Update README.md

README.md

@@ -15,62 +15,30 @@ license_link: https://huggingface.co/datasets/zhs039/frames-oracle-corpus/blob/m
 
 # FRAMES Oracle Corpus
 
-A question-scoped offline corpus derived from
-[google/frames-benchmark](https://huggingface.co/datasets/google/frames-benchmark)
-and its annotated Wikipedia pages. This is a community-prepared subset, not an
-official Google release. It contains data, not saved agent runs or generated answers.
+An offline Wikipedia document corpus derived from [google/frames-benchmark](https://huggingface.co/datasets/google/frames-benchmark), packaging each question with its ground-truth reference articles in Markdown.
 
 ## Contents
 
-All data files are under `corpus/`; the repository root holds only this dataset card, license and repository metadata.
+All dataset files are located under `corpus/`:
 
-- `corpus/frames.tsv`: 822 original questions and reference answers, retaining original row IDs and columns.
-- `corpus/index.json`: `questions` maps question IDs to document IDs; `documents` maps document IDs to title, source URL and Markdown filename.
-- `corpus/doc_0001.md` through `corpus/doc_2500.md`: 2,500 downloaded Wikipedia documents.
+- `corpus/frames.tsv`: 822 questions and reference answers from FRAMES, retaining original row IDs and columns.
+- `corpus/index.json`: Maps `questions` (question IDs to document IDs) and `documents` (document IDs to title, source URL, and Markdown filename).
+- `corpus/doc_0001.md` – `corpus/doc_2500.md`: 2,500 extracted Wikipedia articles in Markdown.
 
-## Preparation and limitations
+## Preparation & Notes
 
-The source is FRAMES `test.tsv` at revision `58d9fb6330f3ab1316d1eca12e5e8ef23dcc22ef`.
-Questions **141** and **540** are excluded because required English biographies
-are unavailable. Question **823** retains its original Simple English Wikipedia
-page; no translation was generated. Question text and reference answers were not rewritten.
-
-Wikipedia HTML was converted to article-body Markdown, retaining tables and
-removing navigation, scripts and reference-list clutter. Combined URL annotations
-were split. The malformed Pokemon annotation in question 644 and double-encoded
-apostrophe in question 764 were resolved to the corresponding English pages.
-Question 650 retains its annotated category-page content.
-
-These are downloaded snapshots, **not** historical revisions contemporary with
-the benchmark. Wikipedia edits and extraction can change the evidence. Keep the
-corpus fixed across comparisons. Full gold documents are available per question;
-this is not a search or retrieval benchmark.
+- **Source revision**: Built from `google/frames-benchmark` (`test.tsv` at commit `58d9fb6330f3ab1316d1eca12e5e8ef23dcc22ef`).
+- **Filtering & link resolution**: Questions 141 and 540 are excluded due to missing English biographies. Question 823 retains its original Simple English Wikipedia article, and Question 650 retains its category-page content. Combined URL annotations were split, and malformed URLs in questions 644 and 764 were resolved to their target English pages.
+- **Formatting**: Wikipedia HTML was converted to clean article-body Markdown, preserving tables while stripping navigation, scripts, and reference clutter. Note that these are static snapshots downloaded after benchmark publication rather than historical Wikipedia revisions.
 
 ## Download
 
 ```bash
-uvx --from huggingface_hub hf download zhs039/frames-oracle-corpus --repo-type dataset --include "corpus/*" --local-dir .
+uvx --from huggingface_hub hf download zhs039/frames-oracle-corpus \
+  --repo-type dataset --include "corpus/*" --local-dir .
 ```
 
-The dataset is public. Load `corpus/frames.tsv` and use `corpus/index.json` to expose only the
-current question's documents.
-
-## Licenses and attribution
-
-The collection is **not uniformly Apache-2.0**.
-
-- **Questions and answers (`corpus/frames.tsv`)**: derived from Google's
-  [FRAMES benchmark](https://huggingface.co/datasets/google/frames-benchmark), whose
-  dataset card declares **Apache License 2.0**. A copy is provided in
-  `LICENSE-APACHE-2.0.txt`. Modifications: rows 141 and 540 removed; retained rows
-  packaged with a local document index.
-- **Wikipedia text (`corpus/doc_*.md`)**: copyright belongs to the respective Wikipedia
-  contributors. Extracted/adapted text is distributed under
-  [CC BY-SA 4.0 International](https://creativecommons.org/licenses/by-sa/4.0/legalcode.en),
-  subject to separately identified material's applicable terms. `corpus/index.json`
-  preserves each article's title and source URL; each source page's history lists
-  its contributors. Extraction modifications are described above. See
-  [Wikimedia Terms of Use, section 7](https://foundation.wikimedia.org/wiki/Policy:Terms_of_Use/en#7._Licensing_of_Content).
+## License & Attribution
+- Questions and answers (corpus/frames.tsv): Derived from Google's FRAMES benchmark under the Apache License 2.0 (rows 141 and 540 removed; packaged with local document mappings).
 
-Redistributors must preserve attribution and comply with each component's
-applicable license; Wikipedia text is not relicensed as Apache-2.0.
+- Wikipedia articles (corpus/doc_*.md): Copyright belongs to the respective Wikipedia contributors and is distributed under CC BY-SA 4.0 International. Article titles and source URLs are preserved in corpus/index.json.