Upload README.md with huggingface_hub

README.md

@@ -4,6 +4,10 @@ configs:
   data_files:
   - split: train
     path: epo_patents/*.parquet
+- config_name: eu_science
+  data_files:
+  - split: train
+    path: eu_science/*.parquet
 - config_name: wikidata_seed
   data_files:
   - split: train
@@ -21,12 +25,16 @@ Volumes below describe what is available beyond these samples.
 
 | Config | Sample | Available |
 |---|---|---|
-| `epo_patents` | 5,000 patents / 72,662 sections / 102M tokens, with source PDFs | 1.08M patents processed (2020-2025); 8.25M-document full-text backfile 1978-2025 (5.49M with description/claims, ~48B words) |
+| `epo_patents` | 5,000 patents / 72,662 sections / 102M tokens, with source PDFs | 1.08M patents processed (2020-2025); 8.25M-document full-text backfile 1978-2025 (~48B words) |
+| `eu_science` | 4,704 papers / 69.5M tokens, with source PDFs | 1.43M documents, ~19B tokens, 204 languages |
 | `wikidata_seed` | 1.45M triples / 42k entities | 558M triples, ~34.5 GB |
 | `personas` | 100k personas / 108 countries | 57.3M personas |
 
-More configs to follow: financial regulation (24 regulators, 12+ jurisdictions), 3GPP
-specifications, European open science.
+More configs to follow: financial regulation (24 regulators, 12+ jurisdictions) and 3GPP
+specifications.
+
+Two configs ship the **source PDFs** alongside the processed text, so extraction quality can be
+judged directly and the pairing can be used for document-understanding training.
 
 ## epo_patents
 
@@ -36,25 +44,31 @@ document **section**, with document metadata repeated on each row (`document_id`
 
 Composition notes: bibliographic front pages, figure lists, and boilerplate disclaimers are
 removed; paragraphs are reassembled across column and page breaks; margin line-numbering is
-dropped. Citation lists are kept but flagged (`is_document_list`) rather than deleted, since
-they carry citation structure. Bibliographic fields come from the document itself (INID codes),
-so `identifier` is the publication number and supplementary fields sit in `metadata` as JSON
-(`application_number`, `filing_date`, `ipc`, `inventors`, `kind_code`, `grant_or_prior_date`,
-`source_file`).
+dropped. Citation lists are kept but flagged (`is_document_list`). Bibliographic fields come
+from the document itself (INID codes), so `identifier` is the publication number and
+supplementary fields sit in `metadata` as JSON.
 
 `date` is the publication year. Some documents are republications after opposition, where the
 original grant year differs — that year is preserved in `metadata`.
 
-### Source PDFs
+**Source PDFs**: `epo_patents/pdfs/epo_pdfs_<year>.tar` (6 files, 3.83 GB). Members are named by
+publication number — `2023/EP1343809B2.pdf` — so they join to `document_id`, even though the
+files as distributed by the EPO are named by application number.
+`epo_patents/epo_5k_pdf_manifest.parquet` maps `document_id -> tar, member, source_file, bytes`.
+
+## eu_science
+
+Scientific literature processed from PDF at scale — open-access papers across many languages,
+one row per document. 4,704 papers drawn evenly across the corpus so the language mix is
+representative rather than clustered.
 
-`epo_patents/pdfs/epo_pdfs_<year>.tar` (6 files, 3.83 GB total) hold the original PDF for every
-patent in the config. Tar members are named by **publication number** — `2023/EP1343809B2.pdf` —
-so they join directly to `document_id`, even though the files as distributed by the EPO are named
-by application number. `epo_patents/epo_5k_pdf_manifest.parquet` maps
-`document_id -> tar, member, source_file, bytes`.
+`language` is empty for ~22% of rows: those documents come from a source pile whose language
+metadata was blank, and the shards they sit in predate language detection. Everything else
+carries a resolved language name.
 
-This pairs each processed text with the page images it came from, for evaluating extraction
-quality or for document-understanding training.
+**Source PDFs**: `eu_science/pdfs/science_pdfs_<nn>.tar` (5 files, 5.27 GB). Members are named
+`<openalex_id>.pdf`, joining directly to `identifier`.
+`eu_science/science_5k_pdf_manifest.parquet` maps `identifier -> tar, member, bytes`.
 
 ## wikidata_seed
 
@@ -77,10 +91,9 @@ Synthetic population personas across 108 countries, demographically grounded: na
 place (with coordinates and settlement tier), marital status, education, employment status,
 occupation, religion, monthly income in PPP dollars, and health conditions.
 
-Location is given as place name with coordinates and settlement tier (administrative subdivision
-codes are not included). This sample is drawn across the full corpus and shuffled; per-country
-counts stay proportional to the underlying population weighting. `persona_id` is globally unique
-(`country`-`id`).
+Location is given as place name with coordinates and settlement tier. This sample is drawn
+across the full corpus and shuffled; per-country counts stay proportional to the underlying
+population weighting. `persona_id` is globally unique (`country`-`id`).
 
 ## Licensing