build NFCorpus from mteb/nfcorpus@ec0fa4fe · rteb_open v0.1.0

README.md

@@ -1,114 +1,96 @@
 ---
-license: apache-2.0
+pretty_name: BEIR · NFCorpus
+license: cc-by-sa-4.0
+language:
+- en
+multilinguality:
+- monolingual
 task_categories:
-- information-retrieval
 - text-retrieval
+task_ids:
+- document-retrieval
 tags:
 - beir
-- nfcorpus
-- information-retrieval
+- mteb
 - retrieval
-- search
+- biomedical
 configs:
 - config_name: corpus
   data_files:
+  - split: test
+    path: corpus/test-*.parquet
+- config_name: qrels
+  data_files:
+  - split: test
+    path: qrels/test-*.parquet
   - split: train
-    path: corpus/train-*
+    path: qrels/train-*.parquet
+  - split: dev
+    path: qrels/dev-*.parquet
 - config_name: queries
   data_files:
+  - split: test
+    path: queries/test-*.parquet
   - split: train
-    path: queries/train-*
-dataset_info:
-- config_name: corpus
-  features:
-  - name: title
-    dtype: string
-  - name: text
-    dtype: string
-  - name: _id
-    dtype: string
-  splits:
-  - name: train
-    num_bytes: 5856698
-    num_examples: 3633
-  download_size: 3178637
-  dataset_size: 5856698
-- config_name: queries
-  features:
-  - name: title
-    dtype: string
-  - name: text
-    dtype: string
-  - name: _id
-    dtype: string
-  splits:
-  - name: train
-    num_bytes: 141303
-    num_examples: 3237
-  download_size: 80445
-  dataset_size: 141303
+    path: queries/train-*.parquet
+  - split: dev
+    path: queries/dev-*.parquet
 ---
-
-# BEIR NFCORPUS Dataset (Migrated)
-
-This is a migrated version of BeIR/nfcorpus that is compatible with datasets library 4.0.0+.
-
-## Dataset Description
-
-This dataset contains the nfcorpus dataset from the BEIR benchmark, converted from the old script-based format to Parquet format.
-
-## Dataset Structure
-
-### Queries
-
-- **Split 'queries'**: 3,237 examples
-  - Features: ['_id', 'title', 'text']
-
-- **Total examples**: 3,237
-
-
-### Corpus
-
-- **Split 'corpus'**: 3,633 examples
-  - Features: ['_id', 'title', 'text']
-
-- **Total examples**: 3,633
-
-
-## Usage
+# NFCorpus — BEIR, unified schema
+
+A normalised copy of the dataset behind the `mteb` task **`NFCorpus`**, one of the tasks of the [BEIR](https://github.com/beir-cellar/beir) benchmark as `mteb` defines it. Same queries, documents
+and relevance judgements as the benchmark evaluates — reshaped into one strict schema shared by every dataset
+in this collection.
+
+|  |  |
+|---|---|
+| Source | [`mteb/nfcorpus`](https://huggingface.co/datasets/mteb/nfcorpus/tree/ec0fa4fe99da2ff19ca1214b7966684033a58814) @ `ec0fa4fe99da` (the revision pinned in `mteb`) |
+| Domain · languages | biomedical · eng |
+| Queries / documents / qrels (all splits) | 3,237 / 3,633 / 134,294 |
+| Qrels per query | min 1 · mean 38.186 · max 475 |
+| Score values | 1 ×11,758, 2 ×576 |
+| Layout | `queries` · `corpus` · `qrels`, split `test`; `queries`/`qrels` also carry `train`, `dev` — one shared corpus |
+| Splits | `qrels`: train, dev, test · `queries`: train, dev, test |
+| License | `cc-by-sa-4.0` |
+
+## Schema
+
+| config | columns | rules |
+|---|---|---|
+| `queries` | `id: string`, `text: string` | ids unique and non-empty; every query has ≥ 1 qrel |
+| `corpus` | `id: string`, `title: string`, `text: string` | `title` is always present (`""` when the source has none) |
+| `qrels` | `query-id: string`, `corpus-id: string`, `score: int32` | referential integrity to both tables; no duplicate pairs; no floats |
+
+Files are Parquet, sorted by id, zstd-compressed, sharded at 500 MB. Every rule above is enforced by a validator
+before publishing; `provenance.json` records the source file hashes, what changed, and the output file hashes.
+
+## What changed from the source
+
+- `dev`: dropped 2,913 queries that have no qrels in this split (mteb does the same at load time)
+- `test`: dropped 2,914 queries that have no qrels in this split (mteb does the same at load time)
+- `train`: dropped 647 queries that have no qrels in this split (mteb does the same at load time)
+- cast `qrels.score` double -> int32
+- kept 40 documents whose text is identical to another document's — a property of the source; note that models embed such documents identically and retrieval ties among them are broken by corpus order
+- renamed `_id` → `id`
+- renamed source splits (`queries`←`queries/queries`, `corpus`←`corpus/corpus`) to `test`
+
+## Load it
 
 ```python
 from datasets import load_dataset
-
-# Load queries (split: queries)
-queries = load_dataset("Hyukkyu/beir-nfcorpus", "queries", split="queries")
-
-# Load corpus (split: corpus)
-corpus = load_dataset("Hyukkyu/beir-nfcorpus", "corpus", split="corpus")
+queries = load_dataset("Hyukkyu/beir-nfcorpus", "queries", split="test")
+corpus  = load_dataset("Hyukkyu/beir-nfcorpus", "corpus", split="test")
+qrels   = load_dataset("Hyukkyu/beir-nfcorpus", "qrels", split="test")
 ```
 
+## Cross-check against the previous layout
 
-## Available Splits
-
-### Queries
-- `queries`: 3,237 examples
-
-### Corpus
-- `corpus`: 3,633 examples
+This repo previously held the same dataset in BEIR's raw two-repo form (`Hyukkyu/beir-nfcorpus` + `Hyukkyu/beir-nfcorpus-qrels`; that layout is kept under the git tag `legacy-layout`). Before replacing it, the build was compared against it: corpus id set **identical** (3,633 documents); qrels rows per split `test` 12,334 = 12,334, `train` 110,575 = 110,575, `dev` 11,385 = 11,385.
 
-## Original Dataset
+## License and attribution
 
-This dataset is migrated from: BeIR/nfcorpus
+The data is redistributed under the source's terms — `cc-by-sa-4.0`. All credit belongs to the
+original authors; see the source repository and the references in `mteb`'s task metadata (https://www.cl.uni-heidelberg.de/statnlpgroup/nfcorpus/).
+This repository is an independent repackaging and is not affiliated with the RTEB or MTEB maintainers.
 
-## Citation
-
-If you use this dataset, please cite the original BEIR paper:
-
-```bibtex
-@article{thakur2021beir,
-  title={BEIR: A Heterogeneous Benchmark for Zero-shot Evaluation of Information Retrieval Models},
-  author={Thakur, Nandan and Reimers, Nils and Ruckle, Andreas and Srivastava, Abhishek and Gurevych, Iryna},
-  journal={arXiv preprint arXiv:2104.08663},
-  year={2021}
-}
-```
+> **License discrepancy.** The upstream dataset card declares `cc-by-sa-4.0` while `mteb`'s `TaskMetadata` declares `not specified`. This card carries the upstream value as the more conservative choice.

corpus/test-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:456c749b15ca4c6baaba5d5e3499f62cc87db34707db786146fc092fbb465830
+size 2159038

provenance.json

@@ -0,0 +1,322 @@
+{
+ "task": "NFCorpus",
+ "benchmark": "beir",
+ "repo": "beir-nfcorpus",
+ "rteb_open_version": "0.1.0",
+ "source": {
+  "path": "mteb/nfcorpus",
+  "revision": "ec0fa4fe99da2ff19ca1214b7966684033a58814",
+  "url": "https://huggingface.co/datasets/mteb/nfcorpus/tree/ec0fa4fe99da2ff19ca1214b7966684033a58814"
+ },
+ "license": {
+  "card": "cc-by-sa-4.0",
+  "mteb_task_metadata": "not specified",
+  "upstream_checked": true
+ },
+ "languages": [
+  "eng"
+ ],
+ "domain": "biomedical",
+ "byte_preserving_text": false,
+ "layout": {
+  "subsets": null,
+  "splits": [
+   "test",
+   "train",
+   "dev"
+  ],
+  "split_map": {
+   "test": "test",
+   "train": "train",
+   "dev": "dev"
+  }
+ },
+ "parts": [
+  {
+   "subset": null,
+   "split": "test",
+   "source_configs": {
+    "queries": "queries/queries",
+    "corpus": "corpus/corpus",
+    "qrels": "default/test"
+   },
+   "normalize": {
+    "dropped_columns": {},
+    "orphan_queries_dropped": 2914,
+    "duplicate_qrels_collapsed": 0,
+    "nul_bytes_removed": 0,
+    "score_cast": "double -> int32",
+    "id_renamed": true,
+    "title_added": false,
+    "null_titles_filled": 0
+   },
+   "stats": {
+    "n_corpus": 3633,
+    "empty_docs": 0,
+    "corpus_with_title": 3633,
+    "avg_doc_chars": 1589.8,
+    "n_queries": 323,
+    "n_qrels": 12334,
+    "qrels_per_query": {
+     "min": 1,
+     "mean": 38.186,
+     "max": 475
+    },
+    "score_values": {
+     "1": 11758,
+     "2": 576
+    },
+    "dangling_qrels": 0,
+    "avg_query_chars": 21.8,
+    "duplicate_text_docs": 40
+   }
+  },
+  {
+   "subset": null,
+   "split": "train",
+   "source_configs": {
+    "queries": "queries/queries",
+    "corpus": "corpus/corpus",
+    "qrels": "default/train"
+   },
+   "normalize": {
+    "dropped_columns": {},
+    "orphan_queries_dropped": 647,
+    "duplicate_qrels_collapsed": 0,
+    "nul_bytes_removed": 0,
+    "score_cast": "double -> int32",
+    "id_renamed": true,
+    "title_added": false,
+    "null_titles_filled": 0
+   },
+   "stats": {
+    "n_corpus": 3633,
+    "empty_docs": 0,
+    "corpus_with_title": 3633,
+    "avg_doc_chars": 1589.8,
+    "n_queries": 2590,
+    "n_qrels": 110575,
+    "qrels_per_query": {
+     "min": 1,
+     "mean": 42.693,
+     "max": 1363
+    },
+    "score_values": {
+     "1": 110575
+    },
+    "dangling_qrels": 0,
+    "avg_query_chars": 21.9,
+    "duplicate_text_docs": 40
+   }
+  },
+  {
+   "subset": null,
+   "split": "dev",
+   "source_configs": {
+    "queries": "queries/queries",
+    "corpus": "corpus/corpus",
+    "qrels": "default/dev"
+   },
+   "normalize": {
+    "dropped_columns": {},
+    "orphan_queries_dropped": 2913,
+    "duplicate_qrels_collapsed": 0,
+    "nul_bytes_removed": 0,
+    "score_cast": "double -> int32",
+    "id_renamed": true,
+    "title_added": false,
+    "null_titles_filled": 0
+   },
+   "stats": {
+    "n_corpus": 3633,
+    "empty_docs": 0,
+    "corpus_with_title": 3633,
+    "avg_doc_chars": 1589.8,
+    "n_queries": 324,
+    "n_qrels": 11385,
+    "qrels_per_query": {
+     "min": 1,
+     "mean": 35.139,
+     "max": 919
+    },
+    "score_values": {
+     "1": 10864,
+     "2": 521
+    },
+    "dangling_qrels": 0,
+    "avg_query_chars": 22.3,
+    "duplicate_text_docs": 40
+   }
+  }
+ ],
+ "files": {
+  "corpus": {
+   "test": [
+    {
+     "path": "corpus/test-00000-of-00001.parquet",
+     "rows": 3633,
+     "bytes": 2159038,
+     "sha256": "456c749b15ca4c6baaba5d5e3499f62cc87db34707db786146fc092fbb465830"
+    }
+   ]
+  },
+  "queries": {
+   "test": [
+    {
+     "path": "queries/test-00000-of-00001.parquet",
+     "rows": 323,
+     "bytes": 6738,
+     "sha256": "138f4b2dd44362311fad6e2aa7911d1be64b91ee07457a3946241d9837575bd6"
+    }
+   ],
+   "train": [
+    {
+     "path": "queries/train-00000-of-00001.parquet",
+     "rows": 2590,
+     "bytes": 43656,
+     "sha256": "26f3560df603d38560ac850c4ca0e42b310ee7865fe23f934673ce2f7d2a4243"
+    }
+   ],
+   "dev": [
+    {
+     "path": "queries/dev-00000-of-00001.parquet",
+     "rows": 324,
+     "bytes": 6856,
+     "sha256": "5c045749115d6d3ffe3358f6588446deef58043b5bc3cab1df8e924d045c0398"
+    }
+   ]
+  },
+  "qrels": {
+   "test": [
+    {
+     "path": "qrels/test-00000-of-00001.parquet",
+     "rows": 12334,
+     "bytes": 21926,
+     "sha256": "67c3f007e34b7d1b211d8276f1c5292751d4dcbba7791f4b315a37b3b937e7af"
+    }
+   ],
+   "train": [
+    {
+     "path": "qrels/train-00000-of-00001.parquet",
+     "rows": 110575,
+     "bytes": 128581,
+     "sha256": "62a747f001310bd406d654f940e186a7c3e2a123d5014b66cc1b81949b0662b6"
+    }
+   ],
+   "dev": [
+    {
+     "path": "qrels/dev-00000-of-00001.parquet",
+     "rows": 11385,
+     "bytes": 21718,
+     "sha256": "f311f8a7a224b0e595c3c41c07a846485e2292d9dceb8b38f11d74c8214df2f4"
+    }
+   ]
+  }
+ },
+ "stats_total": {
+  "n_queries": 3237,
+  "n_corpus": 3633,
+  "n_qrels": 134294
+ },
+ "source_files": [
+  {
+   "path": ".gitattributes",
+   "size": 2406,
+   "sha256": null,
+   "blob_id": "80f8495df3315d75669d8adee26b2c0f409ed338"
+  },
+  {
+   "path": "README.md",
+   "size": 1459,
+   "sha256": null,
+   "blob_id": "de8ed014a0b21956abe2d7b673f7a558f8cbac97"
+  },
+  {
+   "path": "corpus.jsonl",
+   "size": 5969652,
+   "sha256": "ee3db77bc2183e66b9c114e6eca4adf44de381705b1f8c94de8efdd232f0c4ff",
+   "blob_id": "e182728022727291ff1743f59b372ac1c6efdd55"
+  },
+  {
+   "path": "qrels/dev.jsonl",
+   "size": 747747,
+   "sha256": null,
+   "blob_id": "c85fc9ed072525b801d0acdff49c5be7b3c2c17c"
+  },
+  {
+   "path": "qrels/dev.tsv",
+   "size": 258218,
+   "sha256": null,
+   "blob_id": "4c8b9ce3ecc4f26ea4a1d9a6c9e48a9adcc0458c"
+  },
+  {
+   "path": "qrels/test.jsonl",
+   "size": 809908,
+   "sha256": null,
+   "blob_id": "4a3ee3075ada5271b2c09c58b28f54341f20072c"
+  },
+  {
+   "path": "qrels/test.tsv",
+   "size": 279572,
+   "sha256": null,
+   "blob_id": "615081e6e8a76050060caf45d23fe68342d8d360"
+  },
+  {
+   "path": "qrels/train.jsonl",
+   "size": 7259342,
+   "sha256": null,
+   "blob_id": "a79d32ec90473b928109bee7bb48850eeba7da28"
+  },
+  {
+   "path": "qrels/train.tsv",
+   "size": 2504643,
+   "sha256": null,
+   "blob_id": "d61fa14bd70ce45480cf5a02a5d0021b78e2d610"
+  },
+  {
+   "path": "queries.jsonl",
+   "size": 180348,
+   "sha256": "1b893f53f3f62e8f2949a6db619a9cb058e8d97901a5010e108e60c4ef5fd924",
+   "blob_id": "569f21ea926daa15bd6b66d26f062dd8bee349b7"
+  }
+ ],
+ "build": {
+  "timestamp_utc": "2026-09-08T05:17:55Z",
+  "seconds": 10.1,
+  "peak_rss_gb": 0.27,
+  "python": "3.11.0",
+  "pyarrow": "25.0.1",
+  "datasets": "5.0.1",
+  "polars": "1.44.1"
+ },
+ "crosscheck": {
+  "legacy_repo": "Hyukkyu/beir-nfcorpus",
+  "legacy_qrels_repo": "Hyukkyu/beir-nfcorpus-qrels",
+  "corpus": {
+   "legacy_rows": 3633,
+   "build_rows": 3633,
+   "same_id_set": true
+  },
+  "qrels": {
+   "test": {
+    "legacy_split": "test",
+    "legacy_rows": 12334,
+    "build_rows": 12334,
+    "same_rows": true
+   },
+   "train": {
+    "legacy_split": "train",
+    "legacy_rows": 110575,
+    "build_rows": 110575,
+    "same_rows": true
+   },
+   "dev": {
+    "legacy_split": "dev",
+    "legacy_rows": 11385,
+    "build_rows": 11385,
+    "same_rows": true
+   }
+  },
+  "status": "match"
+ }
+}

qrels/dev-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f311f8a7a224b0e595c3c41c07a846485e2292d9dceb8b38f11d74c8214df2f4
+size 21718

qrels/test-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:67c3f007e34b7d1b211d8276f1c5292751d4dcbba7791f4b315a37b3b937e7af
+size 21926

{corpus → qrels}/train-00000-of-00001.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f4cb707cb7748795a2d64af8067fc04a9bb71bb0690bc82363c7c0793d9c0751
-size 3178637
+oid sha256:62a747f001310bd406d654f940e186a7c3e2a123d5014b66cc1b81949b0662b6
+size 128581

queries/dev-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5c045749115d6d3ffe3358f6588446deef58043b5bc3cab1df8e924d045c0398
+size 6856

queries/test-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:138f4b2dd44362311fad6e2aa7911d1be64b91ee07457a3946241d9837575bd6
+size 6738

queries/train-00000-of-00001.parquet

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1005e364488b00b2fec42d02c0fa1d283f0c0044b2d2ae061b16c8eab47a78a1
-size 80445
+oid sha256:26f3560df603d38560ac850c4ca0e42b310ee7865fe23f934673ce2f7d2a4243
+size 43656