Hugging Face's logo

Spaces:
solidprivacy
/
scrub
Running

App Files Community
scrub
File size: 2,903 Bytes
7a24de4
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
# WP37 — DOCX hidden content extraction helper

Status: completed helper/tests/documentation-only  
Repository: `solidprivacy-nl/scrub`

WP37 adds a pure local, read-only helper for detecting and extracting text from high-risk DOCX package parts that are outside the current DOCX reinsert foundation scope.

The helper is:

```text
docx_hidden_content_extractor.py
```

Primary function:

```text
inspect_docx_hidden_content(content: bytes) -> dict
```

---

## 1. Purpose

The helper makes hidden-content risk audit-visible before any cleaner, removal policy or export-blocking policy is implemented.

It focuses on:

- headers;
- footers;
- comments / kantlijncommentaren;
- comment author/person-like metadata where present;
- tracked-change markers such as insertions, deletions and move markers.

This directly follows WP35 and WP36A.

---

## 2. What the helper reports

The helper returns an audit-oriented dictionary with fields such as:

```text
valid_docx
validation_issues
local_only
ai_processing
cloud_processing
extraction_only
cleaning_applied
export_blocking
docx_parts_seen
headers
footers
comments
tracked_changes
detected
warnings
```

Detection flags include:

```text
headers_detected
footers_detected
comments_detected
tracked_changes_detected
```

---

## 3. Supported extraction scope

The helper inspects these DOCX package patterns:

```text
word/header*.xml
word/footer*.xml
word/comments.xml
word/commentsExtended.xml
word/person.xml
word/*.xml for tracked-change markers
```

It extracts text from WordprocessingML text nodes and reports parse errors per part where applicable.

Tracked-change markers include:

```text
w:ins
w:del
w:delText
w:moveFrom
w:moveTo
w:moveFromRangeStart
w:moveFromRangeEnd
w:moveToRangeStart
w:moveToRangeEnd
```

---

## 4. Explicit non-goals

WP37 does not:

- clean DOCX files;
- remove comments;
- remove or accept tracked changes;
- remove metadata;
- block export;
- change export semantics;
- change DOCX reinsert behavior;
- change Scrub Key schema;
- change Streamlit UI;
- add dependencies;
- add cloud processing;
- add real-data fixtures.

The helper is deliberately extraction/audit-only.

---

## 5. Privacy boundary

The helper is local and side-effect free:

- no file-system persistence;
- no network calls;
- no AI calls;
- no cloud processing;
- no mutation of source bytes;
- no export package writing;
- no real-data test fixtures.

It accepts bytes and returns an in-memory audit dictionary.

---

## 6. Recommended next step

Recommended next package:

```text
WP38 — DOCX hygiene audit report
```

Purpose:

- Convert the extraction helper output into a user/support-facing hygiene audit structure.
- Keep it warning/report-only unless a later approved package defines export-blocking semantics.

Do not start comment removal, tracked-change removal or export blocking before the audit policy is explicit and tested.