entity_type": "PERSON",
1774
+ "surface": "Omar Fixture"
1775
+ },
1776
+ {
1777
+ "field": "keywords",
1778
+ "entity_type": "EMAIL",
1779
+ "surface": "Omar.Fixture01@aria-lab.test"
1780
+ }
1781
+ ]
1782
+ },
1783
+ {
1784
+ "doc_id": "DLP-PDF-MEMO-02",
1785
+ "spans": [
1786
+ {
1787
+ "start": 22,
1788
+ "end": 34,
1789
+ "entity_type": "PERSON",
1790
+ "surface": "Noah Fixture",
1791
+ "language": "en",
1792
+ "field": null
1793
+ },
1794
+ {
1795
+ "start": 44,
1796
+ "end": 72,
1797
+ "entity_type": "EMAIL",
1798
+ "surface": "Noah.Fixture02@aria-lab.test",
1799
+ "language": "en",
1800
+ "field": null
1801
+ },
1802
+ {
1803
+ "start": 75,
1804
+ "end": 86,
1805
+ "entity_type": "PHONE",
1806
+ "surface": "09125328205",
1807
+ "language": "en",
1808
+ "field": null
1809
+ }
1810
+ ],
1811
+ "meta": [
1812
+ {
1813
+ "field": "author",
1814
+ "entity_type": "PERSON",
1815
+ "surface": "Noah Fixture"
1816
+ },
1817
+ {
1818
+ "field": "keywords",
1819
+ "entity_type": "EMAIL",
1820
+ "surface": "Noah.Fixture02@aria-lab.test"
1821
+ }
1822
+ ]
1823
+ },
1824
+ {
1825
+ "doc_id": "DLP-PDF-MEMO-03",
1826
+ "spans": [
1827
+ {
1828
+ "start": 22,
1829
+ "end": 33,
1830
+ "entity_type": "PERSON",
1831
+ "surface": "Omar Labian",
1832
+ "language": "en",
1833
+ "field": null
1834
+ },
1835
+ {
1836
+ "start": 43,
1837
+ "end": 70,
1838
+ "entity_type": "EMAIL",
1839
+ "surface": "Omar.Labian03@aria-lab.test",
1840
+ "language": "en",
1841
+ "field": null
1842
+ },
1843
+ {
1844
+ "start": 73,
1845
+ "end": 84,
1846
+ "entity_type": "PHONE",
1847
+ "surface": "09125603855",
1848
+ "language": "en",
1849
+ "field": null
1850
+ }
1851
+ ],
1852
+ "meta": [
1853
+ {
1854
+ "field": "author",
1855
+ "entity_type": "PERSON",
1856
+ "surface": "Omar Labian"
1857
+ },
1858
+ {
1859
+ "field": "keywords",
1860
+ "entity_type": "EMAIL",
1861
+ "surface": "Omar.Labian03@aria-lab.test"
1862
+ }
1863
+ ]
1864
+ },
1865
+ {
1866
+ "doc_id": "DLP-PDF-MEMO-04",
1867
+ "spans": [
1868
+ {
1869
+ "start": 22,
1870
+ "end": 33,
1871
+ "entity_type": "PERSON",
1872
+ "surface": "Noah Labian",
1873
+ "language": "en",
1874
+ "field": null
1875
+ },
1876
+ {
1877
+ "start": 43,
1878
+ "end": 70,
1879
+ "entity_type": "EMAIL",
1880
+ "surface": "Noah.Labian04@aria-lab.test",
1881
+ "language": "en",
1882
+ "field": null
1883
+ },
1884
+ {
1885
+ "start": 73,
1886
+ "end": 84,
1887
+ "entity_type": "PHONE",
1888
+ "surface": "09127591902",
1889
+ "language": "en",
1890
+ "field": null
1891
+ }
1892
+ ],
1893
+ "meta": [
1894
+ {
1895
+ "field": "author",
1896
+ "entity_type": "PERSON",
1897
+ "surface": "Noah Labian"
1898
+ },
1899
+ {
1900
+ "field": "keywords",
1901
+ "entity_type": "EMAIL",
1902
+ "surface": "Noah.Labian04@aria-lab.test"
1903
+ }
1904
+ ]
1905
+ },
1906
+ {
1907
+ "doc_id": "DLP-PDF-MEMO-05",
1908
+ "spans": [
1909
+ {
1910
+ "start": 22,
1911
+ "end": 33,
1912
+ "entity_type": "PERSON",
1913
+ "surface": "Nima Labian",
1914
+ "language": "en",
1915
+ "field": null
1916
+ },
1917
+ {
1918
+ "start": 43,
1919
+ "end": 70,
1920
+ "entity_type": "EMAIL",
1921
+ "surface": "Nima.Labian05@aria-lab.test",
1922
+ "language": "en",
1923
+ "field": null
1924
+ },
1925
+ {
1926
+ "start": 73,
1927
+ "end": 84,
1928
+ "entity_type": "PHONE",
1929
+ "surface": "09129473675",
1930
+ "language": "en",
1931
+ "field": null
1932
+ }
1933
+ ],
1934
+ "meta": [
1935
+ {
1936
+ "field": "author",
1937
+ "entity_type": "PERSON",
1938
+ "surface": "Nima Labian"
1939
+ },
1940
+ {
1941
+ "field": "keywords",
1942
+ "entity_type": "EMAIL",
1943
+ "surface": "Nima.Labian05@aria-lab.test"
1944
+ }
1945
+ ]
1946
+ },
1947
+ {
1948
+ "doc_id": "DLP-PDF-MEMO-06",
1949
+ "spans": [
1950
+ {
1951
+ "start": 22,
1952
+ "end": 35,
1953
+ "entity_type": "PERSON",
1954
+ "surface": "Helen Fixture",
1955
+ "language": "en",
1956
+ "field": null
1957
+ },
1958
+ {
1959
+ "start": 45,
1960
+ "end": 74,
1961
+ "entity_type": "EMAIL",
1962
+ "surface": "Helen.Fixture06@aria-lab.test",
1963
+ "language": "en",
1964
+ "field": null
1965
+ },
1966
+ {
1967
+ "start": 77,
1968
+ "end": 88,
1969
+ "entity_type": "PHONE",
1970
+ "surface": "09126506157",
1971
+ "language": "en",
1972
+ "field": null
1973
+ }
1974
+ ],
1975
+ "meta": [
1976
+ {
1977
+ "field": "author",
1978
+ "entity_type": "PERSON",
1979
+ "surface": "Helen Fixture"
1980
+ },
1981
+ {
1982
+ "field": "keywords",
1983
+ "entity_type": "EMAIL",
1984
+ "surface": "Helen.Fixture06@aria-lab.test"
1985
+ }
1986
+ ]
1987
+ },
1988
+ {
1989
+ "doc_id": "DLP-PDF-MEMO-07",
1990
+ "spans": [
1991
+ {
1992
+ "start": 22,
1993
+ "end": 33,
1994
+ "entity_type": "PERSON",
1995
+ "surface": "Ava Fixture",
1996
+ "language": "en",
1997
+ "field": null
1998
+ },
1999
+ {
2000
+ "start": 43,
2001
+ "end": 70,
2002
+ "entity_type": "EMAIL",
2003
+ "surface": "Ava.Fixture07@aria-lab.test",
2004
+ "language": "en",
2005
+ "field": null
2006
+ },
2007
+ {
2008
+ "start": 73,
2009
+ "end": 84,
2010
+ "entity_type": "PHONE",
2011
+ "surface": "09126633882",
2012
+ "language": "en",
2013
+ "field": null
2014
+ }
2015
+ ],
2016
+ "meta": [
2017
+ {
2018
+ "field": "author",
2019
+ "entity_type": "PERSON",
2020
+ "surface": "Ava Fixture"
2021
+ },
2022
+ {
2023
+ "field": "keywords",
2024
+ "entity_type": "EMAIL",
2025
+ "surface": "Ava.Fixture07@aria-lab.test"
2026
+ }
2027
+ ]
2028
+ },
2029
+ {
2030
+ "doc_id": "DLP-PDF-MEMO-08",
2031
+ "spans": [
2032
+ {
2033
+ "start": 22,
2034
+ "end": 33,
2035
+ "entity_type": "PERSON",
2036
+ "surface": "Nima Labian",
2037
+ "language": "en",
2038
+ "field": null
2039
+ },
2040
+ {
2041
+ "start": 43,
2042
+ "end": 70,
2043
+ "entity_type": "EMAIL",
2044
+ "surface": "Nima.Labian08@aria-lab.test",
2045
+ "language": "en",
2046
+ "field": null
2047
+ },
2048
+ {
2049
+ "start": 73,
2050
+ "end": 84,
2051
+ "entity_type": "PHONE",
2052
+ "surface": "09128661790",
2053
+ "language": "en",
2054
+ "field": null
2055
+ }
2056
+ ],
2057
+ "meta": [
2058
+ {
2059
+ "field": "author",
2060
+ "entity_type": "PERSON",
2061
+ "surface": "Nima Labian"
2062
+ },
2063
+ {
2064
+ "field": "keywords",
2065
+ "entity_type": "EMAIL",
2066
+ "surface": "Nima.Labian08@aria-lab.test"
2067
+ }
2068
+ ]
2069
+ },
2070
+ {
2071
+ "doc_id": "DLP-PDF-MEMO-09",
2072
+ "spans": [
2073
+ {
2074
+ "start": 22,
2075
+ "end": 33,
2076
+ "entity_type": "PERSON",
2077
+ "surface": "Ava Fixture",
2078
+ "language": "en",
2079
+ "field": null
2080
+ },
2081
+ {
2082
+ "start": 43,
2083
+ "end": 70,
2084
+ "entity_type": "EMAIL",
2085
+ "surface": "Ava.Fixture09@aria-lab.test",
2086
+ "language": "en",
2087
+ "field": null
2088
+ },
2089
+ {
2090
+ "start": 73,
2091
+ "end": 84,
2092
+ "entity_type": "PHONE",
2093
+ "surface": "09122509645",
2094
+ "language": "en",
2095
+ "field": null
2096
+ }
2097
+ ],
2098
+ "meta": [
2099
+ {
2100
+ "field": "author",
2101
+ "entity_type": "PERSON",
2102
+ "surface": "Ava Fixture"
2103
+ },
2104
+ {
2105
+ "field": "keywords",
2106
+ "entity_type": "EMAIL",
2107
+ "surface": "Ava.Fixture09@aria-lab.test"
2108
+ }
2109
+ ]
2110
+ },
2111
+ {
2112
+ "doc_id": "DLP-PDF-MEMO-10",
2113
+ "spans": [
2114
+ {
2115
+ "start": 22,
2116
+ "end": 34,
2117
+ "entity_type": "PERSON",
2118
+ "surface": "Nima Fixture",
2119
+ "language": "en",
2120
+ "field": null
2121
+ },
2122
+ {
2123
+ "start": 44,
2124
+ "end": 72,
2125
+ "entity_type": "EMAIL",
2126
+ "surface": "Nima.Fixture10@aria-lab.test",
2127
+ "language": "en",
2128
+ "field": null
2129
+ },
2130
+ {
2131
+ "start": 75,
2132
+ "end": 86,
2133
+ "entity_type": "PHONE",
2134
+ "surface": "09122528399",
2135
+ "language": "en",
2136
+ "field": null
2137
+ }
2138
+ ],
2139
+ "meta": [
2140
+ {
2141
+ "field": "author",
2142
+ "entity_type": "PERSON",
2143
+ "surface": "Nima Fixture"
2144
+ },
2145
+ {
2146
+ "field": "keywords",
2147
+ "entity_type": "EMAIL",
2148
+ "surface": "Nima.Fixture10@aria-lab.test"
2149
+ }
2150
+ ]
2151
+ },
2152
+ {
2153
+ "doc_id": "DLP-HARDNEG-01",
2154
+ "spans": [],
2155
+ "meta": []
2156
+ },
2157
+ {
2158
+ "doc_id": "DLP-HARDNEG-02",
2159
+ "spans": [],
2160
+ "meta": []
2161
+ },
2162
+ {
2163
+ "doc_id": "DLP-HARDNEG-03",
2164
+ "spans": [],
2165
+ "meta": []
2166
+ },
2167
+ {
2168
+ "doc_id": "DLP-HARDNEG-04",
2169
+ "spans": [],
2170
+ "meta": []
2171
+ },
2172
+ {
2173
+ "doc_id": "DLP-HARDNEG-05",
2174
+ "spans": [],
2175
+ "meta": []
2176
+ },
2177
+ {
2178
+ "doc_id": "DLP-HARDNEG-06",
2179
+ "spans": [],
2180
+ "meta": []
2181
+ },
2182
+ {
2183
+ "doc_id": "DLP-HARDNEG-07",
2184
+ "spans": [],
2185
+ "meta": []
2186
+ },
2187
+ {
2188
+ "doc_id": "DLP-HARDNEG-08",
2189
+ "spans": [],
2190
+ "meta": []
2191
+ },
2192
+ {
2193
+ "doc_id": "DLP-HARDNEG-09",
2194
+ "spans": [],
2195
+ "meta": []
2196
+ },
2197
+ {
2198
+ "doc_id": "DLP-HARDNEG-10",
2199
+ "spans": [],
2200
+ "meta": []
2201
+ },
2202
+ {
2203
+ "doc_id": "DLP-TICKET-01",
2204
+ "spans": [
2205
+ {
2206
+ "start": 56,
2207
+ "end": 69,
2208
+ "entity_type": "PERSON",
2209
+ "surface": "Leila Fixture",
2210
+ "language": "en",
2211
+ "field": null
2212
+ },
2213
+ {
2214
+ "start": 72,
2215
+ "end": 84,
2216
+ "entity_type": "PERSON",
2217
+ "surface": "سارا آزمایشی",
2218
+ "language": "fa",
2219
+ "field": null
2220
+ },
2221
+ {
2222
+ "start": 85,
2223
+ "end": 114,
2224
+ "entity_type": "EMAIL",
2225
+ "surface": "Leila.Fixture01@aria-lab.test",
2226
+ "language": "en",
2227
+ "field": null
2228
+ },
2229
+ {
2230
+ "start": 115,
2231
+ "end": 126,
2232
+ "entity_type": "PHONE",
2233
+ "surface": "09126663672",
2234
+ "language": "en",
2235
+ "field": null
2236
+ },
2237
+ {
2238
+ "start": 127,
2239
+ "end": 137,
2240
+ "entity_type": "IR_NATIONAL_ID",
2241
+ "surface": "8810500075",
2242
+ "language": "fa",
2243
+ "field": null
2244
+ }
2245
+ ],
2246
+ "meta": []
2247
+ },
2248
+ {
2249
+ "doc_id": "DLP-TICKET-02",
2250
+ "spans": [
2251
+ {
2252
+ "start": 56,
2253
+ "end": 69,
2254
+ "entity_type": "PERSON",
2255
+ "surface": "Leila Fixture",
2256
+ "language": "en",
2257
+ "field": null
2258
+ },
2259
+ {
2260
+ "start": 72,
2261
+ "end": 85,
2262
+ "entity_type": "PERSON",
2263
+ "surface": "امید نمونه‌ای",
2264
+ "language": "fa",
2265
+ "field": null
2266
+ },
2267
+ {
2268
+ "start": 86,
2269
+ "end": 115,
2270
+ "entity_type": "EMAIL",
2271
+ "surface": "Leila.Fixture02@aria-lab.test",
2272
+ "language": "en",
2273
+ "field": null
2274
+ },
2275
+ {
2276
+ "start": 116,
2277
+ "end": 127,
2278
+ "entity_type": "PHONE",
2279
+ "surface": "09124960362",
2280
+ "language": "en",
2281
+ "field": null
2282
+ },
2283
+ {
2284
+ "start": 128,
2285
+ "end": 138,
2286
+ "entity_type": "IR_NATIONAL_ID",
2287
+ "surface": "7433177132",
2288
+ "language": "fa",
2289
+ "field": null
2290
+ }
2291
+ ],
2292
+ "meta": []
2293
+ },
2294
+ {
2295
+ "doc_id": "DLP-TICKET-03",
2296
+ "spans": [
2297
+ {
2298
+ "start": 56,
2299
+ "end": 67,
2300
+ "entity_type": "PERSON",
2301
+ "surface": "Ava Fixture",
2302
+ "language": "en",
2303
+ "field": null
2304
+ },
2305
+ {
2306
+ "start": 70,
2307
+ "end": 83,
2308
+ "entity_type": "PERSON",
2309
+ "surface": "نیما نمونه‌ای",
2310
+ "language": "fa",
2311
+ "field": null
2312
+ },
2313
+ {
2314
+ "start": 84,
2315
+ "end": 111,
2316
+ "entity_type": "EMAIL",
2317
+ "surface": "Ava.Fixture03@aria-lab.test",
2318
+ "language": "en",
2319
+ "field": null
2320
+ },
2321
+ {
2322
+ "start": 112,
2323
+ "end": 123,
2324
+ "entity_type": "PHONE",
2325
+ "surface": "09129013939",
2326
+ "language": "en",
2327
+ "field": null
2328
+ },
2329
+ {
2330
+ "start": 124,
2331
+ "end": 134,
2332
+ "entity_type": "IR_NATIONAL_ID",
2333
+ "surface": "4751024639",
2334
+ "language": "fa",
2335
+ "field": null
2336
+ }
2337
+ ],
2338
+ "meta": []
2339
+ },
2340
+ {
2341
+ "doc_id": "DLP-TICKET-04",
2342
+ "spans": [
2343
+ {
2344
+ "start": 56,
2345
+ "end": 68,
2346
+ "entity_type": "PERSON",
2347
+ "surface": "Omar Fixture",
2348
+ "language": "en",
2349
+ "field": null
2350
+ },
2351
+ {
2352
+ "start": 71,
2353
+ "end": 84,
2354
+ "entity_type": "PERSON",
2355
+ "surface": "سارا نمونه‌ای",
2356
+ "language": "fa",
2357
+ "field": null
2358
+ },
2359
+ {
2360
+ "start": 85,
2361
+ "end": 113,
2362
+ "entity_type": "EMAIL",
2363
+ "surface": "Omar.Fixture04@aria-lab.test",
2364
+ "language": "en",
2365
+ "field": null
2366
+ },
2367
+ {
2368
+ "start": 114,
2369
+ "end": 125,
2370
+ "entity_type": "PHONE",
2371
+ "surface": "09124265160",
2372
+ "language": "en",
2373
+ "field": null
2374
+ },
2375
+ {
2376
+ "start": 126,
2377
+ "end": 136,
2378
+ "entity_type": "IR_NATIONAL_ID",
2379
+ "surface": "0843397276",
2380
+ "language": "fa",
2381
+ "field": null
2382
+ }
2383
+ ],
2384
+ "meta": []
2385
+ },
2386
+ {
2387
+ "doc_id": "DLP-TICKET-05",
2388
+ "spans": [
2389
+ {
2390
+ "start": 56,
2391
+ "end": 67,
2392
+ "entity_type": "PERSON",
2393
+ "surface": "Ava Fixture",
2394
+ "language": "en",
2395
+ "field": null
2396
+ },
2397
+ {
2398
+ "start": 70,
2399
+ "end": 83,
2400
+ "entity_type": "PERSON",
2401
+ "surface": "امید نمونه‌ای",
2402
+ "language": "fa",
2403
+ "field": null
2404
+ },
2405
+ {
2406
+ "start": 84,
2407
+ "end": 111,
2408
+ "entity_type": "EMAIL",
2409
+ "surface": "Ava.Fixture05@aria-lab.test",
2410
+ "language": "en",
2411
+ "field": null
2412
+ },
2413
+ {
2414
+ "start": 112,
2415
+ "end": 123,
2416
+ "entity_type": "PHONE",
2417
+ "surface": "09127385597",
2418
+ "language": "en",
2419
+ "field": null
2420
+ },
2421
+ {
2422
+ "start": 124,
2423
+ "end": 134,
2424
+ "entity_type": "IR_NATIONAL_ID",
2425
+ "surface": "6714749521",
2426
+ "language": "fa",
2427
+ "field": null
2428
+ }
2429
+ ],
2430
+ "meta": []
2431
+ },
2432
+ {
2433
+ "doc_id": "DLP-TICKET-06",
2434
+ "spans": [
2435
+ {
2436
+ "start": 56,
2437
+ "end": 68,
2438
+ "entity_type": "PERSON",
2439
+ "surface": "Omar Fixture",
2440
+ "language": "en",
2441
+ "field": null
2442
+ },
2443
+ {
2444
+ "start": 71,
2445
+ "end": 83,
2446
+ "entity_type": "PERSON",
2447
+ "surface": "لیلا آزمایشی",
2448
+ "language": "fa",
2449
+ "field": null
2450
+ },
2451
+ {
2452
+ "start": 84,
2453
+ "end": 112,
2454
+ "entity_type": "EMAIL",
2455
+ "surface": "Omar.Fixture06@aria-lab.test",
2456
+ "language": "en",
2457
+ "field": null
2458
+ },
2459
+ {
2460
+ "start": 113,
2461
+ "end": 124,
2462
+ "entity_type": "PHONE",
2463
+ "surface": "09126454625",
2464
+ "language": "en",
2465
+ "field": null
2466
+ },
2467
+ {
2468
+ "start": 125,
2469
+ "end": 135,
2470
+ "entity_type": "IR_NATIONAL_ID",
2471
+ "surface": "7904846901",
2472
+ "language": "fa",
2473
+ "field": null
2474
+ }
2475
+ ],
2476
+ "meta": []
2477
+ },
2478
+ {
2479
+ "doc_id": "DLP-TICKET-07",
2480
+ "spans": [
2481
+ {
2482
+ "start": 56,
2483
+ "end": 69,
2484
+ "entity_type": "PERSON",
2485
+ "surface": "Helen Fixture",
2486
+ "language": "en",
2487
+ "field": null
2488
+ },
2489
+ {
2490
+ "start": 72,
2491
+ "end": 84,
2492
+ "entity_type": "PERSON",
2493
+ "surface": "سارا آزمایشی",
2494
+ "language": "fa",
2495
+ "field": null
2496
+ },
2497
+ {
2498
+ "start": 85,
2499
+ "end": 114,
2500
+ "entity_type": "EMAIL",
2501
+ "surface": "Helen.Fixture07@aria-lab.test",
2502
+ "language": "en",
2503
+ "field": null
2504
+ },
2505
+ {
2506
+ "start": 115,
2507
+ "end": 126,
2508
+ "entity_type": "PHONE",
2509
+ "surface": "09127938081",
2510
+ "language": "en",
2511
+ "field": null
2512
+ },
2513
+ {
2514
+ "start": 127,
2515
+ "end": 137,
2516
+ "entity_type": "IR_NATIONAL_ID",
2517
+ "surface": "6114411269",
2518
+ "language": "fa",
2519
+ "field": null
2520
+ }
2521
+ ],
2522
+ "meta": []
2523
+ },
2524
+ {
2525
+ "doc_id": "DLP-TICKET-08",
2526
+ "spans": [
2527
+ {
2528
+ "start": 56,
2529
+ "end": 67,
2530
+ "entity_type": "PERSON",
2531
+ "surface": "Ava Fixture",
2532
+ "language": "en",
2533
+ "field": null
2534
+ },
2535
+ {
2536
+ "start": 70,
2537
+ "end": 83,
2538
+ "entity_type": "PERSON",
2539
+ "surface": "لیلا نمونه‌ای",
2540
+ "language": "fa",
2541
+ "field": null
2542
+ },
2543
+ {
2544
+ "start": 84,
2545
+ "end": 111,
2546
+ "entity_type": "EMAIL",
2547
+ "surface": "Ava.Fixture08@aria-lab.test",
2548
+ "language": "en",
2549
+ "field": null
2550
+ },
2551
+ {
2552
+ "start": 112,
2553
+ "end": 123,
2554
+ "entity_type": "PHONE",
2555
+ "surface": "09125317904",
2556
+ "language": "en",
2557
+ "field": null
2558
+ },
2559
+ {
2560
+ "start": 124,
2561
+ "end": 134,
2562
+ "entity_type": "IR_NATIONAL_ID",
2563
+ "surface": "2799826970",
2564
+ "language": "fa",
2565
+ "field": null
2566
+ }
2567
+ ],
2568
+ "meta": []
2569
+ },
2570
+ {
2571
+ "doc_id": "DLP-TICKET-09",
2572
+ "spans": [
2573
+ {
2574
+ "start": 56,
2575
+ "end": 68,
2576
+ "entity_type": "PERSON",
2577
+ "surface": "Omar Fixture",
2578
+ "language": "en",
2579
+ "field": null
2580
+ },
2581
+ {
2582
+ "start": 71,
2583
+ "end": 84,
2584
+ "entity_type": "PERSON",
2585
+ "surface": "نیما نمونه‌ای",
2586
+ "language": "fa",
2587
+ "field": null
2588
+ },
2589
+ {
2590
+ "start": 85,
2591
+ "end": 113,
2592
+ "entity_type": "EMAIL",
2593
+ "surface": "Omar.Fixture09@aria-lab.test",
2594
+ "language": "en",
2595
+ "field": null
2596
+ },
2597
+ {
2598
+ "start": 114,
2599
+ "end": 125,
2600
+ "entity_type": "PHONE",
2601
+ "surface": "09125714005",
2602
+ "language": "en",
2603
+ "field": null
2604
+ },
2605
+ {
2606
+ "start": 126,
2607
+ "end": 136,
2608
+ "entity_type": "IR_NATIONAL_ID",
2609
+ "surface": "0468470395",
2610
+ "language": "fa",
2611
+ "field": null
2612
+ }
2613
+ ],
2614
+ "meta": []
2615
+ },
2616
+ {
2617
+ "doc_id": "DLP-TICKET-10",
2618
+ "spans": [
2619
+ {
2620
+ "start": 56,
2621
+ "end": 68,
2622
+ "entity_type": "PERSON",
2623
+ "surface": "Helen Labian",
2624
+ "language": "en",
2625
+ "field": null
2626
+ },
2627
+ {
2628
+ "start": 71,
2629
+ "end": 84,
2630
+ "entity_type": "PERSON",
2631
+ "surface": "سارا نمونه‌ای",
2632
+ "language": "fa",
2633
+ "field": null
2634
+ },
2635
+ {
2636
+ "start": 85,
2637
+ "end": 113,
2638
+ "entity_type": "EMAIL",
2639
+ "surface": "Helen.Labian10@aria-lab.test",
2640
+ "language": "en",
2641
+ "field": null
2642
+ },
2643
+ {
2644
+ "start": 114,
2645
+ "end": 125,
2646
+ "entity_type": "PHONE",
2647
+ "surface": "09126818844",
2648
+ "language": "en",
2649
+ "field": null
2650
+ },
2651
+ {
2652
+ "start": 126,
2653
+ "end": 136,
2654
+ "entity_type": "IR_NATIONAL_ID",
2655
+ "surface": "5049585570",
2656
+ "language": "fa",
2657
+ "field": null
2658
+ }
2659
+ ],
2660
+ "meta": []
2661
+ }
2662
+ ]
demo/examples/protocol.md ADDED
@@ -0,0 +1,23 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # AriaDLP laboratory protocol
2
+
3
+ Synthetic bilingual DLP corpus. Seed 24. Split by template.
4
+
5
+ ## What is labeled
6
+
7
+ Person names from a closed laboratory lexicon, mailbox addresses on `aria-lab.test`,
8
+ Iranian mobile numbers, checksum-valid national IDs, Luhn-valid test PANs, IBANs,
9
+ and `ARIA-24xxx` employee ids.
10
+
11
+ ## Hard negatives
12
+
13
+ Invoice / SKU tokens, ISBN, semantic versions, `Mr. Protocol`, dates, and numbers
14
+ that fail national-ID / Luhn / IBAN checks.
15
+
16
+ ## PDF
17
+
18
+ `en_pdf_memo` is a generated text PDF. Gold includes body spans and metadata fields.
19
+ Acceptance is content removal plus metadata wipe, not drawing a rectangle.
20
+
21
+ ## Limits
22
+
23
+ This corpus does not prove production recall. Unknown items stay on the review list.
demo/examples/sample_preview.json ADDED
@@ -0,0 +1,88 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "run_id": "ariadlp-6a3c5f54da644ce9",
3
+ "status": "scoped_hold",
4
+ "status_label": "Scoped hold \u2014 labeled gold was removed from this export; unknown types stay listed",
5
+ "kpis": {
6
+ "n_docs": 20,
7
+ "n_spans": 70,
8
+ "n_review": 0,
9
+ "n_redacted": 70,
10
+ "policy_precision": 1.0,
11
+ "policy_recall": 1.0,
12
+ "baseline_precision": 0.8333333333333334,
13
+ "baseline_recall": 0.7142857142857143,
14
+ "residual_leak": 0.0,
15
+ "over_redaction": 0.0,
16
+ "leftover_exports": 0,
17
+ "metadata_leaks": 0,
18
+ "elapsed_ms": 52,
19
+ "engine_profile": "presidio-style-checksum-policy",
20
+ "policy_action": "mask",
21
+ "pages": 0,
22
+ "chars": 2897
23
+ },
24
+ "kpi_html": "<div class='kpi-grid'><article class='kpi-card'><p class='kpi-label'>Documents in this run</p><p class='kpi-value'>20</p><p class='kpi-sub'>presidio-style-checksum-policy</p></article><article class='kpi-card'><p class='kpi-label'>Policy spans</p><p class='kpi-value'>70</p><p class='kpi-sub'>checksum policy</p></article><article class='kpi-card'><p class='kpi-label'>Policy precision</p><p class='kpi-value'>100%</p><p class='kpi-sub'>baseline 83%</p></article><article class='kpi-card'><p class='kpi-label'>Policy recall</p><p class='kpi-value'>100%</p><p class='kpi-sub'>baseline 71%</p></article><article class='kpi-card'><p class='kpi-label'>Exports with leftover gold</p><p class='kpi-value'>0</p><p class='kpi-sub'>52 ms</p></article></div>",
25
+ "catalog": [
26
+ {
27
+ "scenario_id": "S01",
28
+ "method": "checksum_policy",
29
+ "precision": 1.0,
30
+ "recall": 1.0,
31
+ "over_redaction": 0.0,
32
+ "residual_leak": 0.0,
33
+ "leftover_exports": 0,
34
+ "metadata_leaks": 0,
35
+ "hardneg_hits": 0,
36
+ "n_docs": 20
37
+ },
38
+ {
39
+ "scenario_id": "S02",
40
+ "method": "checksum_policy",
41
+ "precision": 1.0,
42
+ "recall": 1.0,
43
+ "over_redaction": 0.0,
44
+ "residual_leak": 0.0,
45
+ "leftover_exports": 0,
46
+ "metadata_leaks": 0,
47
+ "hardneg_hits": 0,
48
+ "n_docs": 20
49
+ },
50
+ {
51
+ "scenario_id": "S03",
52
+ "method": "checksum_policy",
53
+ "precision": 1.0,
54
+ "recall": 1.0,
55
+ "over_redaction": 0.0,
56
+ "residual_leak": 0.0,
57
+ "leftover_exports": 0,
58
+ "metadata_leaks": 0,
59
+ "hardneg_hits": 0,
60
+ "n_docs": 10
61
+ },
62
+ {
63
+ "scenario_id": "S04",
64
+ "method": "checksum_policy",
65
+ "precision": 1.0,
66
+ "recall": 1.0,
67
+ "over_redaction": 0.0,
68
+ "residual_leak": 0.0,
69
+ "leftover_exports": 0,
70
+ "metadata_leaks": 0,
71
+ "hardneg_hits": 10,
72
+ "n_docs": 20
73
+ },
74
+ {
75
+ "scenario_id": "S05",
76
+ "method": "checksum_policy",
77
+ "precision": 1.0,
78
+ "recall": 1.0,
79
+ "over_redaction": 0.0,
80
+ "residual_leak": 0.0,
81
+ "leftover_exports": 0,
82
+ "metadata_leaks": 0,
83
+ "hardneg_hits": 10,
84
+ "n_docs": 80
85
+ }
86
+ ],
87
+ "precomputed": true
88
+ }
demo/examples/scenarios.jsonl ADDED
@@ -0,0 +1,5 @@
 
 
 
 
 
 
1
+ {"scenario_id":"S01","title":"English mailbox and invoice CSV","description":"Development templates: support email plus invoice CSV.","description_fa":"قالب‌های توسعه: ایمیل پشتیبانی و سی‌اس‌وی فاکتور.","doc_ids":["DLP-EN-EMAIL-01","DLP-EN-EMAIL-02","DLP-EN-EMAIL-03","DLP-EN-EMAIL-04","DLP-EN-EMAIL-05","DLP-EN-EMAIL-06","DLP-EN-EMAIL-07","DLP-EN-EMAIL-08","DLP-EN-EMAIL-09","DLP-EN-EMAIL-10","DLP-CSV-INV-01","DLP-CSV-INV-02","DLP-CSV-INV-03","DLP-CSV-INV-04","DLP-CSV-INV-05","DLP-CSV-INV-06","DLP-CSV-INV-07","DLP-CSV-INV-08","DLP-CSV-INV-09","DLP-CSV-INV-10"],"split":"development"}
2
+ {"scenario_id":"S02","title":"Persian HR and onboarding","description":"Persian digits / ZWNJ HR letter plus English onboarding with test PAN and IBAN.","description_fa":"نامه منابع انسانی با رقم فارسی و نیم‌فاصله به‌همراه برگه ورود.","doc_ids":["DLP-FA-HR-01","DLP-FA-HR-02","DLP-FA-HR-03","DLP-FA-HR-04","DLP-FA-HR-05","DLP-FA-HR-06","DLP-FA-HR-07","DLP-FA-HR-08","DLP-FA-HR-09","DLP-FA-HR-10","DLP-EN-ONB-01","DLP-EN-ONB-02","DLP-EN-ONB-03","DLP-EN-ONB-04","DLP-EN-ONB-05","DLP-EN-ONB-06","DLP-EN-ONB-07","DLP-EN-ONB-08","DLP-EN-ONB-09","DLP-EN-ONB-10"],"split":"development"}
3
+ {"scenario_id":"S03","title":"Hard negatives","description":"Failed checksums, ISBN, Mr. Protocol, versions. Policy should not over-redact.","description_fa":"checksum نامعتبر، شابک، آقای پروتکل و نسخه. سیاست نباید بیش‌پوشانی کند.","doc_ids":["DLP-HARDNEG-01","DLP-HARDNEG-02","DLP-HARDNEG-03","DLP-HARDNEG-04","DLP-HARDNEG-05","DLP-HARDNEG-06","DLP-HARDNEG-07","DLP-HARDNEG-08","DLP-HARDNEG-09","DLP-HARDNEG-10"],"split":"test"}
4
+ {"scenario_id":"S04","title":"PDF memo and Persian contract","description":"Text PDF with metadata PII plus a Persian contract that keeps invalid national IDs for review.","description_fa":"پی‌دی‌اف متنی با فراداده حساس و قرارداد فارسی با کد ملی نامعتبر برای بازبینی.","doc_ids":["DLP-FA-CTR-01","DLP-FA-CTR-02","DLP-FA-CTR-03","DLP-FA-CTR-04","DLP-FA-CTR-05","DLP-FA-CTR-06","DLP-FA-CTR-07","DLP-FA-CTR-08","DLP-FA-CTR-09","DLP-FA-CTR-10","DLP-PDF-MEMO-01","DLP-PDF-MEMO-02","DLP-PDF-MEMO-03","DLP-PDF-MEMO-04","DLP-PDF-MEMO-05","DLP-PDF-MEMO-06","DLP-PDF-MEMO-07","DLP-PDF-MEMO-08","DLP-PDF-MEMO-09","DLP-PDF-MEMO-10"],"split":"test"}
5
+ {"scenario_id":"S05","title":"Full 80-document catalog","description":"Template-disjoint development/test split, seed 24.","description_fa":"تفکیک قالب توسعه/آزمون، بذر ۲۴.","doc_ids":["DLP-EN-EMAIL-01","DLP-EN-EMAIL-02","DLP-EN-EMAIL-03","DLP-EN-EMAIL-04","DLP-EN-EMAIL-05","DLP-EN-EMAIL-06","DLP-EN-EMAIL-07","DLP-EN-EMAIL-08","DLP-EN-EMAIL-09","DLP-EN-EMAIL-10","DLP-FA-HR-01","DLP-FA-HR-02","DLP-FA-HR-03","DLP-FA-HR-04","DLP-FA-HR-05","DLP-FA-HR-06","DLP-FA-HR-07","DLP-FA-HR-08","DLP-FA-HR-09","DLP-FA-HR-10","DLP-CSV-INV-01","DLP-CSV-INV-02","DLP-CSV-INV-03","DLP-CSV-INV-04","DLP-CSV-INV-05","DLP-CSV-INV-06","DLP-CSV-INV-07","DLP-CSV-INV-08","DLP-CSV-INV-09","DLP-CSV-INV-10","DLP-EN-ONB-01","DLP-EN-ONB-02","DLP-EN-ONB-03","DLP-EN-ONB-04","DLP-EN-ONB-05","DLP-EN-ONB-06","DLP-EN-ONB-07","DLP-EN-ONB-08","DLP-EN-ONB-09","DLP-EN-ONB-10","DLP-FA-CTR-01","DLP-FA-CTR-02","DLP-FA-CTR-03","DLP-FA-CTR-04","DLP-FA-CTR-05","DLP-FA-CTR-06","DLP-FA-CTR-07","DLP-FA-CTR-08","DLP-FA-CTR-09","DLP-FA-CTR-10","DLP-PDF-MEMO-01","DLP-PDF-MEMO-02","DLP-PDF-MEMO-03","DLP-PDF-MEMO-04","DLP-PDF-MEMO-05","DLP-PDF-MEMO-06","DLP-PDF-MEMO-07","DLP-PDF-MEMO-08","DLP-PDF-MEMO-09","DLP-PDF-MEMO-10","DLP-HARDNEG-01","DLP-HARDNEG-02","DLP-HARDNEG-03","DLP-HARDNEG-04","DLP-HARDNEG-05","DLP-HARDNEG-06","DLP-HARDNEG-07","DLP-HARDNEG-08","DLP-HARDNEG-09","DLP-HARDNEG-10","DLP-TICKET-01","DLP-TICKET-02","DLP-TICKET-03","DLP-TICKET-04","DLP-TICKET-05","DLP-TICKET-06","DLP-TICKET-07","DLP-TICKET-08","DLP-TICKET-09","DLP-TICKET-10"],"split":"all"}
demo/examples/schema.json ADDED
@@ -0,0 +1,25 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_id": "aria.dlp.lab_document.v1",
3
+ "fields": [
4
+ "doc_id",
5
+ "template",
6
+ "split",
7
+ "language",
8
+ "format",
9
+ "body",
10
+ "gold_spans",
11
+ "gold_meta",
12
+ "hard_negatives"
13
+ ],
14
+ "entity_types": [
15
+ "PERSON",
16
+ "EMAIL",
17
+ "PHONE",
18
+ "IR_NATIONAL_ID",
19
+ "CREDIT_CARD",
20
+ "IBAN",
21
+ "EMPLOYEE_ID",
22
+ "UNKNOWN"
23
+ ],
24
+ "notes": "Synthetic only. Addresses use aria-lab.test. PANs are Luhn-valid test numbers."
25
+ }
demo/examples/splits.json ADDED
@@ -0,0 +1,17 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "development_templates": [
3
+ "en_onboarding",
4
+ "en_support_email",
5
+ "fa_hr_letter",
6
+ "mixed_invoice_csv"
7
+ ],
8
+ "test_templates": [
9
+ "en_pdf_memo",
10
+ "fa_contract_notice",
11
+ "hardneg_catalog",
12
+ "mixed_ticket"
13
+ ],
14
+ "rule": "Split is by document template, not random rows.",
15
+ "n_documents": 80,
16
+ "seed": 24
17
+ }
demo/requirements.txt ADDED
@@ -0,0 +1,4 @@
 
 
 
 
 
1
+ pandas>=2.2.0,<3
2
+ pydantic>=2.6.0,<3
3
+ plotly>=5.24.0,<7
4
+ pymupdf>=1.24.0,<1.27
demo/src/__init__.py ADDED
@@ -0,0 +1 @@
 
 
1
+ """AriaDLP laboratory package."""
demo/src/analyzer.py ADDED
@@ -0,0 +1,47 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ from __future__ import annotations
2
+
3
+ from src.domain import SPECIFICITY, EntityType, Method, Workflow
4
+ from src.recognizers import default_recognizers
5
+ from src.schemas import Span
6
+
7
+
8
+ def merge_spans(spans: list[Span]) -> list[Span]:
9
+ ordered = sorted(spans, key=lambda item: (item.start, -(item.end - item.start), -SPECIFICITY.get(item.entity_type, 0)))
10
+ kept: list[Span] = []
11
+ for item in ordered:
12
+ overlap = next((other for other in kept if item.start < other.end and item.end > other.start), None)
13
+ if overlap is None:
14
+ kept.append(item)
15
+ continue
16
+ item_score = (SPECIFICITY.get(item.entity_type, 0), item.confidence)
17
+ other_score = (SPECIFICITY.get(overlap.entity_type, 0), overlap.confidence)
18
+ if item_score > other_score:
19
+ kept.remove(overlap)
20
+ kept.append(item)
21
+ kept.sort(key=lambda item: (item.start, item.end))
22
+ for index, item in enumerate(kept, start=1):
23
+ item.finding_id = item.finding_id or f"SPAN-{index:03d}"
24
+ return kept
25
+
26
+
27
+ def analyze_policy(text: str) -> list[Span]:
28
+ raw: list[Span] = []
29
+ for recognizer in default_recognizers():
30
+ raw.extend(recognizer.analyze(text or ""))
31
+ return merge_spans(raw)
32
+
33
+
34
+ def redactable(spans: list[Span], suppressed: set[str] | None = None) -> list[Span]:
35
+ blocked = suppressed or set()
36
+ kept = []
37
+ for item in spans:
38
+ if item.finding_id in blocked:
39
+ item.workflow = Workflow.SUPPRESSED
40
+ continue
41
+ if item.entity_type is EntityType.UNKNOWN or item.workflow is Workflow.REVIEW:
42
+ continue
43
+ if item.method == Method.BASELINE.value:
44
+ kept.append(item)
45
+ continue
46
+ kept.append(item)
47
+ return kept
demo/src/baseline.py ADDED
@@ -0,0 +1,48 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ from __future__ import annotations
2
+
3
+ import re
4
+
5
+ from src.domain import EntityType, Method, Workflow
6
+ from src.schemas import Span
7
+
8
+ EMAIL_RE = re.compile(r"\b[A-Za-z0-9._%+-]+@[A-Za-z0-9.-]+\.[A-Za-z]{2,}\b")
9
+ PHONE_RE = re.compile(r"(?<!\d)(?:\+98|0)9\d{9}(?!\d)")
10
+ TEN_DIGIT_RE = re.compile(r"(?<!\d)\d{10}(?!\d)")
11
+ SIXTEEN_DIGIT_RE = re.compile(r"(?<!\d)\d{16}(?!\d)")
12
+ IBAN_RE = re.compile(r"\bIR\d{24}\b")
13
+ EMPLOYEE_RE = re.compile(r"\b(?:ARIA|EMP)-\d{4,6}\b", re.I)
14
+ HONORIFIC_RE = re.compile(r"\b(?:Mr|Ms|Mrs|Dr)\.?\s+[A-Z][A-Za-z]+(?:\s+[A-Z][A-Za-z]+)?")
15
+
16
+
17
+ def _span(text: str, start: int, end: int, entity_type: EntityType, rule_id: str) -> Span:
18
+ return Span(
19
+ start=start,
20
+ end=end,
21
+ entity_type=entity_type,
22
+ surface=text[start:end],
23
+ confidence=0.4,
24
+ method=Method.BASELINE.value,
25
+ rule_id=rule_id,
26
+ workflow=Workflow.DETECTED,
27
+ reason="Regex-only baseline without checksum or digit fold",
28
+ )
29
+
30
+
31
+ def analyze_regex_only(text: str) -> list[Span]:
32
+ """Naive recognizer: ASCII regex, no Persian digits, no checksum, no hard-negative filter."""
33
+ found: list[Span] = []
34
+ for match in EMAIL_RE.finditer(text):
35
+ found.append(_span(text, match.start(), match.end(), EntityType.EMAIL, "base.email"))
36
+ for match in PHONE_RE.finditer(text):
37
+ found.append(_span(text, match.start(), match.end(), EntityType.PHONE, "base.phone"))
38
+ for match in TEN_DIGIT_RE.finditer(text):
39
+ found.append(_span(text, match.start(), match.end(), EntityType.IR_NATIONAL_ID, "base.ten_digit"))
40
+ for match in SIXTEEN_DIGIT_RE.finditer(text):
41
+ found.append(_span(text, match.start(), match.end(), EntityType.CREDIT_CARD, "base.sixteen_digit"))
42
+ for match in IBAN_RE.finditer(text):
43
+ found.append(_span(text, match.start(), match.end(), EntityType.IBAN, "base.iban"))
44
+ for match in EMPLOYEE_RE.finditer(text):
45
+ found.append(_span(text, match.start(), match.end(), EntityType.EMPLOYEE_ID, "base.employee"))
46
+ for match in HONORIFIC_RE.finditer(text):
47
+ found.append(_span(text, match.start(), match.end(), EntityType.PERSON, "base.honorific"))
48
+ return found
demo/src/checksums.py ADDED
@@ -0,0 +1,106 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ from __future__ import annotations
2
+
3
+ from src.normalize import only_digits, strip_iban
4
+
5
+
6
+ def iran_national_id_ok(value: str) -> bool:
7
+ digits = only_digits(value)
8
+ if len(digits) != 10 or not digits.isdigit():
9
+ return False
10
+ if len(set(digits)) == 1:
11
+ return False
12
+ total = sum(int(digits[index]) * (10 - index) for index in range(9))
13
+ remainder = total % 11
14
+ check = remainder if remainder < 2 else 11 - remainder
15
+ return check == int(digits[9])
16
+
17
+
18
+ def luhn_ok(value: str) -> bool:
19
+ digits = [int(ch) for ch in only_digits(value)]
20
+ if len(digits) < 13 or len(digits) > 19:
21
+ return False
22
+ checksum = 0
23
+ for index, digit in enumerate(reversed(digits)):
24
+ if index % 2 == 1:
25
+ digit *= 2
26
+ if digit > 9:
27
+ digit -= 9
28
+ checksum += digit
29
+ return checksum % 10 == 0
30
+
31
+
32
+ def iban_ok(value: str) -> bool:
33
+ compact = strip_iban(value)
34
+ if len(compact) < 15 or len(compact) > 34:
35
+ return False
36
+ if not compact[:2].isalpha() or not compact[2:4].isdigit():
37
+ return False
38
+ rearranged = compact[4:] + compact[:4]
39
+ numbers = []
40
+ for char in rearranged:
41
+ if char.isdigit():
42
+ numbers.append(char)
43
+ elif char.isalpha():
44
+ numbers.append(str(ord(char) - 55))
45
+ else:
46
+ return False
47
+ try:
48
+ return int("".join(numbers)) % 97 == 1
49
+ except ValueError:
50
+ return False
51
+
52
+
53
+ def make_national_id(rng) -> str:
54
+ for _ in range(80):
55
+ body = "".join(str(rng.randint(0, 9)) for _ in range(9))
56
+ if len(set(body)) == 1:
57
+ continue
58
+ total = sum(int(body[index]) * (10 - index) for index in range(9))
59
+ remainder = total % 11
60
+ check = remainder if remainder < 2 else 11 - remainder
61
+ candidate = f"{body}{check}"
62
+ if iran_national_id_ok(candidate):
63
+ return candidate
64
+ raise RuntimeError("national id generation failed")
65
+
66
+
67
+ def make_luhn(rng, prefix: str = "4111", length: int = 16) -> str:
68
+ body_len = length - 1
69
+ while True:
70
+ rest = "".join(str(rng.randint(0, 9)) for _ in range(body_len - len(prefix)))
71
+ partial = prefix + rest
72
+ for check in range(10):
73
+ candidate = partial + str(check)
74
+ if luhn_ok(candidate):
75
+ return candidate
76
+
77
+
78
+ def make_iban(rng, country: str = "IR") -> str:
79
+ bank = f"{rng.randint(10, 99):02d}{rng.randint(1, 9)}"
80
+ account = f"{rng.randint(1, 10**18 - 1):019d}"
81
+ for check in range(100):
82
+ candidate = f"{country}{check:02d}{bank}{account}"
83
+ if iban_ok(candidate):
84
+ return candidate
85
+ raise RuntimeError("iban generation failed")
86
+
87
+
88
+ def break_national_id(valid: str) -> str:
89
+ digits = list(only_digits(valid))
90
+ digits[-1] = str((int(digits[-1]) + 1) % 10)
91
+ broken = "".join(digits)
92
+ if not iran_national_id_ok(broken):
93
+ return broken
94
+ digits[-2] = str((int(digits[-2]) + 1) % 10)
95
+ return "".join(digits)
96
+
97
+
98
+ def break_luhn(valid: str) -> str:
99
+ digits = list(only_digits(valid))
100
+ digits[-1] = str((int(digits[-1]) + 1) % 10)
101
+ return "".join(digits)
102
+
103
+
104
+ def break_iban(valid: str) -> str:
105
+ compact = strip_iban(valid)
106
+ return compact[:4] + "9" + compact[5:]
demo/src/domain.py ADDED
@@ -0,0 +1,122 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ from __future__ import annotations
2
+
3
+ from enum import Enum
4
+
5
+ PRODUCT = "AriaDLP"
6
+ CODE_VERSION = "1.0.0"
7
+ SCHEMA_ID = "aria.dlp.lab_document.v1"
8
+ RULESET_ID = "dlp-presidio-style-v1"
9
+ RULESET_DATE = "2026-09-21"
10
+ ENGINE_PROFILE = "presidio-style-checksum-policy"
11
+ DEFAULT_SEED = 24
12
+ MAX_BYTES = 10 * 1024 * 1024
13
+ MAX_CHARS = 80_000
14
+ MAX_FILES = 5
15
+ MAX_PAGES = 10
16
+ MAX_DOCS = 200
17
+ N_VARIANTS = 10
18
+ RELATED_SECRET_GUARD = "08-secret-guard"
19
+ RELATED_RAG_BOUNDARY = "23-rag-boundary"
20
+
21
+ LAB_EMAIL_DOMAIN = "aria-lab.test"
22
+ LAB_NAME_MARKERS = ("Labian", "Fixture", "نمونه‌ای", "آزمایشی")
23
+
24
+ DEV_TEMPLATES = frozenset(
25
+ {
26
+ "en_support_email",
27
+ "fa_hr_letter",
28
+ "mixed_invoice_csv",
29
+ "en_onboarding",
30
+ }
31
+ )
32
+ TEST_TEMPLATES = frozenset(
33
+ {
34
+ "fa_contract_notice",
35
+ "en_pdf_memo",
36
+ "hardneg_catalog",
37
+ "mixed_ticket",
38
+ }
39
+ )
40
+
41
+
42
+ class Split(str, Enum):
43
+ DEVELOPMENT = "development"
44
+ TEST = "test"
45
+ USER = "user"
46
+
47
+
48
+ class Method(str, Enum):
49
+ BASELINE = "regex_only"
50
+ POLICY = "checksum_policy"
51
+
52
+
53
+ class PolicyAction(str, Enum):
54
+ MASK = "mask"
55
+ REMOVE = "remove"
56
+
57
+
58
+ class EntityType(str, Enum):
59
+ PERSON = "PERSON"
60
+ EMAIL = "EMAIL"
61
+ PHONE = "PHONE"
62
+ IR_NATIONAL_ID = "IR_NATIONAL_ID"
63
+ CREDIT_CARD = "CREDIT_CARD"
64
+ IBAN = "IBAN"
65
+ EMPLOYEE_ID = "EMPLOYEE_ID"
66
+ UNKNOWN = "UNKNOWN"
67
+
68
+
69
+ class Language(str, Enum):
70
+ EN = "en"
71
+ FA = "fa"
72
+ MIXED = "mixed"
73
+
74
+
75
+ class DocFormat(str, Enum):
76
+ TEXT = "text"
77
+ CSV = "csv"
78
+ PDF = "pdf"
79
+
80
+
81
+ class Verdict(str, Enum):
82
+ REVIEW = "needs_review"
83
+ HOLD = "scoped_hold"
84
+ INSUFFICIENT = "insufficient_evidence"
85
+ INVALID = "invalid_input"
86
+
87
+
88
+ class Feedback(str, Enum):
89
+ CORRECT = "correct"
90
+ FALSE_POSITIVE = "false_positive"
91
+ INSUFFICIENT = "insufficient_evidence"
92
+ NEEDS_REVIEW = "needs_review"
93
+
94
+
95
+ class Workflow(str, Enum):
96
+ DETECTED = "detected"
97
+ REVIEW = "needs_review"
98
+ SUPPRESSED = "suppressed"
99
+
100
+
101
+ class Family(str, Enum):
102
+ EN_SUPPORT_EMAIL = "en_support_email"
103
+ FA_HR_LETTER = "fa_hr_letter"
104
+ MIXED_INVOICE_CSV = "mixed_invoice_csv"
105
+ EN_ONBOARDING = "en_onboarding"
106
+ FA_CONTRACT_NOTICE = "fa_contract_notice"
107
+ EN_PDF_MEMO = "en_pdf_memo"
108
+ HARDNEG_CATALOG = "hardneg_catalog"
109
+ MIXED_TICKET = "mixed_ticket"
110
+ USER = "user"
111
+
112
+
113
+ SPECIFICITY = {
114
+ EntityType.EMAIL: 90,
115
+ EntityType.CREDIT_CARD: 85,
116
+ EntityType.IBAN: 85,
117
+ EntityType.IR_NATIONAL_ID: 80,
118
+ EntityType.PHONE: 75,
119
+ EntityType.EMPLOYEE_ID: 70,
120
+ EntityType.PERSON: 40,
121
+ EntityType.UNKNOWN: 10,
122
+ }
demo/src/explain.py ADDED
@@ -0,0 +1,68 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ from __future__ import annotations
2
+
3
+ from src.domain import Verdict
4
+ from src.i18n import t
5
+ from src.schemas import Finding, Kpis, MethodScore, Span, ValidationReport
6
+
7
+
8
+ def decide_verdict(report: ValidationReport, n_docs: int, leftover: int, n_review: int) -> Verdict:
9
+ if not report.ok:
10
+ return Verdict.INVALID
11
+ if n_docs <= 0:
12
+ return Verdict.INSUFFICIENT
13
+ if leftover > 0 or n_review > 0:
14
+ return Verdict.REVIEW
15
+ return Verdict.HOLD
16
+
17
+
18
+ def status_label(verdict: Verdict, lang: str) -> str:
19
+ mapping = {
20
+ Verdict.REVIEW: t(lang, "status_review"),
21
+ Verdict.HOLD: t(lang, "status_hold"),
22
+ Verdict.INSUFFICIENT: t(lang, "status_insufficient"),
23
+ Verdict.INVALID: t(lang, "status_invalid"),
24
+ }
25
+ return mapping[verdict]
26
+
27
+
28
+ def reasons(policy: MethodScore, baseline: MethodScore, lang: str) -> list[str]:
29
+ return [
30
+ t(lang, "reason_policy").format(p=f"{policy.precision:.2f}", r=f"{policy.recall:.2f}"),
31
+ t(lang, "reason_baseline").format(p=f"{baseline.precision:.2f}", r=f"{baseline.recall:.2f}"),
32
+ t(lang, "reason_over").format(n=f"{policy.over_redaction:.3f}", base=f"{baseline.over_redaction:.3f}"),
33
+ t(lang, "reason_left").format(n=policy.leftover_exports),
34
+ t(lang, "reason_meta").format(n=policy.metadata_leaks),
35
+ t(lang, "reason_scope"),
36
+ t(lang, "reason_no_llm"),
37
+ ]
38
+
39
+
40
+ def narrative(kpis: Kpis, lang: str) -> str:
41
+ return t(lang, "narrative").format(
42
+ docs=kpis.n_docs,
43
+ spans=kpis.n_spans,
44
+ precision=f"{100 * kpis.policy_precision:.0f}%",
45
+ recall=f"{100 * kpis.policy_recall:.0f}%",
46
+ leftover=kpis.leftover_exports,
47
+ )
48
+
49
+
50
+ def finding_from_span(doc_id: str, span: Span, lang: str) -> Finding:
51
+ preview = span.surface if len(span.surface) <= 6 else f"{span.surface[:2]}…{span.surface[-2:]}"
52
+ return Finding(
53
+ finding_id=span.finding_id,
54
+ doc_id=doc_id,
55
+ entity_type=span.entity_type.value,
56
+ start=span.start,
57
+ end=span.end,
58
+ surface_preview=preview,
59
+ confidence=span.confidence,
60
+ method=span.method,
61
+ rule_id=span.rule_id,
62
+ workflow=span.workflow.value,
63
+ reason=span.reason,
64
+ language=span.language,
65
+ observed=t(lang, "observed_span").format(start=span.start, end=span.end, rule=span.rule_id),
66
+ inference=span.reason,
67
+ action=t(lang, "action_span").format(kind=span.entity_type.value),
68
+ )
demo/src/generator.py ADDED
@@ -0,0 +1,462 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ from __future__ import annotations
2
+
3
+ import json
4
+ import random
5
+ from pathlib import Path
6
+
7
+ from src.checksums import break_iban, break_luhn, break_national_id, make_iban, make_luhn, make_national_id
8
+ from src.domain import (
9
+ DEFAULT_SEED,
10
+ DEV_TEMPLATES,
11
+ LAB_EMAIL_DOMAIN,
12
+ N_VARIANTS,
13
+ TEST_TEMPLATES,
14
+ DocFormat,
15
+ EntityType,
16
+ Family,
17
+ Language,
18
+ Split,
19
+ )
20
+ from src.normalize import fold_digits
21
+ from src.schemas import GoldMeta, GoldSpan, LabDocument, Scenario
22
+
23
+
24
+ EN_FIRST = ["Ava", "Noah", "Leila", "Omar", "Helen", "Nima"]
25
+ EN_LAST = ["Labian", "Fixture"]
26
+ FA_FIRST = ["سارا", "نیما", "لیلا", "امید"]
27
+ FA_LAST = ["نمونه‌ای", "آزمایشی"]
28
+
29
+
30
+ class Builder:
31
+ def __init__(self) -> None:
32
+ self.parts: list[str] = []
33
+ self.spans: list[GoldSpan] = []
34
+
35
+ def add(self, text: str) -> "Builder":
36
+ self.parts.append(text)
37
+ return self
38
+
39
+ def pii(self, text: str, entity_type: EntityType, language: str = "en") -> "Builder":
40
+ start = len("".join(self.parts))
41
+ self.parts.append(text)
42
+ self.spans.append(GoldSpan(start=start, end=start + len(text), entity_type=entity_type, surface=text, language=language))
43
+ return self
44
+
45
+ def text(self) -> str:
46
+ return "".join(self.parts)
47
+
48
+
49
+ def to_persian_digits(value: str) -> str:
50
+ table = str.maketrans("0123456789", "۰۱۲۳۴۵۶۷۸۹")
51
+ return value.translate(table)
52
+
53
+
54
+ def _person(rng: random.Random, lang: str) -> tuple[str, str, str]:
55
+ if lang == "fa":
56
+ first, last = rng.choice(FA_FIRST), rng.choice(FA_LAST)
57
+ return first, last, f"{first} {last}"
58
+ first, last = rng.choice(EN_FIRST), rng.choice(EN_LAST)
59
+ return first, last, f"{first} {last}"
60
+
61
+
62
+ def _email(first: str, last: str, index: int) -> str:
63
+ slug = f"{fold_digits(first)}.{fold_digits(last)}{index:02d}".encode("ascii", "ignore").decode("ascii")
64
+ if not slug or slug.startswith("."):
65
+ slug = f"user{index:02d}"
66
+ slug = "".join(ch if ch.isalnum() or ch == "." else "x" for ch in slug)
67
+ return f"{slug}@{LAB_EMAIL_DOMAIN}"
68
+
69
+
70
+ def _phone(rng: random.Random) -> str:
71
+ return f"0912{rng.randint(1000000, 9999999)}"
72
+
73
+
74
+ def _employee(index: int) -> str:
75
+ return f"ARIA-24{index:03d}"
76
+
77
+
78
+ def _en_support_email(rng: random.Random, index: int) -> LabDocument:
79
+ first, last, name = _person(rng, "en")
80
+ email = _email(first, last, index)
81
+ phone = _phone(rng)
82
+ emp = _employee(index)
83
+ builder = Builder()
84
+ builder.add("From: ").pii(name, EntityType.PERSON).add(" <").pii(email, EntityType.EMAIL).add(">\n")
85
+ builder.add("Phone: ").pii(phone, EntityType.PHONE).add("\n")
86
+ builder.add("Employee: ").pii(emp, EntityType.EMPLOYEE_ID).add("\n")
87
+ builder.add(f"Please reset the laboratory mailbox for {first}. Ticket ORD-2024-{index:04d}. Version 2.4.1.\n")
88
+ return LabDocument(
89
+ doc_id=f"DLP-EN-EMAIL-{index:02d}",
90
+ template="en_support_email",
91
+ family=Family.EN_SUPPORT_EMAIL.value,
92
+ split=Split.DEVELOPMENT,
93
+ language=Language.EN,
94
+ format=DocFormat.TEXT,
95
+ title="English support mailbox",
96
+ body=builder.text(),
97
+ gold_spans=builder.spans,
98
+ hard_negatives=[f"ORD-2024-{index:04d}", "2.4.1"],
99
+ description="English support email with name, mailbox, phone, and employee id.",
100
+ description_fa="ایمیل پشتیبانی انگلیسی با نام، صندوق، تلفن و شناسه کارمند.",
101
+ )
102
+
103
+
104
+ def _fa_hr_letter(rng: random.Random, index: int) -> LabDocument:
105
+ first, last, name = _person(rng, "fa")
106
+ nid = make_national_id(rng)
107
+ phone = _phone(rng)
108
+ email = _email("user", last, index)
109
+ nid_fa = to_persian_digits(nid)
110
+ phone_fa = to_persian_digits(phone)
111
+ builder = Builder()
112
+ builder.add("خانم ").pii(name, EntityType.PERSON, "fa").add("،\n")
113
+ builder.add("کد\u200cملی ").pii(nid_fa, EntityType.IR_NATIONAL_ID, "fa").add("\n")
114
+ builder.add("تلفن همراه ").pii(phone_fa, EntityType.PHONE, "fa").add("\n")
115
+ builder.add("ایمیل ").pii(email, EntityType.EMAIL).add("\n")
116
+ builder.add(f"شماره پیگیری INV-1403-{index:04d} و نسخه ۳٫۲٫۱ فقط ارجاع داخلی‌اند.\n")
117
+ return LabDocument(
118
+ doc_id=f"DLP-FA-HR-{index:02d}",
119
+ template="fa_hr_letter",
120
+ family=Family.FA_HR_LETTER.value,
121
+ split=Split.DEVELOPMENT,
122
+ language=Language.FA,
123
+ format=DocFormat.TEXT,
124
+ title="Persian HR letter",
125
+ body=builder.text(),
126
+ gold_spans=builder.spans,
127
+ hard_negatives=[f"INV-1403-{index:04d}"],
128
+ description="Persian HR note with ZWNJ, Persian digits, national id, and phone.",
129
+ description_fa="نامه منابع انسانی با نیم‌فاصله، رقم فارسی، کد ملی و تلفن.",
130
+ )
131
+
132
+
133
+ def _mixed_invoice_csv(rng: random.Random, index: int) -> LabDocument:
134
+ first, last, name = _person(rng, "en")
135
+ email = _email(first, last, index)
136
+ phone = _phone(rng)
137
+ invoice = f"INV-2024-{1000 + index}"
138
+ bad_nid = break_national_id(make_national_id(rng))
139
+ builder = Builder()
140
+ builder.add("invoice_id,customer,email,phone,lookup\n")
141
+ builder.add(f"{invoice},").pii(name, EntityType.PERSON).add(",").pii(email, EntityType.EMAIL)
142
+ builder.add(",").pii(phone, EntityType.PHONE).add(f",{bad_nid}\n")
143
+ return LabDocument(
144
+ doc_id=f"DLP-CSV-INV-{index:02d}",
145
+ template="mixed_invoice_csv",
146
+ family=Family.MIXED_INVOICE_CSV.value,
147
+ split=Split.DEVELOPMENT,
148
+ language=Language.MIXED,
149
+ format=DocFormat.CSV,
150
+ title="Invoice CSV",
151
+ body=builder.text(),
152
+ gold_spans=builder.spans,
153
+ hard_negatives=[invoice, bad_nid],
154
+ description="CSV with a customer row and a checksum-failing lookup number.",
155
+ description_fa="سی‌اس‌وی مشتری با شماره جست‌وجوی بدون checksum معتبر.",
156
+ )
157
+
158
+
159
+ def _en_onboarding(rng: random.Random, index: int) -> LabDocument:
160
+ first, last, name = _person(rng, "en")
161
+ card = make_luhn(rng)
162
+ iban = make_iban(rng)
163
+ email = _email(first, last, index)
164
+ builder = Builder()
165
+ builder.add("Name: ").pii(name, EntityType.PERSON).add("\n")
166
+ builder.add("Work email: ").pii(email, EntityType.EMAIL).add("\n")
167
+ builder.add("Travel advance test PAN: ").pii(card, EntityType.CREDIT_CARD).add("\n")
168
+ builder.add("Payroll IBAN: ").pii(iban, EntityType.IBAN).add("\n")
169
+ builder.add("ISBN 978-3-16-148410-0 is a catalog example, not an account.\n")
170
+ return LabDocument(
171
+ doc_id=f"DLP-EN-ONB-{index:02d}",
172
+ template="en_onboarding",
173
+ family=Family.EN_ONBOARDING.value,
174
+ split=Split.DEVELOPMENT,
175
+ language=Language.EN,
176
+ format=DocFormat.TEXT,
177
+ title="Onboarding worksheet",
178
+ body=builder.text(),
179
+ gold_spans=builder.spans,
180
+ hard_negatives=["978-3-16-148410-0", "9783161484100"],
181
+ description="Onboarding sheet with test PAN and IBAN plus an ISBN hard negative.",
182
+ description_fa="برگه ورود با کارت و شبا آزمایشی و ISBN به‌عنوان منفی سخت.",
183
+ )
184
+
185
+
186
+ def _fa_contract_notice(rng: random.Random, index: int) -> LabDocument:
187
+ first, last, name = _person(rng, "fa")
188
+ nid = make_national_id(rng)
189
+ phone = to_persian_digits(_phone(rng))
190
+ invalid = break_national_id(make_national_id(rng))
191
+ builder = Builder()
192
+ builder.add("موضوع: اخطار قراردادی\n")
193
+ builder.add("نام: ").pii(name, EntityType.PERSON, "fa").add("\n")
194
+ builder.add("کد ملی ").pii(nid, EntityType.IR_NATIONAL_ID, "fa").add("\n")
195
+ builder.add("تلفن ").pii(phone, EntityType.PHONE, "fa").add("\n")
196
+ builder.add(f"کد ملی {invalid} در سامانه پذیرفته نشد و برای بازبینی مانده است.\n")
197
+ builder.add("تاریخ ۱۴۰۳/۰۵/۱۲ و نسخه ۱.۰.۰ ارجاع سند هستند.\n")
198
+ return LabDocument(
199
+ doc_id=f"DLP-FA-CTR-{index:02d}",
200
+ template="fa_contract_notice",
201
+ family=Family.FA_CONTRACT_NOTICE.value,
202
+ split=Split.TEST,
203
+ language=Language.FA,
204
+ format=DocFormat.TEXT,
205
+ title="Persian contract notice",
206
+ body=builder.text(),
207
+ gold_spans=builder.spans,
208
+ hard_negatives=[invalid, "1403/05/12", "۱۴۰۳/۰۵/۱۲", "1.0.0"],
209
+ description="Persian contract with a failed national-id left for review.",
210
+ description_fa="اخطار قراردادی با کد ملی نامعتبر برای بازبینی.",
211
+ )
212
+
213
+
214
+ def _en_pdf_memo(rng: random.Random, index: int) -> LabDocument:
215
+ first, last, name = _person(rng, "en")
216
+ email = _email(first, last, index)
217
+ phone = _phone(rng)
218
+ builder = Builder()
219
+ builder.add("Internal memo\n")
220
+ builder.add("Author: ").pii(name, EntityType.PERSON).add("\n")
221
+ builder.add("Contact: ").pii(email, EntityType.EMAIL).add(" / ").pii(phone, EntityType.PHONE).add("\n")
222
+ builder.add("Mr. Protocol is the software module, not a person.\n")
223
+ return LabDocument(
224
+ doc_id=f"DLP-PDF-MEMO-{index:02d}",
225
+ template="en_pdf_memo",
226
+ family=Family.EN_PDF_MEMO.value,
227
+ split=Split.TEST,
228
+ language=Language.EN,
229
+ format=DocFormat.PDF,
230
+ title="PDF memo",
231
+ body=builder.text(),
232
+ metadata={"author": name, "keywords": email, "title": f"Memo {index}"},
233
+ gold_spans=builder.spans,
234
+ gold_meta=[
235
+ GoldMeta(field="author", entity_type=EntityType.PERSON, surface=name),
236
+ GoldMeta(field="keywords", entity_type=EntityType.EMAIL, surface=email),
237
+ ],
238
+ hard_negatives=["Mr. Protocol"],
239
+ description="Text PDF whose metadata also repeats the author mailbox.",
240
+ description_fa="پی‌دی‌اف متنی که فراداده هم نام و ایمیل را تکرار می‌کند.",
241
+ )
242
+
243
+
244
+ def _hardneg_catalog(rng: random.Random, index: int) -> LabDocument:
245
+ bad_card = break_luhn(make_luhn(rng))
246
+ bad_iban = break_iban(make_iban(rng))
247
+ bad_nid = break_national_id(make_national_id(rng))
248
+ body = (
249
+ f"Catalog row {index}\n"
250
+ f"ISBN 978-0-306-40615-7\n"
251
+ f"Invoice {bad_nid}\n"
252
+ f"Card-like {bad_card}\n"
253
+ f"IBAN-like {bad_iban}\n"
254
+ f"Mr. Protocol handler v2.4.1\n"
255
+ f"Date 2026/09/21 SKU-9{index:03d}\n"
256
+ )
257
+ return LabDocument(
258
+ doc_id=f"DLP-HARDNEG-{index:02d}",
259
+ template="hardneg_catalog",
260
+ family=Family.HARDNEG_CATALOG.value,
261
+ split=Split.TEST,
262
+ language=Language.EN,
263
+ format=DocFormat.TEXT,
264
+ title="Hard-negative catalog",
265
+ body=body,
266
+ gold_spans=[],
267
+ hard_negatives=[bad_card, bad_iban, bad_nid, "Mr. Protocol", "2.4.1", "978-0-306-40615-7"],
268
+ description="Lookalikes that must not be redacted: failed checksums, ISBN, protocol title.",
269
+ description_fa="شبیه‌سازی‌هایی که نباید پوشانده شوند: checksum نامعتبر، شابک، عنوان پروتکل.",
270
+ )
271
+
272
+
273
+ def _mixed_ticket(rng: random.Random, index: int) -> LabDocument:
274
+ first, last, name = _person(rng, "en")
275
+ fa_first, fa_last, fa_name = _person(rng, "fa")
276
+ email = _email(first, last, index)
277
+ phone = _phone(rng)
278
+ nid = make_national_id(rng)
279
+ invoice = f"INV-MIX-{index:04d}"
280
+ builder = Builder()
281
+ builder.add("ticket_id,reporter,email,phone,national_id\n")
282
+ builder.add(f"{invoice},").pii(name, EntityType.PERSON).add(" / ").pii(fa_name, EntityType.PERSON, "fa")
283
+ builder.add(",").pii(email, EntityType.EMAIL).add(",").pii(phone, EntityType.PHONE)
284
+ builder.add(",").pii(nid, EntityType.IR_NATIONAL_ID, "fa").add("\n")
285
+ return LabDocument(
286
+ doc_id=f"DLP-TICKET-{index:02d}",
287
+ template="mixed_ticket",
288
+ family=Family.MIXED_TICKET.value,
289
+ split=Split.TEST,
290
+ language=Language.MIXED,
291
+ format=DocFormat.CSV,
292
+ title="Mixed-language ticket CSV",
293
+ body=builder.text(),
294
+ gold_spans=builder.spans,
295
+ hard_negatives=[invoice],
296
+ description="Mixed English/Persian ticket row.",
297
+ description_fa="ردیف تیکت دوزبانه.",
298
+ )
299
+
300
+
301
+ BUILDERS = {
302
+ "en_support_email": _en_support_email,
303
+ "fa_hr_letter": _fa_hr_letter,
304
+ "mixed_invoice_csv": _mixed_invoice_csv,
305
+ "en_onboarding": _en_onboarding,
306
+ "fa_contract_notice": _fa_contract_notice,
307
+ "en_pdf_memo": _en_pdf_memo,
308
+ "hardneg_catalog": _hardneg_catalog,
309
+ "mixed_ticket": _mixed_ticket,
310
+ }
311
+
312
+
313
+ def generate_documents(seed: int = DEFAULT_SEED) -> list[LabDocument]:
314
+ rng = random.Random(seed)
315
+ docs: list[LabDocument] = []
316
+ for template, builder in BUILDERS.items():
317
+ for index in range(1, N_VARIANTS + 1):
318
+ docs.append(builder(rng, index))
319
+ assert docs[-1].template == template
320
+ return docs
321
+
322
+
323
+ def documents_by_id(seed: int = DEFAULT_SEED) -> dict[str, LabDocument]:
324
+ return {item.doc_id: item for item in generate_documents(seed)}
325
+
326
+
327
+ def scenarios(seed: int = DEFAULT_SEED) -> list[Scenario]:
328
+ docs = generate_documents(seed)
329
+
330
+ def ids(*templates: str) -> list[str]:
331
+ return [item.doc_id for item in docs if item.template in templates]
332
+
333
+ return [
334
+ Scenario(
335
+ scenario_id="S01",
336
+ title="English mailbox and invoice CSV",
337
+ description="Development templates: support email plus invoice CSV.",
338
+ description_fa="قالب‌های توسعه: ایمیل پشتیبانی و سی‌اس‌وی فاکتور.",
339
+ doc_ids=ids("en_support_email", "mixed_invoice_csv"),
340
+ split="development",
341
+ ),
342
+ Scenario(
343
+ scenario_id="S02",
344
+ title="Persian HR and onboarding",
345
+ description="Persian digits / ZWNJ HR letter plus English onboarding with test PAN and IBAN.",
346
+ description_fa="نامه منابع انسانی با رقم فارسی و نیم‌فاصله به‌همراه برگه ورود.",
347
+ doc_ids=ids("fa_hr_letter", "en_onboarding"),
348
+ split="development",
349
+ ),
350
+ Scenario(
351
+ scenario_id="S03",
352
+ title="Hard negatives",
353
+ description="Failed checksums, ISBN, Mr. Protocol, versions. Policy should not over-redact.",
354
+ description_fa="checksum نامعتبر، شابک، آقای پروتکل و نسخه. سیاست نباید بیش‌پوشانی کند.",
355
+ doc_ids=ids("hardneg_catalog"),
356
+ split="test",
357
+ ),
358
+ Scenario(
359
+ scenario_id="S04",
360
+ title="PDF memo and Persian contract",
361
+ description="Text PDF with metadata PII plus a Persian contract that keeps invalid national IDs for review.",
362
+ description_fa="پی‌دی‌اف متنی با فراداده حساس و قرارداد فارسی با کد ملی نامعتبر برای بازبینی.",
363
+ doc_ids=ids("en_pdf_memo", "fa_contract_notice"),
364
+ split="test",
365
+ ),
366
+ Scenario(
367
+ scenario_id="S05",
368
+ title="Full 80-document catalog",
369
+ description="Template-disjoint development/test split, seed 24.",
370
+ description_fa="تفکیک قالب توسعه/آزمون، بذر ۲۴.",
371
+ doc_ids=[item.doc_id for item in docs],
372
+ split="all",
373
+ ),
374
+ ]
375
+
376
+
377
+ def scenario_by_id(scenario_id: str, seed: int = DEFAULT_SEED) -> Scenario:
378
+ return next(item for item in scenarios(seed) if item.scenario_id == scenario_id)
379
+
380
+
381
+ def write_examples(folder: Path, seed: int = DEFAULT_SEED) -> None:
382
+ folder.mkdir(parents=True, exist_ok=True)
383
+ docs = generate_documents(seed)
384
+ (folder / "documents.jsonl").write_text(
385
+ "\n".join(item.model_dump_json() for item in docs) + "\n",
386
+ encoding="utf-8",
387
+ newline="\n",
388
+ )
389
+ (folder / "scenarios.jsonl").write_text(
390
+ "\n".join(item.model_dump_json() for item in scenarios(seed)) + "\n",
391
+ encoding="utf-8",
392
+ newline="\n",
393
+ )
394
+ schema = {
395
+ "schema_id": "aria.dlp.lab_document.v1",
396
+ "fields": [
397
+ "doc_id",
398
+ "template",
399
+ "split",
400
+ "language",
401
+ "format",
402
+ "body",
403
+ "gold_spans",
404
+ "gold_meta",
405
+ "hard_negatives",
406
+ ],
407
+ "entity_types": [item.value for item in EntityType],
408
+ "notes": "Synthetic only. Addresses use aria-lab.test. PANs are Luhn-valid test numbers.",
409
+ }
410
+ (folder / "schema.json").write_text(json.dumps(schema, indent=2), encoding="utf-8", newline="\n")
411
+ split_map = {
412
+ "development_templates": sorted(DEV_TEMPLATES),
413
+ "test_templates": sorted(TEST_TEMPLATES),
414
+ "rule": "Split is by document template, not random rows.",
415
+ "n_documents": len(docs),
416
+ "seed": seed,
417
+ }
418
+ (folder / "splits.json").write_text(json.dumps(split_map, indent=2), encoding="utf-8", newline="\n")
419
+ gold = [
420
+ {
421
+ "doc_id": item.doc_id,
422
+ "spans": [span.model_dump() for span in item.gold_spans],
423
+ "meta": [meta.model_dump() for meta in item.gold_meta],
424
+ }
425
+ for item in docs
426
+ ]
427
+ (folder / "gold_spans.json").write_text(json.dumps(gold, indent=2, ensure_ascii=False), encoding="utf-8", newline="\n")
428
+ protocol = """# AriaDLP laboratory protocol
429
+
430
+ Synthetic bilingual DLP corpus. Seed 24. Split by template.
431
+
432
+ ## What is labeled
433
+
434
+ Person names from a closed laboratory lexicon, mailbox addresses on `aria-lab.test`,
435
+ Iranian mobile numbers, checksum-valid national IDs, Luhn-valid test PANs, IBANs,
436
+ and `ARIA-24xxx` employee ids.
437
+
438
+ ## Hard negatives
439
+
440
+ Invoice / SKU tokens, ISBN, semantic versions, `Mr. Protocol`, dates, and numbers
441
+ that fail national-ID / Luhn / IBAN checks.
442
+
443
+ ## PDF
444
+
445
+ `en_pdf_memo` is a generated text PDF. Gold includes body spans and metadata fields.
446
+ Acceptance is content removal plus metadata wipe, not drawing a rectangle.
447
+
448
+ ## Limits
449
+
450
+ This corpus does not prove production recall. Unknown items stay on the review list.
451
+ """
452
+ (folder / "protocol.md").write_text(protocol, encoding="utf-8", newline="\n")
453
+ feature = """# Feature card
454
+
455
+ - Task: span detection and reversible laboratory redaction
456
+ - Languages: English, Persian (digits, ZWNJ)
457
+ - Formats: text, CSV, text PDF (≤10 pages)
458
+ - Baseline: ASCII regex without checksum
459
+ - Policy: Presidio-style recognizer registry + checksum + lexicon
460
+ - Split: template-disjoint
461
+ """
462
+ (folder / "feature_card.md").write_text(feature, encoding="utf-8", newline="\n")
demo/src/i18n.py ADDED
@@ -0,0 +1,143 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ from __future__ import annotations
2
+
3
+ TEXTS = {
4
+ "en": {
5
+ "title": "AriaDLP — Persian / English Redaction Lab",
6
+ "subtitle": "Find laboratory personal data in text, CSV, and text PDFs, then preview a mask or remove export. Detection is scoped. This demo does not claim that every sensitive value was removed.",
7
+ "kind": "Document-analysis decision support · CPU · Presidio-style recognizers · no live LLM",
8
+ "run_sample": "Run sample",
9
+ "cancel": "Cancel run",
10
+ "reset": "Reset defaults",
11
+ "load_sample": "Load selected scenario",
12
+ "apply_review": "Apply span review",
13
+ "language": "Language",
14
+ "scenario": "Scenario",
15
+ "policy": "Export policy",
16
+ "sample_note": "Synthetic bilingual corpus, seed 24. Addresses use aria-lab.test. PANs are Luhn-valid test numbers. Scanned PDFs are out of scope.",
17
+ "live_note": "Live CPU run. Regex-only is the baseline. Policy adds checksums, Persian digits, ZWNJ, and a name lexicon.",
18
+ "kpi_docs": "Documents in this run",
19
+ "kpi_spans": "Policy spans",
20
+ "kpi_precision": "Policy precision",
21
+ "kpi_recall": "Policy recall",
22
+ "kpi_left": "Exports with leftover gold",
23
+ "optimized": "checksum policy",
24
+ "baseline": "regex-only baseline",
25
+ "empty": "No live result yet. Load a scenario and run the sample.",
26
+ "status": "Run status",
27
+ "status_review": "Needs review — leftover gold, unknown checksums, or suppressed spans remain",
28
+ "status_hold": "Scoped hold — labeled gold was removed from this export; unknown types stay listed",
29
+ "status_insufficient": "Insufficient evidence — no laboratory document to score",
30
+ "status_invalid": "Invalid input — see the field locator",
31
+ "upload_hint": "Paste text or upload UTF-8 TXT/CSV/JSON or a text PDF (10 MB, 10 pages). Real identity documents and scanned PDFs are rejected. S00 is the interactive editor.",
32
+ "advanced": "Advanced / interactive document",
33
+ "editor_text": "Paste laboratory text",
34
+ "uploads": "Upload text, CSV, or text PDF",
35
+ "feedback": "Analyst note (not used to retrain)",
36
+ "correct": "Correct",
37
+ "false_positive": "False positive",
38
+ "insufficient": "Insufficient evidence",
39
+ "needs_review": "Needs review",
40
+ "compare_plot": "Precision, recall, and over-redaction",
41
+ "type_plot": "Policy precision / recall by entity type",
42
+ "left_plot": "Leftover gold values by document",
43
+ "method_body": """
44
+ AriaDLP is a **document redaction workbench** for bilingual laboratory text.
45
+
46
+ **What it decides.** Character spans for mailbox, phone, Iranian national ID, test PAN, IBAN, employee id, and a closed name lexicon. Policy uses checksums and Persian/Latin digit folding. Regex-only is the baseline.
47
+
48
+ **What it does not do.** It does not run a GPU NER. It does not call an LLM on raw text. Scanned PDFs and production identity data are out of scope. Official Presidio Persian coverage is **not** assumed; recognizers here are custom.
49
+
50
+ **Export.** `mask` replaces a span with `[TYPE:hash]`. `remove` deletes it. PDF acceptance is content-stream redaction plus metadata wipe, not a painted rectangle. Preview the export before download.
51
+
52
+ **Split.** Development vs test is by document template, not random rows. Seed 24.
53
+ """,
54
+ "reason_policy": "Policy span precision {p} / recall {r} on labeled gold.",
55
+ "reason_baseline": "Regex-only precision {p} / recall {r}; it misses Persian digits and over-hits failed checksums.",
56
+ "reason_over": "Policy over-redaction {n}; baseline {base}.",
57
+ "reason_left": "Exports still containing a labeled gold value: {n}.",
58
+ "reason_meta": "Metadata fields still containing gold: {n}.",
59
+ "reason_scope": "A clean lab suite is not a claim that all sensitive data was removed.",
60
+ "reason_no_llm": "No paid API. An optional later LLM would only see already-redacted text.",
61
+ "narrative": "Scored {docs} documents and {spans} policy spans. Precision {precision}, recall {recall}, leftover exports {leftover}.",
62
+ "observed": "Observed",
63
+ "inference": "Inference",
64
+ "action": "Suggested action",
65
+ "observed_span": "Offsets {start}:{end} cited by `{rule}`.",
66
+ "action_span": "Review `{kind}` then keep or suppress before download.",
67
+ "before": "Original with detector spans",
68
+ "after": "Redacted preview",
69
+ "preview_note": "Preview the export. Download only after the span table looks right.",
70
+ },
71
+ "fa": {
72
+ "title": "آریادی‌ال‌پی — آزمایشگاه پوشاندن داده فارسی / انگلیسی",
73
+ "subtitle": "داده شخصی آزمایشگاهی را در متن، CSV و PDF متن‌دار پیدا کنید و پیش‌نمایش mask یا حذف را ببینید. تشخیص محدود به دامنه کنترل‌شده است. این دمو ادعا نمی‌کند همه داده حساس حذف شده است.",
74
+ "kind": "تصمیم‌یار تحلیل سند · CPU · شناساگر به سبک Presidio · بدون مدل زبانی زنده",
75
+ "run_sample": "اجرای نمونه",
76
+ "cancel": "لغو اجرا",
77
+ "reset": "بازگشت به پیش‌فرض",
78
+ "load_sample": "بارگذاری سناریوی انتخابی",
79
+ "apply_review": "اعمال بازبینی بازه",
80
+ "language": "زبان",
81
+ "scenario": "سناریو",
82
+ "policy": "سیاست خروجی",
83
+ "sample_note": "پیکره دوزبانه ساختگی، بذر ۲۴. نشانی‌ها روی aria-lab.test. کارت‌ها شماره آزمایشی Luhn. PDF اسکن‌شده خارج از محدوده است.",
84
+ "live_note": "اجرای زنده CPU. مبنا فقط regex است. سیاست checksum، رقم فارسی، نیم‌فاصله و واژه‌نامه نام را اضافه می‌کند.",
85
+ "kpi_docs": "اسناد این اجرا",
86
+ "kpi_spans": "بازه‌های سیاست",
87
+ "kpi_precision": "دقت سیاست",
88
+ "kpi_recall": "یادآوری سیاست",
89
+ "kpi_left": "خروجی با باقی‌مانده طلایی",
90
+ "optimized": "سیاست checksum",
91
+ "baseline": "مبنا فقط regex",
92
+ "empty": "هنوز نتیجه زنده‌ای نیست. سناریو را بارگذاری و نمونه را اجرا کنید.",
93
+ "status": "وضعیت اجرا",
94
+ "status_review": "نیازمند بررسی — باقی‌مانده طلایی، checksum نامعلوم یا بازه حذف‌شده",
95
+ "status_hold": "نگهداشت محدود — برچسب طلایی از این خروجی حذف شد؛ انواع ناشناخته در فهرست مانده‌اند",
96
+ "status_insufficient": "شواهد ناکافی — سندی برای امتیازدهی نیست",
97
+ "status_invalid": "ورودی نامعتبر — محل فیلد را ببینید",
98
+ "upload_hint": "متن بچسبانید یا TXT/CSV/JSON یا PDF متن‌دار بارگذاری کنید (۱۰ مگابایت، ۱۰ صفحه). سند هویتی واقعی و PDF اسکن‌شده رد می‌شود. S00 ویرایشگر تعاملی است.",
99
+ "advanced": "سند پیشرفته / تعاملی",
100
+ "editor_text": "متن آزمایشگاهی",
101
+ "uploads": "بارگذاری متن، CSV یا PDF متن‌دار",
102
+ "feedback": "یادداشت تحلیل‌گر (وارد آموزش نمی‌شود)",
103
+ "correct": "صحیح",
104
+ "false_positive": "خطای مثبت",
105
+ "insufficient": "شواهد ناکافی",
106
+ "needs_review": "نیازمند بررسی",
107
+ "compare_plot": "دقت، یادآوری و بیش‌پوشانی",
108
+ "type_plot": "دقت / یادآوری سیاست به تفکیک نوع",
109
+ "left_plot": "باقی‌مانده طلایی به تفکیک سند",
110
+ "method_body": """
111
+ آریادی‌ال‌پی میز کار پوشاندن سند برای متن آزمایشگاهی دوزبانه است.
112
+
113
+ **چه چیزی را تصمیم می‌گیرد.** بازه کاراکتری برای ایمیل، تلفن، کد ملی، کارت آزمایشی، شبا، شناسه کارمند و واژه‌نامه بسته نام. سیاست از checksum و تا کردن رقم فارسی/لاتین استفاده می‌کند. مبنا فقط regex است.
114
+
115
+ **چه کاری نمی‌کند.** NER روی GPU اجرا نمی‌شود. مدل زبانی روی متن خام صدا زده نمی‌شود. PDF اسکن‌شده و داده هویتی واقعی خارج از محدوده‌اند. پوشش فارسی رسمی Presidio فرض نشده؛ شناساگرها اختصاصی‌اند.
116
+
117
+ **خروجی.** `mask` بازه را با `[TYPE:hash]` عوض می‌کند. `remove` آن را حذف می‌کند. پذیرش PDF حذف واقعی محتوا و پاک‌کردن فراداده است، نه کشیدن مستطیل. پیش‌نمایش را قبل از دانلود ببینید.
118
+
119
+ **تفکیک.** توسعه و آزمون بر اساس قالب سند است نه ردیف تصادفی. بذر ۲۴.
120
+ """,
121
+ "reason_policy": "دقت سیاست {p} / یادآوری {r} روی برچسب طلایی.",
122
+ "reason_baseline": "دقت regex {p} / یادآوری {r}؛ رقم فارسی را از دست می‌دهد و checksum نامعتبر را بیش‌تشخیص می‌دهد.",
123
+ "reason_over": "بیش‌پوشانی سیاست {n}؛ مبنا {base}.",
124
+ "reason_left": "خروجی‌هایی که هنوز مقدار طلایی دارند: {n}.",
125
+ "reason_meta": "فیلد فراداده با مقدار طلایی: {n}.",
126
+ "reason_scope": "سوئیت تمیز آزمایشگاه به‌معنی حذف همه داده حساس نیست.",
127
+ "reason_no_llm": "بدون API پولی. مدل بعدی ��قط متن ازپیش‌پوشانده را می‌بیند.",
128
+ "narrative": "{docs} سند و {spans} بازه سیاست. دقت {precision}، یادآوری {recall}، خروجی باقی‌مانده {leftover}.",
129
+ "observed": "واقعیت مشاهده‌شده",
130
+ "inference": "استنباط",
131
+ "action": "اقدام پیشنهادی",
132
+ "observed_span": "بازه {start}:{end} با قاعده `{rule}`.",
133
+ "action_span": "`{kind}` را بازبینی کنید؛ سپس نگه دارید یا از خروجی حذف کنید.",
134
+ "before": "اصل با بازه‌های آشکارساز",
135
+ "after": "پیش‌نمایش پوشانده‌شده",
136
+ "preview_note": "پیش‌نمایش خروجی. فقط وقتی جدول بازه درست است دانلود کنید.",
137
+ },
138
+ }
139
+
140
+
141
+ def t(lang: str, key: str) -> str:
142
+ bundle = TEXTS.get(lang) or TEXTS["en"]
143
+ return bundle.get(key) or TEXTS["en"].get(key, key)
demo/src/metrics.py ADDED
@@ -0,0 +1,212 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ from __future__ import annotations
2
+
3
+ from collections import defaultdict
4
+
5
+ from src.domain import ENGINE_PROFILE, EntityType, Method, PolicyAction
6
+ from src.redaction import leftover_values
7
+ from src.schemas import GoldSpan, Kpis, LabDocument, MethodScore, Span
8
+
9
+
10
+ def iou(left: tuple[int, int], right: tuple[int, int]) -> float:
11
+ start = max(left[0], right[0])
12
+ end = min(left[1], right[1])
13
+ inter = max(0, end - start)
14
+ union = max(left[1], right[1]) - min(left[0], right[0])
15
+ return inter / union if union else 0.0
16
+
17
+
18
+ def _match_pairs(pred: list[Span], gold: list[GoldSpan], threshold: float = 0.5) -> tuple[int, int, int]:
19
+ used: set[int] = set()
20
+ tp = 0
21
+ for item in pred:
22
+ if item.entity_type is EntityType.UNKNOWN:
23
+ continue
24
+ for index, truth in enumerate(gold):
25
+ if index in used:
26
+ continue
27
+ if item.entity_type != truth.entity_type:
28
+ continue
29
+ if iou((item.start, item.end), (truth.start, truth.end)) >= threshold:
30
+ used.add(index)
31
+ tp += 1
32
+ break
33
+ pred_n = sum(1 for item in pred if item.entity_type is not EntityType.UNKNOWN)
34
+ return tp, pred_n, len(gold)
35
+
36
+
37
+ def _exact_pairs(pred: list[Span], gold: list[GoldSpan]) -> tuple[int, int, int]:
38
+ used: set[int] = set()
39
+ tp = 0
40
+ for item in pred:
41
+ if item.entity_type is EntityType.UNKNOWN:
42
+ continue
43
+ for index, truth in enumerate(gold):
44
+ if index in used:
45
+ continue
46
+ if (
47
+ item.entity_type == truth.entity_type
48
+ and item.start == truth.start
49
+ and item.end == truth.end
50
+ ):
51
+ used.add(index)
52
+ tp += 1
53
+ break
54
+ pred_n = sum(1 for item in pred if item.entity_type is not EntityType.UNKNOWN)
55
+ return tp, pred_n, len(gold)
56
+
57
+
58
+ def _ratio(num: int, den: int) -> float:
59
+ return num / den if den else 1.0
60
+
61
+
62
+ def over_redaction_rate(pred: list[Span], gold: list[GoldSpan], text_len: int) -> float:
63
+ if text_len <= 0:
64
+ return 0.0
65
+ gold_chars: set[int] = set()
66
+ for item in gold:
67
+ gold_chars.update(range(item.start, item.end))
68
+ extra = 0
69
+ for item in pred:
70
+ if item.entity_type is EntityType.UNKNOWN:
71
+ continue
72
+ for pos in range(item.start, item.end):
73
+ if pos not in gold_chars:
74
+ extra += 1
75
+ return extra / text_len
76
+
77
+
78
+ def score_document(
79
+ document: LabDocument,
80
+ pred: list[Span],
81
+ export_text: str,
82
+ export_meta: dict[str, str] | None = None,
83
+ ) -> dict[str, float | int | list[str]]:
84
+ tp, pred_n, gold_n = _match_pairs(pred, document.gold_spans)
85
+ etp, epred, egold = _exact_pairs(pred, document.gold_spans)
86
+ gold_values = [item.surface for item in document.gold_spans] + [item.surface for item in document.gold_meta]
87
+ leaked = leftover_values(export_text, gold_values)
88
+ meta_blob = " ".join((export_meta or {}).values())
89
+ leaked.extend(leftover_values(meta_blob, [item.surface for item in document.gold_meta]))
90
+ hard_hits = leftover_values(" ".join(item.surface for item in pred), document.hard_negatives)
91
+ meta_leaks = leftover_values(meta_blob, [item.surface for item in document.gold_meta])
92
+ return {
93
+ "tp": tp,
94
+ "pred_n": pred_n,
95
+ "gold_n": gold_n,
96
+ "exact_tp": etp,
97
+ "exact_pred": epred,
98
+ "exact_gold": egold,
99
+ "over_redaction": over_redaction_rate(pred, document.gold_spans, len(document.body)),
100
+ "leaked": leaked,
101
+ "hard_hits": hard_hits,
102
+ "meta_leaks": meta_leaks,
103
+ }
104
+
105
+
106
+ def aggregate_scores(method: str, rows: list[dict[str, float | int | list[str]]], n_docs: int) -> MethodScore:
107
+ tp = sum(int(row["tp"]) for row in rows)
108
+ pred_n = sum(int(row["pred_n"]) for row in rows)
109
+ gold_n = sum(int(row["gold_n"]) for row in rows)
110
+ exact_tp = sum(int(row["exact_tp"]) for row in rows)
111
+ leftover = sum(1 for row in rows if row["leaked"])
112
+ meta = sum(len(row["meta_leaks"]) for row in rows)
113
+ hard = sum(len(row["hard_hits"]) for row in rows)
114
+ over = sum(float(row["over_redaction"]) for row in rows) / n_docs if n_docs else 0.0
115
+ residual = leftover / n_docs if n_docs else 0.0
116
+ return MethodScore(
117
+ method=method,
118
+ n_docs=n_docs,
119
+ n_pred=pred_n,
120
+ n_gold=gold_n,
121
+ precision=_ratio(tp, pred_n) if pred_n else 1.0,
122
+ recall=_ratio(tp, gold_n) if gold_n else 1.0,
123
+ exact_precision=_ratio(exact_tp, pred_n) if pred_n else 1.0,
124
+ exact_recall=_ratio(exact_tp, gold_n) if gold_n else 1.0,
125
+ over_redaction=over,
126
+ residual_leak=residual,
127
+ leftover_exports=leftover,
128
+ metadata_leaks=meta,
129
+ hardneg_hits=hard,
130
+ )
131
+
132
+
133
+ def per_group(documents: list[LabDocument], preds: list[list[Span]], key: str) -> dict[str, dict[str, float]]:
134
+ buckets: dict[str, list[tuple[LabDocument, list[Span]]]] = defaultdict(list)
135
+ for document, pred in zip(documents, preds):
136
+ label = document.language.value if key == "language" else "all"
137
+ if key == "type":
138
+ continue
139
+ buckets[label].append((document, pred))
140
+ if key == "type":
141
+ types = {span.entity_type.value for doc in documents for span in doc.gold_spans}
142
+ result = {}
143
+ for etype in sorted(types):
144
+ gold_n = pred_n = tp = 0
145
+ for document, pred in zip(documents, preds):
146
+ gold = [span for span in document.gold_spans if span.entity_type.value == etype]
147
+ chosen = [span for span in pred if span.entity_type.value == etype]
148
+ hit, pn, gn = _match_pairs(chosen, gold)
149
+ tp += hit
150
+ pred_n += pn
151
+ gold_n += gn
152
+ result[etype] = {"precision": _ratio(tp, pred_n) if pred_n else 1.0, "recall": _ratio(tp, gold_n) if gold_n else 1.0}
153
+ return result
154
+ result = {}
155
+ for label, pairs in buckets.items():
156
+ tp = pred_n = gold_n = 0
157
+ for document, pred in pairs:
158
+ hit, pn, gn = _match_pairs(pred, document.gold_spans)
159
+ tp += hit
160
+ pred_n += pn
161
+ gold_n += gn
162
+ result[label] = {"precision": _ratio(tp, pred_n) if pred_n else 1.0, "recall": _ratio(tp, gold_n) if gold_n else 1.0}
163
+ return result
164
+
165
+
166
+ def kpis_from(
167
+ n_docs: int,
168
+ policy: MethodScore,
169
+ baseline: MethodScore,
170
+ n_spans: int,
171
+ n_review: int,
172
+ n_redacted: int,
173
+ elapsed_ms: int,
174
+ pages: int,
175
+ chars: int,
176
+ action: PolicyAction,
177
+ ) -> Kpis:
178
+ return Kpis(
179
+ n_docs=n_docs,
180
+ n_spans=n_spans,
181
+ n_review=n_review,
182
+ n_redacted=n_redacted,
183
+ policy_precision=policy.precision,
184
+ policy_recall=policy.recall,
185
+ baseline_precision=baseline.precision,
186
+ baseline_recall=baseline.recall,
187
+ residual_leak=policy.residual_leak,
188
+ over_redaction=policy.over_redaction,
189
+ leftover_exports=policy.leftover_exports,
190
+ metadata_leaks=policy.metadata_leaks,
191
+ elapsed_ms=elapsed_ms,
192
+ engine_profile=ENGINE_PROFILE,
193
+ policy_action=action.value,
194
+ pages=pages,
195
+ chars=chars,
196
+ )
197
+
198
+
199
+ def comparison_rows(policy: MethodScore, baseline: MethodScore) -> list[dict[str, float | int | str]]:
200
+ return [
201
+ {
202
+ "method": item.method,
203
+ "precision": round(item.precision, 3),
204
+ "recall": round(item.recall, 3),
205
+ "over_redaction": round(item.over_redaction, 4),
206
+ "residual_leak": round(item.residual_leak, 3),
207
+ "leftover_exports": item.leftover_exports,
208
+ "metadata_leaks": item.metadata_leaks,
209
+ "hardneg_hits": item.hardneg_hits,
210
+ }
211
+ for item in (baseline, policy)
212
+ ]
demo/src/normalize.py ADDED
@@ -0,0 +1,28 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ from __future__ import annotations
2
+
3
+ import re
4
+
5
+ DIGIT_TABLE = str.maketrans("۰۱۲۳۴۵۶۷۸۹٠١٢٣٤٥٦٧٨٩", "01234567890123456789")
6
+ ZWNJ = "\u200c"
7
+ NBSP = "\u00a0"
8
+ HALF_SPACE_RE = re.compile(r"[\u200c\u200b\u00a0]")
9
+
10
+
11
+ def fold_digits(text: str) -> str:
12
+ return (text or "").translate(DIGIT_TABLE)
13
+
14
+
15
+ def squeeze_spaces(text: str) -> str:
16
+ return HALF_SPACE_RE.sub(" ", text or "")
17
+
18
+
19
+ def compact(text: str) -> str:
20
+ return fold_digits(squeeze_spaces(text or ""))
21
+
22
+
23
+ def only_digits(text: str) -> str:
24
+ return "".join(ch for ch in fold_digits(text or "") if ch.isdigit())
25
+
26
+
27
+ def strip_iban(text: str) -> str:
28
+ return re.sub(r"[\s\-]", "", fold_digits(text or "")).upper()
demo/src/parsers.py ADDED
@@ -0,0 +1,62 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ from __future__ import annotations
2
+
3
+ from pathlib import Path
4
+ from typing import Any
5
+
6
+ from src.domain import DocFormat, MAX_BYTES, MAX_CHARS, MAX_FILES
7
+ from src.pdfio import inspect_pdf
8
+ from src.schemas import ValidationReport
9
+
10
+
11
+ def read_upload(path: Path, report: ValidationReport) -> dict[str, Any] | None:
12
+ data = path.read_bytes()
13
+ if len(data) > MAX_BYTES:
14
+ report.add("too_large", f"File exceeds {MAX_BYTES} bytes.", path.name)
15
+ return None
16
+ suffix = path.suffix.lower()
17
+ if suffix == ".pdf":
18
+ extracted = inspect_pdf(data, report)
19
+ return {
20
+ "name": path.name,
21
+ "format": DocFormat.PDF,
22
+ "text": extracted["text"],
23
+ "raw": data,
24
+ "metadata": extracted["metadata"],
25
+ "pages": extracted["pages"],
26
+ }
27
+ try:
28
+ text = data.decode("utf-8")
29
+ except UnicodeDecodeError:
30
+ report.add("encoding", "Only UTF-8 text, CSV, or text PDFs are accepted.", path.name)
31
+ return None
32
+ if "\x00" in text:
33
+ report.add("binary", "Binary files are out of scope.", path.name)
34
+ return None
35
+ fmt = DocFormat.CSV if suffix == ".csv" else DocFormat.TEXT
36
+ return {"name": path.name, "format": fmt, "text": text, "raw": None, "metadata": {}, "pages": 0}
37
+
38
+
39
+ def parse_uploads(uploads: list[Any] | None, report: ValidationReport) -> list[dict[str, Any]]:
40
+ items: list[dict[str, Any]] = []
41
+ for upload in uploads or []:
42
+ path = Path(str(getattr(upload, "name", upload)))
43
+ parsed = read_upload(path, report)
44
+ if parsed:
45
+ items.append(parsed)
46
+ if len(items) > MAX_FILES:
47
+ report.add("too_many_files", f"At most {MAX_FILES} files are accepted.", "uploads")
48
+ total_chars = sum(len(item.get("text") or "") for item in items)
49
+ if total_chars > MAX_CHARS:
50
+ report.add("too_large_text", f"Combined text exceeds {MAX_CHARS} characters.", "uploads")
51
+ return items
52
+
53
+
54
+ def parse_pasted(text: str, report: ValidationReport) -> dict[str, Any] | None:
55
+ body = text or ""
56
+ if len(body) > MAX_CHARS:
57
+ report.add("too_large_text", f"Pasted text exceeds {MAX_CHARS} characters.", "text")
58
+ return None
59
+ if "\x00" in body:
60
+ report.add("binary", "Binary content is out of scope.", "text")
61
+ return None
62
+ return {"name": "pasted.txt", "format": DocFormat.TEXT, "text": body, "raw": None, "metadata": {}, "pages": 0}
demo/src/pdfio.py ADDED
@@ -0,0 +1,99 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ from __future__ import annotations
2
+
3
+ from io import BytesIO
4
+ from typing import Any
5
+
6
+ import fitz
7
+
8
+ from src.domain import MAX_PAGES
9
+ from src.normalize import fold_digits
10
+ from src.schemas import ValidationReport
11
+
12
+
13
+ def build_pdf(text: str, metadata: dict[str, str] | None = None) -> bytes:
14
+ doc = fitz.open()
15
+ page = doc.new_page()
16
+ # Keep a single text block so search_for can locate laboratory values.
17
+ page.insert_text((72, 72), text or " ", fontsize=11)
18
+ if metadata:
19
+ doc.set_metadata({key: value for key, value in metadata.items() if value})
20
+ payload = doc.tobytes()
21
+ doc.close()
22
+ return payload
23
+
24
+
25
+ def inspect_pdf(data: bytes, report: ValidationReport | None = None) -> dict[str, Any]:
26
+ issues = report or ValidationReport(ok=True)
27
+ try:
28
+ doc = fitz.open(stream=data, filetype="pdf")
29
+ except Exception:
30
+ issues.add("pdf_unreadable", "The PDF could not be opened.", "pdf")
31
+ return {"text": "", "pages": 0, "metadata": {}, "scanned": False, "report": issues}
32
+ pages = doc.page_count
33
+ if pages > MAX_PAGES:
34
+ issues.add("too_many_pages", f"PDFs are limited to {MAX_PAGES} text pages.", "pdf")
35
+ texts = []
36
+ for page in doc:
37
+ texts.append(page.get_text("text") or "")
38
+ body = "\n".join(texts).strip()
39
+ metadata = {key: str(value or "") for key, value in (doc.metadata or {}).items()}
40
+ scanned = pages > 0 and len(body) < max(12, pages * 8)
41
+ if scanned:
42
+ issues.add(
43
+ "scanned_pdf",
44
+ "Image-only / scanned PDFs are out of scope. This MVP reads a text layer only.",
45
+ "pdf",
46
+ severity="warning",
47
+ )
48
+ result = {
49
+ "text": body,
50
+ "pages": pages,
51
+ "metadata": metadata,
52
+ "scanned": scanned,
53
+ "report": issues,
54
+ }
55
+ doc.close()
56
+ return result
57
+
58
+
59
+ def redact_pdf(data: bytes, values: list[str], extra_metadata_clear: bool = True) -> tuple[bytes, dict[str, Any]]:
60
+ doc = fitz.open(stream=data, filetype="pdf")
61
+ hits = 0
62
+ folded_needles = [fold_digits(value) for value in values if value and len(value) >= 4]
63
+ for page in doc:
64
+ for value in values:
65
+ if not value or len(value) < 3:
66
+ continue
67
+ for rect in page.search_for(value):
68
+ page.add_redact_annot(rect, fill=(0, 0, 0))
69
+ hits += 1
70
+ page.apply_redactions()
71
+ if extra_metadata_clear:
72
+ empty = {key: "" for key in (doc.metadata or {})}
73
+ doc.set_metadata(empty)
74
+ try:
75
+ doc.del_xml_metadata()
76
+ except Exception:
77
+ pass
78
+ payload = doc.tobytes()
79
+ verify = inspect_pdf(payload)
80
+ leftover = []
81
+ extracted = fold_digits(verify["text"])
82
+ meta_blob = fold_digits(" ".join(verify["metadata"].values()))
83
+ for needle, original in zip(folded_needles, [v for v in values if v and len(v) >= 4]):
84
+ if needle in extracted or needle in meta_blob:
85
+ leftover.append(original)
86
+ doc.close()
87
+ return payload, {"hits": hits, "leftover": leftover, "pages": verify["pages"], "metadata": verify["metadata"], "text": verify["text"]}
88
+
89
+
90
+ def write_pdf(path, data: bytes) -> None:
91
+ path.write_bytes(data)
92
+
93
+
94
+ def pdf_from_path(path) -> bytes:
95
+ return path.read_bytes()
96
+
97
+
98
+ def buffer(data: bytes) -> BytesIO:
99
+ return BytesIO(data)
demo/src/pipeline.py ADDED
@@ -0,0 +1,325 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ from __future__ import annotations
2
+
3
+ import json
4
+ import time
5
+ from pathlib import Path
6
+ from typing import Any
7
+ from uuid import uuid4
8
+
9
+ from src.analyzer import analyze_policy, redactable
10
+ from src.baseline import analyze_regex_only
11
+ from src.domain import (
12
+ CODE_VERSION,
13
+ DEFAULT_SEED,
14
+ ENGINE_PROFILE,
15
+ PRODUCT,
16
+ DocFormat,
17
+ EntityType,
18
+ Family,
19
+ Language,
20
+ Method,
21
+ PolicyAction,
22
+ Split,
23
+ Workflow,
24
+ )
25
+ from src.explain import decide_verdict, finding_from_span, narrative, reasons, status_label
26
+ from src.generator import documents_by_id, scenario_by_id, scenarios
27
+ from src.metrics import aggregate_scores, comparison_rows, kpis_from, per_group, score_document
28
+ from src.parsers import parse_pasted, parse_uploads
29
+ from src.pdfio import build_pdf, redact_pdf
30
+ from src.redaction import apply_text, leftover_values
31
+ from src.reporting import input_hash, write_downloads
32
+ from src.schemas import LabDocument, Span, ValidationReport
33
+ from src.validation import validate_documents, validate_limits
34
+
35
+ ROOT = Path(__file__).resolve().parents[1]
36
+
37
+
38
+ def _user_document(name: str, text: str, fmt: DocFormat, metadata: dict[str, str] | None = None) -> LabDocument:
39
+ return LabDocument(
40
+ doc_id=f"USER-{name}",
41
+ template="user",
42
+ family=Family.USER.value,
43
+ split=Split.USER,
44
+ language=Language.MIXED,
45
+ format=fmt,
46
+ title=name,
47
+ body=text,
48
+ metadata=metadata or {},
49
+ description="Uploaded or pasted laboratory document",
50
+ description_fa="سند آزمایشگاهی بارگذاری‌شده یا چسبانده‌شده",
51
+ )
52
+
53
+
54
+ def _selected_documents(
55
+ scenario_id: str,
56
+ pasted: str,
57
+ uploads: list[Any] | None,
58
+ seed: int,
59
+ ) -> tuple[list[LabDocument], list[dict[str, Any]], ValidationReport, str]:
60
+ report = ValidationReport(ok=True)
61
+ parsed = parse_uploads(uploads, report)
62
+ extra: list[dict[str, Any]] = []
63
+ if pasted and pasted.strip():
64
+ item = parse_pasted(pasted, report)
65
+ if item:
66
+ parsed.insert(0, item)
67
+ if parsed:
68
+ documents = [
69
+ _user_document(item["name"], item["text"], item["format"], item.get("metadata") or {})
70
+ for item in parsed
71
+ ]
72
+ extra = parsed
73
+ validate_limits(len(parsed), sum(int(item.get("pages") or 0) for item in parsed), sum(len(item["text"]) for item in parsed), report)
74
+ report_docs = validate_documents(documents)
75
+ if not report_docs.ok:
76
+ report.issues.extend(report_docs.issues)
77
+ report.ok = False
78
+ return documents, extra, report, "upload"
79
+ if scenario_id == "S00":
80
+ report.add("empty", "Interactive mode needs pasted text or a file.", "text")
81
+ return [], [], report, "interactive"
82
+ scenario = scenario_by_id(scenario_id, seed)
83
+ catalog = documents_by_id(seed)
84
+ documents = [catalog[item] for item in scenario.doc_ids if item in catalog]
85
+ return documents, extra, validate_documents(documents), scenario_id
86
+
87
+
88
+ def _export_document(
89
+ document: LabDocument,
90
+ spans: list[Span],
91
+ action: PolicyAction,
92
+ upload: dict[str, Any] | None,
93
+ ) -> tuple[str, dict[str, str], bytes | None, list[str]]:
94
+ active = redactable(spans)
95
+ text_out = apply_text(document.body, active, action)
96
+ meta_out = {key: "" for key in document.metadata}
97
+ pdf_bytes = None
98
+ leftover: list[str] = []
99
+ values = [item.surface for item in active]
100
+ if document.format is DocFormat.PDF:
101
+ source = upload["raw"] if upload and upload.get("raw") else build_pdf(document.body, document.metadata)
102
+ pdf_bytes, info = redact_pdf(source, values)
103
+ text_out = info["text"] or text_out
104
+ meta_out = {key: str(value or "") for key, value in (info.get("metadata") or {}).items()}
105
+ leftover = info.get("leftover") or []
106
+ leftover.extend(leftover_values(text_out, [span.surface for span in document.gold_spans]))
107
+ leftover.extend(leftover_values(" ".join(meta_out.values()), [item.surface for item in document.gold_meta]))
108
+ return text_out, meta_out, pdf_bytes, leftover
109
+
110
+
111
+ def run_pipeline(
112
+ scenario_id: str = "S01",
113
+ pasted: str = "",
114
+ uploads: list[Any] | None = None,
115
+ lang: str = "en",
116
+ seed: int = DEFAULT_SEED,
117
+ action: PolicyAction | str = PolicyAction.MASK,
118
+ suppressed: list[str] | None = None,
119
+ ) -> dict[str, Any]:
120
+ started = time.perf_counter()
121
+ policy_action = action if isinstance(action, PolicyAction) else PolicyAction(action)
122
+ documents, extras, report, source = _selected_documents(scenario_id, pasted, uploads, seed)
123
+ run_id = f"ariadlp-{uuid4().hex[:16]}"
124
+ if not report.ok:
125
+ empty = aggregate_scores(Method.POLICY.value, [], 0)
126
+ kpis = kpis_from(0, empty, empty, 0, 0, 0, 0, 0, 0, policy_action)
127
+ return {
128
+ "run_id": run_id,
129
+ "scenario_id": scenario_id,
130
+ "verdict": decide_verdict(report, 0, 0, 0),
131
+ "status_label": status_label(decide_verdict(report, 0, 0, 0), lang),
132
+ "error": report.issues[0].message if report.issues else "Invalid input",
133
+ "kpis": kpis,
134
+ "validation": report,
135
+ "findings": [],
136
+ "finding_rows": [],
137
+ "doc_rows": [],
138
+ "span_rows": [],
139
+ "comparison": [],
140
+ "scores": {Method.POLICY: empty, Method.BASELINE: empty},
141
+ "reasons": [item.message for item in report.issues],
142
+ "narrative": "",
143
+ "documents": [],
144
+ "exports": {},
145
+ "before_html": "",
146
+ "after_html": "",
147
+ "redacted_text": "",
148
+ "engine_profile": ENGINE_PROFILE,
149
+ "input_hash": "",
150
+ "challenger_notes": [],
151
+ }
152
+
153
+ uploads_by_name = {item["name"]: item for item in extras}
154
+ policy_preds: list[list[Span]] = []
155
+ baseline_preds: list[list[Span]] = []
156
+ policy_rows: list[dict] = []
157
+ baseline_rows: list[dict] = []
158
+ findings = []
159
+ exports: dict[str, Any] = {}
160
+ suppressed_ids = set(suppressed or [])
161
+ pages = 0
162
+ chars = 0
163
+ first_before = ""
164
+ first_after = ""
165
+ first_redacted = ""
166
+ first_spans: list[Span] = []
167
+
168
+ for document in documents:
169
+ upload = uploads_by_name.get(document.title) or uploads_by_name.get(document.doc_id.replace("USER-", ""))
170
+ policy = analyze_policy(document.body)
171
+ baseline = analyze_regex_only(document.body)
172
+ if suppressed_ids:
173
+ for span in policy:
174
+ if span.finding_id in suppressed_ids:
175
+ span.workflow = Workflow.SUPPRESSED
176
+ policy_preds.append(policy)
177
+ baseline_preds.append(baseline)
178
+ export_text, export_meta, pdf_bytes, leftover = _export_document(document, policy, policy_action, upload)
179
+ exports[document.doc_id] = {
180
+ "text": export_text,
181
+ "metadata": export_meta,
182
+ "pdf": pdf_bytes,
183
+ "leftover": leftover,
184
+ "format": document.format.value,
185
+ }
186
+ policy_rows.append(score_document(document, policy, export_text, export_meta))
187
+ baseline_export = apply_text(document.body, redactable(baseline), policy_action)
188
+ baseline_rows.append(score_document(document, baseline, baseline_export, document.metadata))
189
+ findings.extend(finding_from_span(document.doc_id, span, lang) for span in policy)
190
+ chars += len(document.body)
191
+ if document.format is DocFormat.PDF:
192
+ pages += 1
193
+ if not first_redacted:
194
+ first_before = document.body
195
+ first_after = export_text
196
+ first_redacted = export_text
197
+ first_spans = policy
198
+
199
+ policy_score = aggregate_scores(Method.POLICY.value, policy_rows, len(documents))
200
+ policy_score.by_type = per_group(documents, policy_preds, "type")
201
+ policy_score.by_language = per_group(documents, policy_preds, "language")
202
+ baseline_score = aggregate_scores(Method.BASELINE.value, baseline_rows, len(documents))
203
+ n_review = sum(1 for item in findings if item.workflow == Workflow.REVIEW.value or item.entity_type == EntityType.UNKNOWN.value)
204
+ n_redacted = sum(1 for item in findings if item.workflow == Workflow.DETECTED.value)
205
+ elapsed = int((time.perf_counter() - started) * 1000)
206
+ kpis = kpis_from(
207
+ len(documents),
208
+ policy_score,
209
+ baseline_score,
210
+ len(findings),
211
+ n_review,
212
+ n_redacted,
213
+ elapsed,
214
+ pages,
215
+ chars,
216
+ policy_action,
217
+ )
218
+ verdict = decide_verdict(report, len(documents), policy_score.leftover_exports, n_review)
219
+ from src.redaction import highlight_html, highlight_redacted_html
220
+
221
+ rtl = lang == "fa"
222
+ finding_rows = [item.model_dump() for item in findings]
223
+ span_rows = [
224
+ {
225
+ "keep": item.workflow != Workflow.SUPPRESSED.value,
226
+ "finding_id": item.finding_id,
227
+ "doc_id": item.doc_id,
228
+ "type": item.entity_type,
229
+ "start": item.start,
230
+ "end": item.end,
231
+ "preview": item.surface_preview,
232
+ "confidence": item.confidence,
233
+ "workflow": item.workflow,
234
+ "rule_id": item.rule_id,
235
+ }
236
+ for item in findings
237
+ ]
238
+ doc_rows = [
239
+ {
240
+ "doc_id": document.doc_id,
241
+ "template": document.template,
242
+ "language": document.language.value,
243
+ "format": document.format.value,
244
+ "split": document.split.value,
245
+ "n_gold": len(document.gold_spans),
246
+ "leftover": len(exports[document.doc_id]["leftover"]),
247
+ }
248
+ for document in documents
249
+ ]
250
+ hashed = input_hash(
251
+ {
252
+ "scenario": scenario_id,
253
+ "docs": [item.doc_id for item in documents],
254
+ "action": policy_action.value,
255
+ "version": CODE_VERSION,
256
+ }
257
+ )
258
+ return {
259
+ "run_id": run_id,
260
+ "product": PRODUCT,
261
+ "scenario_id": scenario_id if source != "upload" else "upload",
262
+ "source": source,
263
+ "verdict": verdict,
264
+ "status_label": status_label(verdict, lang),
265
+ "error": None,
266
+ "kpis": kpis,
267
+ "validation": report,
268
+ "findings": findings,
269
+ "finding_rows": finding_rows,
270
+ "doc_rows": doc_rows,
271
+ "span_rows": span_rows,
272
+ "comparison": comparison_rows(policy_score, baseline_score),
273
+ "scores": {Method.POLICY: policy_score, Method.BASELINE: baseline_score},
274
+ "reasons": reasons(policy_score, baseline_score, lang),
275
+ "narrative": narrative(kpis, lang),
276
+ "documents": documents,
277
+ "exports": exports,
278
+ "before_html": highlight_html(first_before, first_spans, rtl=rtl),
279
+ "after_html": highlight_redacted_html(first_after, rtl=rtl),
280
+ "redacted_text": first_redacted,
281
+ "original_text": first_before,
282
+ "engine_profile": ENGINE_PROFILE,
283
+ "input_hash": hashed,
284
+ "policy_action": policy_action.value,
285
+ "challenger_notes": [
286
+ {
287
+ "doc_id": item.doc_id,
288
+ "note": "Unknown / failed-checksum items stay on the review list and are not auto-redacted.",
289
+ }
290
+ for item in findings
291
+ if item.entity_type == EntityType.UNKNOWN.value
292
+ ][:20],
293
+ }
294
+
295
+
296
+ def evaluate_catalog(seed: int = DEFAULT_SEED) -> list[dict[str, Any]]:
297
+ rows = []
298
+ for scenario in scenarios(seed):
299
+ result = run_pipeline(scenario.scenario_id, "", None, "en", seed, PolicyAction.MASK)
300
+ policy = result["scores"][Method.POLICY]
301
+ rows.append(
302
+ {
303
+ "scenario_id": scenario.scenario_id,
304
+ "method": Method.POLICY.value,
305
+ "precision": round(policy.precision, 3),
306
+ "recall": round(policy.recall, 3),
307
+ "over_redaction": round(policy.over_redaction, 4),
308
+ "residual_leak": round(policy.residual_leak, 3),
309
+ "leftover_exports": policy.leftover_exports,
310
+ "metadata_leaks": policy.metadata_leaks,
311
+ "hardneg_hits": policy.hardneg_hits,
312
+ "n_docs": policy.n_docs,
313
+ }
314
+ )
315
+ return rows
316
+
317
+
318
+ def result_to_downloads(result: dict[str, Any], folder: Path) -> dict[str, str]:
319
+ return write_downloads(result, folder)
320
+
321
+
322
+ def load_sample(scenario_id: str = "S01") -> LabDocument:
323
+ catalog = documents_by_id()
324
+ scenario = scenario_by_id(scenario_id)
325
+ return catalog[scenario.doc_ids[0]]
demo/src/plots.py ADDED
@@ -0,0 +1,101 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ from __future__ import annotations
2
+
3
+ from typing import Any
4
+
5
+ import pandas as pd
6
+ import plotly.graph_objects as go
7
+
8
+ from src.i18n import t
9
+ from src.schemas import Kpis, MethodScore
10
+
11
+
12
+ def empty_plot(title: str = "—") -> go.Figure:
13
+ fig = go.Figure()
14
+ fig.update_layout(template="plotly_white", height=340, title=title)
15
+ return fig
16
+
17
+
18
+ def kpi_html(kpis: dict[str, Any] | Kpis, lang: str) -> str:
19
+ data = kpis.model_dump() if isinstance(kpis, Kpis) else kpis
20
+
21
+ def card(label: str, value: str, sub: str) -> str:
22
+ return (
23
+ f"<article class='kpi-card'><p class='kpi-label'>{label}</p>"
24
+ f"<p class='kpi-value'>{value}</p><p class='kpi-sub'>{sub}</p></article>"
25
+ )
26
+
27
+ return (
28
+ "<div class='kpi-grid'>"
29
+ + card(t(lang, "kpi_docs"), str(data.get("n_docs") or 0), data.get("engine_profile") or "policy")
30
+ + card(t(lang, "kpi_spans"), str(data.get("n_spans") or 0), t(lang, "optimized"))
31
+ + card(
32
+ t(lang, "kpi_precision"),
33
+ f"{100 * float(data.get('policy_precision') or 0):.0f}%",
34
+ f"baseline {100 * float(data.get('baseline_precision') or 0):.0f}%",
35
+ )
36
+ + card(
37
+ t(lang, "kpi_recall"),
38
+ f"{100 * float(data.get('policy_recall') or 0):.0f}%",
39
+ f"baseline {100 * float(data.get('baseline_recall') or 0):.0f}%",
40
+ )
41
+ + card(
42
+ t(lang, "kpi_left"),
43
+ str(data.get("leftover_exports") or 0),
44
+ f"{data.get('elapsed_ms') or 0} ms",
45
+ )
46
+ + "</div>"
47
+ )
48
+
49
+
50
+ def compare_figure(scores: dict[str, MethodScore] | dict[str, Any], title: str) -> go.Figure:
51
+ if not scores:
52
+ return empty_plot(title)
53
+ names, precision, recall, over = [], [], [], []
54
+ for key in ("regex_only", "checksum_policy"):
55
+ item = None
56
+ for raw_key, value in scores.items():
57
+ name = raw_key.value if hasattr(raw_key, "value") else str(raw_key)
58
+ if name == key or (hasattr(raw_key, "value") and raw_key.value == key):
59
+ item = value
60
+ break
61
+ if str(raw_key).endswith(key):
62
+ item = value
63
+ if item is None:
64
+ continue
65
+ names.append(key)
66
+ precision.append(round(100 * float(item.precision if hasattr(item, "precision") else item.get("precision", 0)), 1))
67
+ recall.append(round(100 * float(item.recall if hasattr(item, "recall") else item.get("recall", 0)), 1))
68
+ over.append(round(100 * float(item.over_redaction if hasattr(item, "over_redaction") else item.get("over_redaction", 0)), 2))
69
+ fig = go.Figure()
70
+ fig.add_bar(name="precision %", x=names, y=precision, marker_color="#0f766e")
71
+ fig.add_bar(name="recall %", x=names, y=recall, marker_color="#1d4ed8")
72
+ fig.add_bar(name="over-redaction %", x=names, y=over, marker_color="#b45309")
73
+ fig.update_layout(template="plotly_white", height=380, barmode="group", title=title)
74
+ return fig
75
+
76
+
77
+ def type_figure(score: MethodScore | dict[str, Any] | None, title: str) -> go.Figure:
78
+ if score is None:
79
+ return empty_plot(title)
80
+ by_type = score.by_type if hasattr(score, "by_type") else (score or {}).get("by_type") or {}
81
+ if not by_type:
82
+ return empty_plot(title)
83
+ names = list(by_type.keys())
84
+ precision = [round(100 * float(by_type[name].get("precision", 0)), 1) for name in names]
85
+ recall = [round(100 * float(by_type[name].get("recall", 0)), 1) for name in names]
86
+ fig = go.Figure()
87
+ fig.add_bar(name="precision %", x=names, y=precision, marker_color="#4338ca")
88
+ fig.add_bar(name="recall %", x=names, y=recall, marker_color="#7c3aed")
89
+ fig.update_layout(template="plotly_white", height=360, barmode="group", title=title)
90
+ return fig
91
+
92
+
93
+ def leftover_figure(doc_rows: list[dict[str, Any]], title: str) -> go.Figure:
94
+ if not doc_rows:
95
+ return empty_plot(title)
96
+ frame = pd.DataFrame(doc_rows)
97
+ if "leftover" not in frame.columns:
98
+ return empty_plot(title)
99
+ fig = go.Figure(go.Bar(x=list(frame["doc_id"].astype(str)), y=list(frame["leftover"]), marker_color="#be123c"))
100
+ fig.update_layout(template="plotly_white", height=360, title=title)
101
+ return fig
demo/src/recognizers.py ADDED
@@ -0,0 +1,303 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ from __future__ import annotations
2
+
3
+ import re
4
+ from dataclasses import dataclass
5
+
6
+ from src.checksums import iban_ok, iran_national_id_ok, luhn_ok
7
+ from src.domain import EntityType, Method, Workflow
8
+ from src.normalize import compact, fold_digits, only_digits
9
+ from src.schemas import Span
10
+
11
+ EMAIL_RE = re.compile(r"\b[A-Za-z0-9._%+-]+@[A-Za-z0-9.-]+\.[A-Za-z]{2,}\b")
12
+ PHONE_ASCII_RE = re.compile(r"(?<!\d)(?:\+98|0098|0)9\d{9}(?!\d)")
13
+ PHONE_LOOSE_RE = re.compile(r"(?<!\d)(?:\+\d{1,3}[\s-]?)?(?:0?9\d{2}[\s-]?\d{3}[\s-]?\d{4})(?!\d)")
14
+ NID_RE = re.compile(r"(?<!\d)\d{10}(?!\d)")
15
+ CARD_RE = re.compile(r"(?<!\d)(?:\d[ -]?){13,19}(?!\d)")
16
+ IBAN_RE = re.compile(r"\b[A-Za-z]{2}\d{2}[A-Za-z0-9]{11,30}\b")
17
+ EMPLOYEE_RE = re.compile(r"\b(?:ARIA|EMP)-\d{4,6}\b", re.I)
18
+ ISBN_RE = re.compile(r"\b97[89](?:-?\d){10}\b")
19
+ VERSION_RE = re.compile(r"\bv?\d+\.\d+\.\d+\b", re.I)
20
+ INVOICE_RE = re.compile(r"\b(?:INV|ORD|SKU|TRK)[-_A-Za-z0-9]{3,20}\b", re.I)
21
+ DATE_RE = re.compile(r"\b(?:13|14|19|20)\d{2}[/-]\d{1,2}[/-]\d{1,2}\b")
22
+ PROTOCOL_NAME_RE = re.compile(r"\bMr\.?\s+Protocol\b", re.I)
23
+ HONORIFIC_EN = re.compile(r"\b(?:Mr|Ms|Mrs|Dr)\.?\s+([A-Z][a-z]+(?:\s+[A-Z][a-z]+){0,2})")
24
+ HONORIFIC_FA = re.compile(r"(?:آقای|خانم|دکتر)\s+([\u0600-\u06FF\u200c]+(?:\s+[\u0600-\u06FF\u200c]+){0,2})")
25
+ LABEL_NAME_EN = re.compile(r"(?i)\b(?:name|employee name|full name)\s*[:：]\s*([A-Z][A-Za-z\s]{2,40})")
26
+ LABEL_NAME_FA = re.compile(r"(?:نام(?:\s+و\s+نام\s+خانوادگی)?|نام\s+کارمند)\s*[:：]\s*([\u0600-\u06FF\u200c\s]{2,40})")
27
+ NID_CONTEXT_RE = re.compile(r"(?:کد[\u200c\s]*ملی|national\s+id|national\s+code)\s*[:：]?\s*", re.I)
28
+
29
+ LAB_FIRST = {
30
+ "Ava",
31
+ "Noah",
32
+ "Leila",
33
+ "Omar",
34
+ "Helen",
35
+ "Nima",
36
+ "سارا",
37
+ "نیما",
38
+ "لیلا",
39
+ "امید",
40
+ "هلن",
41
+ "رضا",
42
+ }
43
+ LAB_LAST = {
44
+ "Labian",
45
+ "Fixture",
46
+ "Rahimi",
47
+ "Azadi",
48
+ "نمونه‌ای",
49
+ "آزمایشی",
50
+ "رحیمی",
51
+ "آزادی",
52
+ }
53
+
54
+
55
+ @dataclass(frozen=True)
56
+ class Recognizer:
57
+ recognizer_id: str
58
+ entity_type: EntityType
59
+ language: str = "all"
60
+
61
+ def analyze(self, text: str) -> list[Span]:
62
+ raise NotImplementedError
63
+
64
+
65
+ def _span(
66
+ text: str,
67
+ start: int,
68
+ end: int,
69
+ entity_type: EntityType,
70
+ rule_id: str,
71
+ confidence: float,
72
+ reason: str,
73
+ workflow: Workflow = Workflow.DETECTED,
74
+ language: str = "en",
75
+ ) -> Span:
76
+ return Span(
77
+ start=start,
78
+ end=end,
79
+ entity_type=entity_type,
80
+ surface=text[start:end],
81
+ confidence=confidence,
82
+ method=Method.POLICY.value,
83
+ rule_id=rule_id,
84
+ language=language,
85
+ workflow=workflow,
86
+ reason=reason,
87
+ )
88
+
89
+
90
+ def _window(text: str, start: int, radius: int = 32) -> str:
91
+ return text[max(0, start - radius) : min(len(text), start + radius)]
92
+
93
+
94
+ class EmailRecognizer(Recognizer):
95
+ def __init__(self) -> None:
96
+ super().__init__("email", EntityType.EMAIL)
97
+
98
+ def analyze(self, text: str) -> list[Span]:
99
+ found = []
100
+ for match in EMAIL_RE.finditer(text):
101
+ found.append(
102
+ _span(text, match.start(), match.end(), EntityType.EMAIL, "email.rfc_like", 0.92, "Email pattern")
103
+ )
104
+ return found
105
+
106
+
107
+ class PhoneRecognizer(Recognizer):
108
+ def __init__(self) -> None:
109
+ super().__init__("phone", EntityType.PHONE)
110
+
111
+ def analyze(self, text: str) -> list[Span]:
112
+ folded = fold_digits(text)
113
+ found = []
114
+ for match in PHONE_ASCII_RE.finditer(folded):
115
+ found.append(
116
+ _span(
117
+ text,
118
+ match.start(),
119
+ match.end(),
120
+ EntityType.PHONE,
121
+ "phone.ir_mobile",
122
+ 0.9,
123
+ "Iranian mobile after digit fold",
124
+ language="fa" if text[match.start() : match.end()] != folded[match.start() : match.end()] else "en",
125
+ )
126
+ )
127
+ if not found:
128
+ for match in PHONE_LOOSE_RE.finditer(compact(text)):
129
+ raw = only_digits(match.group(0))
130
+ if len(raw) in {10, 11, 12, 13}:
131
+ # Offsets on compact() can drift; fall back to digit fold alignment.
132
+ continue
133
+ return found
134
+
135
+
136
+ class NationalIdRecognizer(Recognizer):
137
+ def __init__(self) -> None:
138
+ super().__init__("iran_national_id", EntityType.IR_NATIONAL_ID)
139
+
140
+ def analyze(self, text: str) -> list[Span]:
141
+ folded = fold_digits(text)
142
+ found: list[Span] = []
143
+ for match in NID_RE.finditer(folded):
144
+ value = match.group(0)
145
+ left = _window(folded, match.start())
146
+ hinted = bool(NID_CONTEXT_RE.search(left))
147
+ if iran_national_id_ok(value):
148
+ found.append(
149
+ _span(
150
+ text,
151
+ match.start(),
152
+ match.end(),
153
+ EntityType.IR_NATIONAL_ID,
154
+ "nid.iran_checksum",
155
+ 0.96,
156
+ "Iranian national ID checksum passed",
157
+ language="fa",
158
+ )
159
+ )
160
+ elif hinted:
161
+ found.append(
162
+ _span(
163
+ text,
164
+ match.start(),
165
+ match.end(),
166
+ EntityType.UNKNOWN,
167
+ "nid.context_invalid",
168
+ 0.35,
169
+ "National-ID context with failed checksum; left for review",
170
+ workflow=Workflow.REVIEW,
171
+ language="fa",
172
+ )
173
+ )
174
+ return found
175
+
176
+
177
+ class CardRecognizer(Recognizer):
178
+ def __init__(self) -> None:
179
+ super().__init__("credit_card", EntityType.CREDIT_CARD)
180
+
181
+ def analyze(self, text: str) -> list[Span]:
182
+ found = []
183
+ for match in CARD_RE.finditer(fold_digits(text)):
184
+ digits = only_digits(match.group(0))
185
+ if digits.startswith(("978", "979")):
186
+ continue
187
+ if 13 <= len(digits) <= 19 and luhn_ok(digits):
188
+ found.append(
189
+ _span(
190
+ text,
191
+ match.start(),
192
+ match.end(),
193
+ EntityType.CREDIT_CARD,
194
+ "card.luhn",
195
+ 0.95,
196
+ "Luhn-valid laboratory PAN",
197
+ )
198
+ )
199
+ return found
200
+
201
+
202
+ class IbanRecognizer(Recognizer):
203
+ def __init__(self) -> None:
204
+ super().__init__("iban", EntityType.IBAN)
205
+
206
+ def analyze(self, text: str) -> list[Span]:
207
+ found = []
208
+ folded = fold_digits(text)
209
+ for match in IBAN_RE.finditer(folded):
210
+ if iban_ok(match.group(0)):
211
+ found.append(
212
+ _span(text, match.start(), match.end(), EntityType.IBAN, "iban.mod97", 0.95, "IBAN mod-97 passed")
213
+ )
214
+ return found
215
+
216
+
217
+ class EmployeeRecognizer(Recognizer):
218
+ def __init__(self) -> None:
219
+ super().__init__("employee_id", EntityType.EMPLOYEE_ID)
220
+
221
+ def analyze(self, text: str) -> list[Span]:
222
+ return [
223
+ _span(text, match.start(), match.end(), EntityType.EMPLOYEE_ID, "employee.aria", 0.88, "Laboratory employee id")
224
+ for match in EMPLOYEE_RE.finditer(text)
225
+ ]
226
+
227
+
228
+ class PersonRecognizer(Recognizer):
229
+ def __init__(self) -> None:
230
+ super().__init__("person", EntityType.PERSON)
231
+
232
+ def analyze(self, text: str) -> list[Span]:
233
+ found: list[Span] = []
234
+ if PROTOCOL_NAME_RE.search(text):
235
+ # Explicit hard negative: software protocol, not a person.
236
+ pass
237
+ for match in HONORIFIC_EN.finditer(text):
238
+ surface = match.group(0)
239
+ if PROTOCOL_NAME_RE.search(surface):
240
+ continue
241
+ found.append(_span(text, match.start(), match.end(), EntityType.PERSON, "person.honorific_en", 0.72, "English honorific"))
242
+ for match in HONORIFIC_FA.finditer(text):
243
+ found.append(
244
+ _span(
245
+ text,
246
+ match.start(),
247
+ match.end(),
248
+ EntityType.PERSON,
249
+ "person.honorific_fa",
250
+ 0.74,
251
+ "Persian honorific",
252
+ language="fa",
253
+ )
254
+ )
255
+ for match in LABEL_NAME_EN.finditer(text):
256
+ start, end = match.start(1), match.end(1)
257
+ found.append(_span(text, start, end, EntityType.PERSON, "person.label_en", 0.7, "Labeled English name"))
258
+ for match in LABEL_NAME_FA.finditer(text):
259
+ start, end = match.start(1), match.end(1)
260
+ found.append(
261
+ _span(text, start, end, EntityType.PERSON, "person.label_fa", 0.72, "Labeled Persian name", language="fa")
262
+ )
263
+ for first in LAB_FIRST:
264
+ for last in LAB_LAST:
265
+ needle = f"{first} {last}"
266
+ start = 0
267
+ while True:
268
+ index = text.find(needle, start)
269
+ if index < 0:
270
+ break
271
+ found.append(
272
+ _span(
273
+ text,
274
+ index,
275
+ index + len(needle),
276
+ EntityType.PERSON,
277
+ "person.lab_lexicon",
278
+ 0.8,
279
+ "Laboratory name lexicon",
280
+ language="fa" if any(ord(ch) > 127 for ch in needle) else "en",
281
+ )
282
+ )
283
+ start = index + len(needle)
284
+ return found
285
+
286
+
287
+ def hard_negative_surfaces(text: str) -> list[str]:
288
+ surfaces: list[str] = []
289
+ for pattern in (ISBN_RE, VERSION_RE, INVOICE_RE, DATE_RE, PROTOCOL_NAME_RE):
290
+ surfaces.extend(match.group(0) for match in pattern.finditer(text))
291
+ return surfaces
292
+
293
+
294
+ def default_recognizers() -> list[Recognizer]:
295
+ return [
296
+ EmailRecognizer(),
297
+ PhoneRecognizer(),
298
+ NationalIdRecognizer(),
299
+ CardRecognizer(),
300
+ IbanRecognizer(),
301
+ EmployeeRecognizer(),
302
+ PersonRecognizer(),
303
+ ]
demo/src/redaction.py ADDED
@@ -0,0 +1,76 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ from __future__ import annotations
2
+
3
+ import hashlib
4
+ import html
5
+ import re
6
+
7
+ from src.domain import PolicyAction, RULESET_ID
8
+ from src.normalize import fold_digits
9
+ from src.schemas import Span
10
+
11
+
12
+ def fingerprint(value: str) -> str:
13
+ digest = hashlib.sha256(f"{RULESET_ID}:{value}".encode("utf-8")).hexdigest()
14
+ return digest[:10]
15
+
16
+
17
+ def mask_token(span: Span) -> str:
18
+ return f"[{span.entity_type.value}:{fingerprint(span.surface)}]"
19
+
20
+
21
+ def apply_text(text: str, spans: list[Span], action: PolicyAction) -> str:
22
+ if not text:
23
+ return ""
24
+ ordered = sorted(spans, key=lambda item: item.start)
25
+ chunks: list[str] = []
26
+ cursor = 0
27
+ for item in ordered:
28
+ if item.start < cursor or item.end > len(text) or item.start >= item.end:
29
+ continue
30
+ chunks.append(text[cursor : item.start])
31
+ chunks.append("" if action is PolicyAction.REMOVE else mask_token(item))
32
+ cursor = item.end
33
+ chunks.append(text[cursor:])
34
+ return "".join(chunks)
35
+
36
+
37
+ def leftover_values(export_text: str, values: list[str]) -> list[str]:
38
+ haystack = fold_digits(export_text or "")
39
+ leaked: list[str] = []
40
+ for value in values:
41
+ needle = fold_digits(value or "")
42
+ if needle and len(needle) >= 4 and needle in haystack:
43
+ leaked.append(value)
44
+ return leaked
45
+
46
+
47
+ def highlight_html(text: str, spans: list[Span], rtl: bool = False) -> str:
48
+ direction = "rtl-mode" if rtl else ""
49
+ if not text:
50
+ return f"<pre class='doc-view {direction}'></pre>"
51
+ ordered = sorted(spans, key=lambda item: (item.start, item.end))
52
+ chunks: list[str] = []
53
+ cursor = 0
54
+ for item in ordered:
55
+ if item.start < cursor or item.end > len(text):
56
+ continue
57
+ chunks.append(html.escape(text[cursor : item.start]))
58
+ title = html.escape(f"{item.entity_type.value} {item.confidence:.2f}")
59
+ chunks.append(f"<mark class='pii-span' title='{title}'>{html.escape(text[item.start:item.end])}</mark>")
60
+ cursor = item.end
61
+ chunks.append(html.escape(text[cursor:]))
62
+ return f"<pre class='doc-view {direction}'>{''.join(chunks)}</pre>"
63
+
64
+
65
+ def redact_csv_cells(text: str, spans: list[Span], action: PolicyAction) -> str:
66
+ return apply_text(text, spans, action)
67
+
68
+
69
+ TOKEN_RE = re.compile(r"\[[A-Z_]+:[0-9a-f]{10}\]")
70
+
71
+
72
+ def highlight_redacted_html(text: str, rtl: bool = False) -> str:
73
+ direction = "rtl-mode" if rtl else ""
74
+ escaped = html.escape(text or "")
75
+ marked = TOKEN_RE.sub(lambda match: f"<mark class='redact-span'>{match.group(0)}</mark>", escaped)
76
+ return f"<pre class='doc-view {direction}'>{marked}</pre>"
demo/src/reporting.py ADDED
@@ -0,0 +1,70 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ from __future__ import annotations
2
+
3
+ import hashlib
4
+ import json
5
+ from pathlib import Path
6
+ from typing import Any
7
+
8
+ import pandas as pd
9
+
10
+ from src.domain import CODE_VERSION, PRODUCT, RULESET_ID
11
+
12
+
13
+ def input_hash(payload: dict[str, Any]) -> str:
14
+ raw = json.dumps(payload, sort_keys=True, ensure_ascii=False, default=str)
15
+ return hashlib.sha256(raw.encode("utf-8")).hexdigest()[:16]
16
+
17
+
18
+ def public_report(result: dict[str, Any]) -> dict[str, Any]:
19
+ kpis = result["kpis"]
20
+ return {
21
+ "product": PRODUCT,
22
+ "code_version": CODE_VERSION,
23
+ "ruleset": RULESET_ID,
24
+ "run_id": result["run_id"],
25
+ "scenario_id": result.get("scenario_id"),
26
+ "input_hash": result.get("input_hash"),
27
+ "verdict": result["verdict"].value if hasattr(result["verdict"], "value") else result["verdict"],
28
+ "kpis": kpis.model_dump() if hasattr(kpis, "model_dump") else kpis,
29
+ "comparison": result.get("comparison") or [],
30
+ "finding_count": len(result.get("finding_rows") or []),
31
+ "engine_profile": result.get("engine_profile"),
32
+ "policy_action": result.get("policy_action"),
33
+ "limits": (
34
+ "Detector-scope redaction only. This report does not claim that every sensitive "
35
+ "value was removed. Scanned PDFs and production identity data are out of scope."
36
+ ),
37
+ }
38
+
39
+
40
+ def write_downloads(result: dict[str, Any], folder: Path) -> dict[str, str]:
41
+ folder.mkdir(parents=True, exist_ok=True)
42
+ run_id = result["run_id"]
43
+ files = {
44
+ "documents": folder / f"{run_id}_documents.csv",
45
+ "findings": folder / f"{run_id}_findings.csv",
46
+ "spans": folder / f"{run_id}_spans.csv",
47
+ "comparison": folder / f"{run_id}_comparison.csv",
48
+ "report": folder / f"{run_id}_report.json",
49
+ "redacted": folder / f"{run_id}_redacted.txt",
50
+ }
51
+ pd.DataFrame(result.get("doc_rows") or []).to_csv(files["documents"], index=False)
52
+ pd.DataFrame(result.get("finding_rows") or []).to_csv(files["findings"], index=False)
53
+ pd.DataFrame(result.get("span_rows") or []).to_csv(files["spans"], index=False)
54
+ pd.DataFrame(result.get("comparison") or []).to_csv(files["comparison"], index=False)
55
+ files["report"].write_text(
56
+ json.dumps(public_report(result), indent=2, ensure_ascii=False, default=str),
57
+ encoding="utf-8",
58
+ newline="\n",
59
+ )
60
+ files["redacted"].write_text(result.get("redacted_text") or "", encoding="utf-8", newline="\n")
61
+ exports = result.get("exports") or {}
62
+ for doc_id, payload in exports.items():
63
+ if payload.get("pdf"):
64
+ pdf_path = folder / f"{run_id}_{doc_id}.pdf"
65
+ pdf_path.write_bytes(payload["pdf"])
66
+ files[f"pdf_{doc_id}"] = pdf_path
67
+ text_path = folder / f"{run_id}_{doc_id}.txt"
68
+ text_path.write_text(payload.get("text") or "", encoding="utf-8", newline="\n")
69
+ files[f"text_{doc_id}"] = text_path
70
+ return {key: str(path) for key, path in files.items()}
demo/src/schemas.py ADDED
@@ -0,0 +1,175 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ from __future__ import annotations
2
+
3
+ from typing import Any
4
+
5
+ from pydantic import BaseModel, Field
6
+
7
+ from src.domain import (
8
+ DocFormat,
9
+ EntityType,
10
+ Language,
11
+ Method,
12
+ PolicyAction,
13
+ Split,
14
+ Verdict,
15
+ Workflow,
16
+ )
17
+
18
+
19
+ class Issue(BaseModel):
20
+ code: str
21
+ message: str
22
+ locator: str | None = None
23
+ severity: str = "error"
24
+
25
+
26
+ class ValidationReport(BaseModel):
27
+ ok: bool
28
+ issues: list[Issue] = Field(default_factory=list)
29
+
30
+ def add(self, code: str, message: str, locator: str | None = None, severity: str = "error") -> None:
31
+ self.issues.append(Issue(code=code, message=message, locator=locator, severity=severity))
32
+ if severity == "error":
33
+ self.ok = False
34
+
35
+
36
+ class Span(BaseModel):
37
+ start: int
38
+ end: int
39
+ entity_type: EntityType
40
+ surface: str
41
+ confidence: float
42
+ method: str
43
+ rule_id: str
44
+ language: str = "en"
45
+ workflow: Workflow = Workflow.DETECTED
46
+ finding_id: str = ""
47
+ reason: str = ""
48
+ page: int | None = None
49
+ field: str | None = None
50
+
51
+
52
+ class GoldSpan(BaseModel):
53
+ start: int
54
+ end: int
55
+ entity_type: EntityType
56
+ surface: str
57
+ language: str = "en"
58
+ field: str | None = None
59
+
60
+
61
+ class GoldMeta(BaseModel):
62
+ field: str
63
+ entity_type: EntityType
64
+ surface: str
65
+
66
+
67
+ class LabDocument(BaseModel):
68
+ doc_id: str
69
+ template: str
70
+ family: str
71
+ split: Split
72
+ language: Language
73
+ format: DocFormat
74
+ title: str
75
+ body: str
76
+ metadata: dict[str, str] = Field(default_factory=dict)
77
+ gold_spans: list[GoldSpan] = Field(default_factory=list)
78
+ gold_meta: list[GoldMeta] = Field(default_factory=list)
79
+ hard_negatives: list[str] = Field(default_factory=list)
80
+ description: str = ""
81
+ description_fa: str = ""
82
+
83
+
84
+ class Scenario(BaseModel):
85
+ scenario_id: str
86
+ title: str
87
+ description: str
88
+ description_fa: str
89
+ doc_ids: list[str] = Field(default_factory=list)
90
+ split: str = "mixed"
91
+
92
+
93
+ class MethodScore(BaseModel):
94
+ method: str
95
+ n_docs: int = 0
96
+ n_pred: int = 0
97
+ n_gold: int = 0
98
+ precision: float = 0.0
99
+ recall: float = 0.0
100
+ exact_precision: float = 0.0
101
+ exact_recall: float = 0.0
102
+ over_redaction: float = 0.0
103
+ residual_leak: float = 0.0
104
+ leftover_exports: int = 0
105
+ metadata_leaks: int = 0
106
+ hardneg_hits: int = 0
107
+ by_type: dict[str, dict[str, float]] = Field(default_factory=dict)
108
+ by_language: dict[str, dict[str, float]] = Field(default_factory=dict)
109
+
110
+
111
+ class Kpis(BaseModel):
112
+ n_docs: int = 0
113
+ n_spans: int = 0
114
+ n_review: int = 0
115
+ n_redacted: int = 0
116
+ policy_precision: float = 0.0
117
+ policy_recall: float = 0.0
118
+ baseline_precision: float = 0.0
119
+ baseline_recall: float = 0.0
120
+ residual_leak: float = 0.0
121
+ over_redaction: float = 0.0
122
+ leftover_exports: int = 0
123
+ metadata_leaks: int = 0
124
+ elapsed_ms: int = 0
125
+ engine_profile: str = ""
126
+ policy_action: str = PolicyAction.MASK.value
127
+ pages: int = 0
128
+ chars: int = 0
129
+
130
+
131
+ class CatalogRow(BaseModel):
132
+ scenario_id: str
133
+ method: str
134
+ precision: float = 0.0
135
+ recall: float = 0.0
136
+ over_redaction: float = 0.0
137
+ residual_leak: float = 0.0
138
+ leftover_exports: int = 0
139
+ metadata_leaks: int = 0
140
+ hardneg_hits: int = 0
141
+ n_docs: int = 0
142
+
143
+
144
+ class Finding(BaseModel):
145
+ finding_id: str
146
+ doc_id: str
147
+ entity_type: str
148
+ start: int
149
+ end: int
150
+ surface_preview: str
151
+ confidence: float
152
+ method: str
153
+ rule_id: str
154
+ workflow: str
155
+ reason: str
156
+ language: str = "en"
157
+ observed: str = ""
158
+ inference: str = ""
159
+ action: str = ""
160
+
161
+
162
+ class RunReport(BaseModel):
163
+ run_id: str
164
+ product: str
165
+ schema_id: str
166
+ ruleset_id: str
167
+ code_version: str
168
+ scenario_id: str
169
+ verdict: Verdict
170
+ status_label: str
171
+ kpis: dict[str, Any]
172
+ notes: list[str] = Field(default_factory=list)
173
+ input_hash: str = ""
174
+ method: str = Method.POLICY.value
175
+ policy_action: str = PolicyAction.MASK.value
demo/src/ui.py ADDED
@@ -0,0 +1,315 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ from __future__ import annotations
2
+
3
+ import json
4
+ import tempfile
5
+ from pathlib import Path
6
+ from zipfile import ZipFile
7
+
8
+ import gradio as gr
9
+ import pandas as pd
10
+
11
+ from src.domain import PolicyAction
12
+ from src.generator import scenario_by_id, scenarios
13
+ from src.i18n import TEXTS, t
14
+ from src.pipeline import result_to_downloads, run_pipeline
15
+ from src.plots import compare_figure, empty_plot, kpi_html, leftover_figure, type_figure
16
+
17
+ ROOT = Path(__file__).resolve().parents[1]
18
+ SAMPLE_PREVIEW = ROOT / "examples" / "sample_preview.json"
19
+ CSS = """
20
+ .gradio-container {max-width: 1360px !important; margin: 0 auto;}
21
+ .aria-header h1 {margin-bottom: 0.2rem;}
22
+ .aria-header p {opacity: 0.85;}
23
+ .kpi-grid {display:grid; grid-template-columns: repeat(auto-fit, minmax(170px, 1fr)); gap: 12px;}
24
+ .kpi-card {background: linear-gradient(180deg, #eef2ff, #ffffff); border: 1px solid #c7d2fe; border-radius: 16px; padding: 14px 16px;}
25
+ .kpi-label {font-size: 0.8rem; color: #3730a3; margin: 0;}
26
+ .kpi-value {font-size: 1.55rem; font-weight: 700; color: #1e1b4b; margin: 4px 0;}
27
+ .kpi-sub {margin: 0; font-size: 0.78rem; color: #334155;}
28
+ .rtl-mode {direction: rtl; text-align: right;}
29
+ .ltr-id {direction: ltr; text-align: left; font-family: ui-monospace, Consolas, monospace;}
30
+ .doc-view {white-space: pre-wrap; font-family: ui-monospace, Consolas, Tahoma, sans-serif; font-size: 0.92rem; line-height: 1.5;}
31
+ .pii-span {background: #fde68a; padding: 0 2px; border-radius: 4px;}
32
+ .redact-span {background: #c7d2fe; padding: 0 2px; border-radius: 4px;}
33
+ """
34
+
35
+
36
+ def _preview_payload() -> dict:
37
+ if SAMPLE_PREVIEW.exists():
38
+ return json.loads(SAMPLE_PREVIEW.read_text(encoding="utf-8"))
39
+ return {}
40
+
41
+
42
+ def _scenario_choices() -> list[str]:
43
+ return ["S00"] + [item.scenario_id for item in scenarios()]
44
+
45
+
46
+ def _feedback_choices(lang: str) -> list[str]:
47
+ return [t(lang, "correct"), t(lang, "false_positive"), t(lang, "insufficient"), t(lang, "needs_review")]
48
+
49
+
50
+ def _scenario_meta(scenario_id: str) -> tuple[str, str]:
51
+ if scenario_id == "S00":
52
+ return (
53
+ "Interactive editor. Paste laboratory text or upload a text/CSV/text PDF.",
54
+ "ویرایشگر تعاملی. متن آزمایشگاهی بچسبانید یا فایل متنی/CSV/PDF متن‌دار بارگذاری کنید.",
55
+ )
56
+ item = scenario_by_id(scenario_id)
57
+ return item.description, item.description_fa
58
+
59
+
60
+ def apply_language(lang: str):
61
+ rtl = lang == "fa"
62
+ preview = _preview_payload()
63
+ header = (
64
+ f"<div class='aria-header {'rtl-mode' if rtl else ''}'>"
65
+ f"<h1>{t(lang, 'title')}</h1><p>{t(lang, 'subtitle')}</p><p>{t(lang, 'kind')}</p></div>"
66
+ )
67
+ return (
68
+ gr.update(value=header),
69
+ gr.update(value=t(lang, "run_sample")),
70
+ gr.update(value=t(lang, "cancel")),
71
+ gr.update(value=t(lang, "reset")),
72
+ gr.update(value=t(lang, "load_sample")),
73
+ gr.update(value=t(lang, "apply_review")),
74
+ gr.update(label=t(lang, "scenario")),
75
+ gr.update(label=t(lang, "policy")),
76
+ gr.update(value=t(lang, "sample_note")),
77
+ gr.update(value=preview.get("kpi_html") or kpi_html(preview.get("kpis") or {}, lang)),
78
+ gr.update(open=False, label=t(lang, "advanced")),
79
+ gr.update(value=t(lang, "method_body"), rtl=rtl),
80
+ gr.update(value=t(lang, "upload_hint")),
81
+ gr.update(label=t(lang, "uploads")),
82
+ gr.update(label=t(lang, "editor_text")),
83
+ gr.update(label=t(lang, "feedback"), choices=_feedback_choices(lang), value=_feedback_choices(lang)[3]),
84
+ )
85
+
86
+
87
+ def load_sample_ui(scenario_id: str):
88
+ desc_en, desc_fa = _scenario_meta(scenario_id)
89
+ sample = ""
90
+ if scenario_id == "S00":
91
+ sample = "From: Ava Labian <ava.labian01@aria-lab.test>\nPhone: 09121234567\n"
92
+ return desc_en, desc_fa, sample
93
+
94
+
95
+ def reset_defaults():
96
+ desc_en, desc_fa = _scenario_meta("S01")
97
+ return desc_en, desc_fa, "", "S01", "mask", None
98
+
99
+
100
+ def _suppressed_from_table(table) -> list[str]:
101
+ if table is None:
102
+ return []
103
+ frame = table if isinstance(table, pd.DataFrame) else pd.DataFrame(table)
104
+ if frame.empty or "finding_id" not in frame.columns:
105
+ return []
106
+ keep_col = frame["keep"] if "keep" in frame.columns else True
107
+ suppressed = []
108
+ for finding_id, keep in zip(frame["finding_id"], keep_col if "keep" in frame.columns else [True] * len(frame)):
109
+ flag = str(keep).strip().lower()
110
+ if flag in {"false", "0", "no", ""}:
111
+ suppressed.append(str(finding_id))
112
+ return suppressed
113
+
114
+
115
+ def _detail_markdown(state: dict, finding_id: str, lang: str) -> str:
116
+ result = (state or {}).get("result") or {}
117
+ findings = result.get("findings") or []
118
+ item = next((row for row in findings if row.finding_id == finding_id), None)
119
+ if item is None:
120
+ if findings:
121
+ item = findings[0]
122
+ else:
123
+ return t(lang, "empty")
124
+ return (
125
+ f"**{item.finding_id}** · `{item.doc_id}`\n\n"
126
+ f"`{item.entity_type}` · `{item.rule_id}` · {item.workflow}\n\n"
127
+ f"**{t(lang, 'observed')}.** {item.observed}\n\n"
128
+ f"**{t(lang, 'inference')}.** {item.inference}\n\n"
129
+ f"**{t(lang, 'action')}.** {item.action}"
130
+ )
131
+
132
+
133
+ def _pack_result(result: dict) -> dict:
134
+ files = result_to_downloads(result, Path(tempfile.mkdtemp(prefix="dlp-")))
135
+ zip_path = Path(files["report"]).parent / f"{result['run_id']}_bundle.zip"
136
+ with ZipFile(zip_path, "w") as zipped:
137
+ for path in files.values():
138
+ zipped.write(path, arcname=Path(path).name)
139
+ return files, str(zip_path)
140
+
141
+
142
+ def run_live(lang, scenario_id, pasted, policy, uploads, span_table, previous_state, apply_edits):
143
+ suppressed = _suppressed_from_table(span_table) if apply_edits else []
144
+ result = run_pipeline(
145
+ scenario_id=scenario_id,
146
+ pasted=pasted or "",
147
+ uploads=uploads,
148
+ lang=lang,
149
+ action=PolicyAction(policy),
150
+ suppressed=suppressed,
151
+ )
152
+ if result.get("error") and result["verdict"].value == "invalid_input":
153
+ raise gr.Error(result["error"])
154
+ reason_md = "\n".join(f"- {line}" for line in result["reasons"] if line)
155
+ status = (
156
+ f"**{t(lang, 'status')}:** {result['status_label']}\n\n"
157
+ f"{t(lang, 'live_note')}\n\n{result['narrative']}\n\n{t(lang, 'preview_note')}"
158
+ )
159
+ files, zip_path = _pack_result(result)
160
+ history = pd.DataFrame()
161
+ if previous_state:
162
+ history = pd.DataFrame(
163
+ [
164
+ {"run": "previous", **previous_state.get("kpis", {})},
165
+ {"run": "current", **result["kpis"].model_dump()},
166
+ ]
167
+ )
168
+ finding_ids = [item.finding_id for item in result["findings"]] or ["—"]
169
+ preview = _preview_payload()
170
+ return (
171
+ kpi_html(result["kpis"], lang),
172
+ status,
173
+ reason_md,
174
+ result["before_html"],
175
+ result["after_html"],
176
+ pd.DataFrame(result["span_rows"]),
177
+ pd.DataFrame(result["doc_rows"]),
178
+ pd.DataFrame(result["finding_rows"]),
179
+ pd.DataFrame(result["comparison"]),
180
+ pd.DataFrame(preview.get("catalog") or []),
181
+ compare_figure(result["scores"], t(lang, "compare_plot")),
182
+ type_figure(next(score for key, score in result["scores"].items() if str(getattr(key, "value", key)) == "checksum_policy"), t(lang, "type_plot")),
183
+ leftover_figure(result["doc_rows"], t(lang, "left_plot")),
184
+ gr.update(choices=finding_ids, value=finding_ids[0]),
185
+ _detail_markdown({"result": result}, finding_ids[0], lang),
186
+ zip_path,
187
+ files["report"],
188
+ history,
189
+ {"kpis": result["kpis"].model_dump(), "result": result},
190
+ json.dumps([item.model_dump() for item in result["validation"].issues], indent=2),
191
+ json.dumps(result.get("challenger_notes") or [], indent=2, ensure_ascii=False),
192
+ f"run `{result['run_id']}` · hash `{result['input_hash']}` · policy `{result['policy_action']}`",
193
+ result.get("redacted_text") or "",
194
+ )
195
+
196
+
197
+ def select_finding(lang, finding_id, state):
198
+ return _detail_markdown(state or {}, finding_id, lang)
199
+
200
+
201
+ def build_demo() -> gr.Blocks:
202
+ preview = _preview_payload()
203
+ first = scenarios()[0]
204
+ preview_kpis = preview.get("kpis") or {
205
+ "n_docs": 20,
206
+ "n_spans": 60,
207
+ "policy_precision": 1.0,
208
+ "policy_recall": 1.0,
209
+ "baseline_precision": 0.6,
210
+ "baseline_recall": 0.5,
211
+ "leftover_exports": 0,
212
+ "elapsed_ms": 40,
213
+ "engine_profile": "presidio-style-checksum-policy",
214
+ }
215
+
216
+ with gr.Blocks(theme=gr.themes.Soft(primary_hue="indigo", secondary_hue="blue"), css=CSS) as demo:
217
+ lang = gr.Radio(["en", "fa"], value="en", label=TEXTS["en"]["language"])
218
+ header = gr.HTML()
219
+ with gr.Row():
220
+ run_btn = gr.Button(t("en", "run_sample"), variant="primary")
221
+ cancel_btn = gr.Button(t("en", "cancel"), variant="stop")
222
+ reset_btn = gr.Button(t("en", "reset"))
223
+ load_btn = gr.Button(t("en", "load_sample"))
224
+ apply_btn = gr.Button(t("en", "apply_review"))
225
+ note = gr.Markdown(t("en", "sample_note"))
226
+ kpis = gr.HTML(preview.get("kpi_html") or kpi_html(preview_kpis, "en"))
227
+ status = gr.Markdown(t("en", "empty"))
228
+ run_md = gr.Markdown("run —")
229
+
230
+ with gr.Row():
231
+ with gr.Column(scale=1):
232
+ scenario = gr.Dropdown(choices=_scenario_choices(), value=first.scenario_id, label=t("en", "scenario"))
233
+ policy = gr.Radio(["mask", "remove"], value="mask", label=t("en", "policy"))
234
+ desc_en = gr.Textbox(value=first.description, label="Scenario (EN)", lines=4)
235
+ desc_fa = gr.Textbox(value=first.description_fa, label="سناریو", lines=3, rtl=True)
236
+ findings = gr.Dataframe(label="Span review (toggle keep)")
237
+ finding_pick = gr.Dropdown(choices=["—"], value="—", label="Finding")
238
+ with gr.Column(scale=1):
239
+ reasons = gr.Markdown()
240
+ before = gr.HTML()
241
+ with gr.Column(scale=1):
242
+ after = gr.HTML()
243
+ detail = gr.Markdown()
244
+ redacted_box = gr.Textbox(label="Redacted preview (plain)", lines=8)
245
+ feedback = gr.Radio(choices=_feedback_choices("en"), value=_feedback_choices("en")[3], label=t("en", "feedback"))
246
+ with gr.Row():
247
+ zip_out = gr.File(label="CSV / JSON / PDF bundle")
248
+ json_out = gr.File(label="JSON report")
249
+
250
+ with gr.Tabs():
251
+ with gr.Tab("Data"):
252
+ hint = gr.Markdown(t("en", "upload_hint"))
253
+ with gr.Accordion(t("en", "advanced"), open=False) as advanced:
254
+ uploads = gr.File(
255
+ label=t("en", "uploads"),
256
+ file_count="multiple",
257
+ file_types=[".txt", ".csv", ".pdf", ".json", ".jsonl"],
258
+ )
259
+ pasted = gr.Textbox(value="", label=t("en", "editor_text"), lines=8)
260
+ with gr.Tab("Findings"):
261
+ docs = gr.Dataframe(label="Documents")
262
+ finding_table = gr.Dataframe(label="Findings")
263
+ type_plot = gr.Plot(value=empty_plot("—"))
264
+ with gr.Tab("Evidence"):
265
+ leftover_plot = gr.Plot(value=empty_plot("—"))
266
+ with gr.Tab("Compare"):
267
+ compare_plot = gr.Plot(value=empty_plot("—"))
268
+ comparison = gr.Dataframe(label="Regex-only vs checksum policy")
269
+ catalog = gr.Dataframe(label="Recorded catalog")
270
+ history = gr.Dataframe(label="Previous vs current KPIs")
271
+ with gr.Tab("Method and limits"):
272
+ method_md = gr.Markdown(t("en", "method_body"))
273
+ issues = gr.Code(language="json", label="Validation issues")
274
+ challenger = gr.Code(language="json", label="Review-list challenger (no paid API)")
275
+ gr.Markdown(
276
+ "Organization: [AriaAICompany](https://huggingface.co/AriaAICompany) · "
277
+ "Dataset: [dlp-lab](https://huggingface.co/datasets/AriaAICompany/dlp-lab) · "
278
+ "Site: [aria-ai.ir](https://aria-ai.ir/) · "
279
+ "Not secret scanning (project 8), not tenant ACL (project 23)."
280
+ )
281
+
282
+ previous = gr.State(None)
283
+ outputs_lang = [
284
+ header, run_btn, cancel_btn, reset_btn, load_btn, apply_btn, scenario, policy,
285
+ note, kpis, advanced, method_md, hint, uploads, pasted, feedback,
286
+ ]
287
+ lang.change(apply_language, inputs=[lang], outputs=outputs_lang)
288
+ demo.load(apply_language, inputs=[lang], outputs=outputs_lang)
289
+ scenario.change(load_sample_ui, inputs=[scenario], outputs=[desc_en, desc_fa, pasted])
290
+ load_btn.click(load_sample_ui, inputs=[scenario], outputs=[desc_en, desc_fa, pasted])
291
+ reset_btn.click(reset_defaults, outputs=[desc_en, desc_fa, pasted, scenario, policy, uploads])
292
+
293
+ outputs_run = [
294
+ kpis, status, reasons, before, after, findings, docs, finding_table, comparison, catalog,
295
+ compare_plot, type_plot, leftover_plot, finding_pick, detail, zip_out, json_out,
296
+ history, previous, issues, challenger, run_md, redacted_box,
297
+ ]
298
+ click = run_btn.click(
299
+ lambda lang, scenario, pasted, policy, uploads, findings, previous: run_live(
300
+ lang, scenario, pasted, policy, uploads, findings, previous, False
301
+ ),
302
+ inputs=[lang, scenario, pasted, policy, uploads, findings, previous],
303
+ outputs=outputs_run,
304
+ )
305
+ apply_click = apply_btn.click(
306
+ lambda lang, scenario, pasted, policy, uploads, findings, previous: run_live(
307
+ lang, scenario, pasted, policy, uploads, findings, previous, True
308
+ ),
309
+ inputs=[lang, scenario, pasted, policy, uploads, findings, previous],
310
+ outputs=outputs_run,
311
+ )
312
+ cancel_btn.click(None, cancels=[click, apply_click])
313
+ finding_pick.change(select_finding, inputs=[lang, finding_pick, previous], outputs=[detail])
314
+
315
+ return demo
demo/src/validation.py ADDED
@@ -0,0 +1,42 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ from __future__ import annotations
2
+
3
+ from src.domain import MAX_BYTES, MAX_CHARS, MAX_DOCS, MAX_FILES, MAX_PAGES
4
+ from src.schemas import LabDocument, ValidationReport
5
+
6
+
7
+ def validate_text(name: str, text: str, report: ValidationReport) -> None:
8
+ if text is None:
9
+ report.add("empty", "Text is missing.", name)
10
+ return
11
+ if len(text) > MAX_CHARS:
12
+ report.add("too_large_text", f"Text exceeds {MAX_CHARS} characters.", name)
13
+ if "\x00" in text:
14
+ report.add("binary", "Binary content is out of scope.", name)
15
+ lowered = text.lower()
16
+ if "begin private key" in lowered or "aws_secret_access_key" in lowered:
17
+ report.add("credential_rejected", "Live credentials are not accepted in the public demo.", name)
18
+
19
+
20
+ def validate_documents(documents: list[LabDocument]) -> ValidationReport:
21
+ report = ValidationReport(ok=True)
22
+ if not documents:
23
+ report.add("empty", "No laboratory document is selected.", "documents")
24
+ return report
25
+ if len(documents) > MAX_DOCS:
26
+ report.add("too_many_docs", f"At most {MAX_DOCS} documents are scored.", "documents")
27
+ total = 0
28
+ for item in documents:
29
+ validate_text(item.doc_id, item.body, report)
30
+ total += len(item.body.encode("utf-8", errors="replace"))
31
+ if total > MAX_BYTES:
32
+ report.add("too_large", f"Corpus exceeds {MAX_BYTES} bytes.", "documents")
33
+ return report
34
+
35
+
36
+ def validate_limits(n_files: int, pages: int, chars: int, report: ValidationReport) -> None:
37
+ if n_files > MAX_FILES:
38
+ report.add("too_many_files", f"At most {MAX_FILES} uploads.", "uploads")
39
+ if pages > MAX_PAGES:
40
+ report.add("too_many_pages", f"PDFs are limited to {MAX_PAGES} pages.", "pdf")
41
+ if chars > MAX_CHARS:
42
+ report.add("too_large_text", f"Combined text exceeds {MAX_CHARS} characters.", "text")
demo/tests/conftest.py ADDED
@@ -0,0 +1,8 @@
 
 
 
 
 
 
 
 
 
1
+ from __future__ import annotations
2
+
3
+ import sys
4
+ from pathlib import Path
5
+
6
+ ROOT = Path(__file__).resolve().parents[1]
7
+ if str(ROOT) not in sys.path:
8
+ sys.path.insert(0, str(ROOT))
demo/tests/test_pipeline.py ADDED
@@ -0,0 +1,159 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ from __future__ import annotations
2
+
3
+ from src.analyzer import analyze_policy
4
+ from src.baseline import analyze_regex_only
5
+ from src.checksums import iban_ok, iran_national_id_ok, luhn_ok, make_iban, make_luhn, make_national_id
6
+ from src.domain import (
7
+ DEFAULT_SEED,
8
+ DEV_TEMPLATES,
9
+ MAX_BYTES,
10
+ MAX_CHARS,
11
+ TEST_TEMPLATES,
12
+ EntityType,
13
+ Method,
14
+ PolicyAction,
15
+ Workflow,
16
+ )
17
+ from src.generator import generate_documents, scenarios
18
+ from src.normalize import fold_digits
19
+ from src.pdfio import build_pdf, inspect_pdf, redact_pdf
20
+ from src.pipeline import run_pipeline
21
+ from src.redaction import leftover_values
22
+ from src.schemas import ValidationReport
23
+ from src.validation import validate_text
24
+
25
+
26
+ def test_catalog_sizes():
27
+ docs = generate_documents()
28
+ assert len(docs) == 80
29
+ assert len(scenarios()) == 5
30
+ assert {item.scenario_id for item in scenarios()} == {"S01", "S02", "S03", "S04", "S05"}
31
+
32
+
33
+ def test_split_is_by_template():
34
+ development = {item.template for item in generate_documents() if item.split.value == "development"}
35
+ test = {item.template for item in generate_documents() if item.split.value == "test"}
36
+ assert development == DEV_TEMPLATES
37
+ assert test == TEST_TEMPLATES
38
+ assert development.isdisjoint(test)
39
+
40
+
41
+ def test_checksums_roundtrip():
42
+ class R:
43
+ def __init__(self) -> None:
44
+ self.n = 7
45
+
46
+ def randint(self, a, b):
47
+ self.n = (self.n * 1103515245 + 12345) % 2**31
48
+ return a + self.n % (b - a + 1)
49
+
50
+ rng = R()
51
+ nid = make_national_id(rng)
52
+ card = make_luhn(rng)
53
+ iban = make_iban(rng)
54
+ assert iran_national_id_ok(nid)
55
+ assert luhn_ok(card)
56
+ assert iban_ok(iban)
57
+
58
+
59
+ def test_policy_reads_persian_digits_baseline_does_not():
60
+ result = run_pipeline("S02", lang="en", seed=DEFAULT_SEED)
61
+ policy = result["scores"][Method.POLICY]
62
+ baseline = result["scores"][Method.BASELINE]
63
+ assert policy.recall >= 0.95
64
+ assert baseline.recall < policy.recall
65
+ fa = next(item for item in result["documents"] if item.template == "fa_hr_letter")
66
+ pred = analyze_policy(fa.body)
67
+ types = {item.entity_type for item in pred}
68
+ assert EntityType.PHONE in types
69
+ assert EntityType.IR_NATIONAL_ID in types
70
+ base = analyze_regex_only(fa.body)
71
+ assert not any(item.entity_type is EntityType.PHONE for item in base)
72
+
73
+
74
+ def test_hard_negatives_are_not_policy_redacted():
75
+ result = run_pipeline("S03", lang="en", seed=DEFAULT_SEED)
76
+ policy = result["scores"][Method.POLICY]
77
+ baseline = result["scores"][Method.BASELINE]
78
+ assert policy.hardneg_hits == 0
79
+ assert policy.precision == 1.0
80
+ assert policy.leftover_exports == 0
81
+ assert baseline.hardneg_hits >= 1 or baseline.over_redaction > 0
82
+
83
+
84
+ def test_invalid_national_id_stays_on_review_list():
85
+ result = run_pipeline("S04", lang="en", seed=DEFAULT_SEED)
86
+ unknown = [item for item in result["findings"] if item.entity_type == EntityType.UNKNOWN.value]
87
+ assert unknown
88
+ assert all(item.workflow == Workflow.REVIEW.value for item in unknown)
89
+ exported = fold_digits(" ".join(payload["text"] for payload in result["exports"].values()))
90
+ for document in result["documents"]:
91
+ for item in document.hard_negatives:
92
+ digits = "".join(ch for ch in fold_digits(item) if ch.isdigit())
93
+ if len(digits) == 10:
94
+ assert digits in exported
95
+
96
+
97
+ def test_pdf_redaction_removes_text_and_metadata():
98
+ docs = generate_documents()
99
+ memo = next(item for item in docs if item.template == "en_pdf_memo")
100
+ values = [span.surface for span in memo.gold_spans]
101
+ raw = build_pdf(memo.body, memo.metadata)
102
+ before = inspect_pdf(raw)
103
+ assert memo.gold_meta[0].surface in before["metadata"].get("author", "")
104
+ redacted, info = redact_pdf(raw, values)
105
+ after = inspect_pdf(redacted)
106
+ assert leftover_values(after["text"], values) == []
107
+ assert leftover_values(" ".join(after["metadata"].values()), [item.surface for item in memo.gold_meta]) == []
108
+ assert info["leftover"] == []
109
+
110
+
111
+ def test_full_catalog_policy_beats_baseline_precision():
112
+ result = run_pipeline("S05", lang="en", seed=DEFAULT_SEED)
113
+ policy = result["scores"][Method.POLICY]
114
+ baseline = result["scores"][Method.BASELINE]
115
+ assert policy.precision >= 0.9
116
+ assert policy.recall >= 0.9
117
+ assert policy.precision > baseline.precision
118
+ assert policy.leftover_exports == 0
119
+ assert policy.metadata_leaks == 0
120
+
121
+
122
+ def test_mask_hides_gold_remove_deletes_it():
123
+ masked = run_pipeline("S01", lang="en", action=PolicyAction.MASK)
124
+ removed = run_pipeline("S01", lang="en", action=PolicyAction.REMOVE)
125
+ gold = masked["documents"][0].gold_spans[0].surface
126
+ assert gold not in masked["exports"][masked["documents"][0].doc_id]["text"]
127
+ assert gold not in removed["exports"][removed["documents"][0].doc_id]["text"]
128
+ assert "[" in masked["redacted_text"]
129
+ assert gold not in removed["redacted_text"]
130
+
131
+
132
+ def test_empty_interactive_is_invalid():
133
+ result = run_pipeline("S00", "", None, "en")
134
+ assert result["verdict"].value == "invalid_input"
135
+ assert result["error"]
136
+
137
+
138
+ def test_oversized_text_is_rejected():
139
+ report = ValidationReport(ok=True)
140
+ validate_text("user_text", "x" * (MAX_CHARS + 1), report)
141
+ assert not report.ok
142
+ assert any(item.code == "too_large_text" for item in report.issues)
143
+
144
+
145
+ def test_credential_marker_is_rejected():
146
+ report = ValidationReport(ok=True)
147
+ validate_text("user_text", "-----BEGIN PRIVATE KEY----- lab", report)
148
+ assert not report.ok
149
+ assert any(item.code == "credential_rejected" for item in report.issues)
150
+
151
+
152
+ def test_caps():
153
+ assert MAX_BYTES == 10 * 1024 * 1024
154
+ assert MAX_CHARS == 80_000
155
+
156
+
157
+ def test_mr_protocol_is_not_a_person():
158
+ spans = analyze_policy("Restart Mr. Protocol handler v2.4.1")
159
+ assert not any(item.entity_type is EntityType.PERSON for item in spans)
showcase/README.md ADDED
@@ -0,0 +1,18 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ title: AriaDLP Redaction Lab
3
+ emoji: 🔏
4
+ colorFrom: blue
5
+ colorTo: indigo
6
+ sdk: static
7
+ short_description: Find and redact lab PII in text, CSV, and PDF
8
+ pinned: true
9
+ license: apache-2.0
10
+ tags:
11
+ - cybersecurity
12
+ - pii-detection
13
+ - redaction
14
+ - document-analysis
15
+ - text-analysis
16
+ ---
17
+
18
+ Organization project card for AriaDLP. The live Gradio workbench is hosted on the personal PRO account when `cpu-basic` quota allows.
showcase/index.html ADDED
@@ -0,0 +1,64 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ <!DOCTYPE html>
2
+ <html lang="en">
3
+ <head>
4
+ <meta charset="utf-8" />
5
+ <meta name="viewport" content="width=device-width, initial-scale=1" />
6
+ <title>AriaDLP — Aria AI</title>
7
+ <style>
8
+ :root { --indigo:#4338ca; --blue:#1d4ed8; --ink:#0f172a; --muted:#475569; --bg:#f8fafc; }
9
+ * { box-sizing: border-box; }
10
+ body { margin:0; font-family:"Segoe UI","Vazirmatn",Tahoma,sans-serif; color:var(--ink); background:radial-gradient(circle at top left,#e0e7ff,var(--bg) 42%); }
11
+ header, main, footer { width:min(1120px, calc(100% - 32px)); margin: 0 auto; }
12
+ header { padding:36px 0 12px; }
13
+ .eyebrow { color:var(--indigo); font-weight:700; letter-spacing:.04em; text-transform:uppercase; font-size:.78rem; }
14
+ h1 { font-size:clamp(1.8rem,3vw,2.4rem); margin:8px 0; }
15
+ .lead { font-size:1.1rem; color:var(--muted); max-width:46rem; }
16
+ .actions { display:flex; flex-wrap:wrap; gap:10px; margin:20px 0 8px; }
17
+ a.btn { display:inline-block; text-decoration:none; padding:10px 16px; border-radius:999px; font-weight:600; }
18
+ a.primary { background:var(--indigo); color:#fff; }
19
+ a.ghost { background:#fff; color:var(--indigo); border:1px solid #c7d2fe; }
20
+ .grid { display:grid; grid-template-columns:repeat(auto-fit,minmax(220px,1fr)); gap:12px; margin:24px 0; }
21
+ .card { background:#fff; border:1px solid #e2e8f0; border-radius:18px; padding:16px 18px; }
22
+ .card strong { display:block; font-size:1.4rem; color:var(--indigo); }
23
+ .note { background:#eef2ff; border:1px solid #c7d2fe; border-radius:16px; padding:14px 16px; color:#312e81; }
24
+ iframe { width:100%; height:86vh; min-height:720px; border:1px solid #c7d2fe; border-radius:20px; background:#fff; margin:16px 0 28px; }
25
+ footer { padding:12px 0 40px; color:var(--muted); font-size:.92rem; }
26
+ .fa { direction:rtl; text-align:right; }
27
+ </style>
28
+ </head>
29
+ <body>
30
+ <header>
31
+ <p class="eyebrow">Aria AI · Project 24 / 25</p>
32
+ <h1>AriaDLP — Persian / English Redaction Lab</h1>
33
+ <p class="lead">
34
+ Find laboratory personal data in text, CSV, and text PDFs.
35
+ Preview mask or remove before download. Detection is scoped.
36
+ A clean suite is not a claim that every sensitive value was removed.
37
+ </p>
38
+ <div class="actions">
39
+ <a class="btn primary" href="https://huggingface.co/spaces/alirezaaminzadeh/aria-dlp" target="_blank" rel="noreferrer">Open live demo</a>
40
+ <a class="btn ghost" href="https://huggingface.co/datasets/AriaAICompany/dlp-lab" target="_blank" rel="noreferrer">Dataset</a>
41
+ <a class="btn ghost" href="https://huggingface.co/AriaAICompany" target="_blank" rel="noreferrer">Organization</a>
42
+ <a class="btn ghost" href="https://huggingface.co/collections/AriaAICompany/aria-ai-cybersecurity-6ab01faa3995a3bb737b58a7" target="_blank" rel="noreferrer">Collection</a>
43
+ </div>
44
+ </header>
45
+ <main>
46
+ <section class="grid">
47
+ <article class="card"><strong>80 docs</strong>Template-disjoint development / test split</article>
48
+ <article class="card"><strong>2 methods</strong>Regex-only vs checksum policy</article>
49
+ <article class="card"><strong>3 formats</strong>Text · CSV · text PDF</article>
50
+ <article class="card"><strong>CPU</strong>No live LLM · scoped redaction</article>
51
+ </section>
52
+ <p class="note">
53
+ Organization Gradio hosting needs Team/Enterprise. This page is the AriaAICompany project card.
54
+ The iframe is the personal PRO workbench when that Space is running.
55
+ </p>
56
+ <section class="fa card">
57
+ <h2>آریادی‌ال‌پی</h2>
58
+ <p>میز کار پوشاندن داده آزمایشگاهی دوزبانه. مبنا فقط regex است. سیاست checksum و رقم فارسی دارد. پوشش فارسی آماده Presidio فرض نشده. سوئیت تمیز ادعای حذف کامل نیست. PDF اسکن‌شده خارج از محدوده است.</p>
59
+ </section>
60
+ <iframe title="AriaDLP live demo" src="https://alirezaaminzadeh-aria-dlp.hf.space" loading="lazy"></iframe>
61
+ </main>
62
+ <footer>Aria AI · Cybersecurity · Roadmap project 24 · Apache-2.0 / CC-BY-4.0</footer>
63
+ </body>
64
+ </html>