Datasets:

Bayanat-Labs
/

arabic-wikisource-pd

The text, chapters, footnotes, verse lines, page lists and categories come from the official Wikimedia dumps of Arabic Wikisource (arwikisource), run of 1 October 2026 (https://dumps.wikimedia.org/arwikisource/20261001/): pages-articles.xml.bz2, page_props.sql.gz, categorylinks.sql.gz and linktarget.sql.gz. The SHA-1 of each file was checked against the value Wikimedia publishes and a SHA-256 recorded in PROVENANCE.jsonl.

Licence of the transcribed text: CC BY-SA 4.0, https://creativecommons.org/licenses/by-sa/4.0/ (Arabic Wikisource's site licence; full text in LICENSE-CC-BY-SA-4.0).
Authors of the transcriptions: the Arabic Wikisource contributors listed in the history of each page. Each record carries history_url (the history of the work's root page) and source_pages, the main-namespace and Page-namespace pages used, each with its URL, revision id and revision time. Contributor account names are not included in this dataset.
Changes: wikitext was converted to plain text; headers, navigation, running heads, page numbers and editing templates were removed; footnotes were moved to a separate field; the pages of a work were joined into chapters and the chapters into one text.
Scans: no scan images or page images are included. Page-level transcriptions with their proofreading status are released in the Bayanat Labs Arabic Document Atlas (https://huggingface.co/datasets/Bayanat-Labs/arabic-document-atlas, config community_text).

Wikidata (people, death dates, genres, publication dates)

Statements were read with the Wikidata API (wbgetentities) from 30 to 31 October 2026; names of translators read from titles and first lines of texts were looked up by their Arabic Wikipedia title and the Wikidata site search. Wikidata data is available under CC0 1.0 (https://creativecommons.org/publicdomain/zero/1.0/).

Bayanat Labs tables

authors, pd_decisions, stats/ and qa/, and the fields computed by Bayanat Labs in the works configs: CC BY 4.0 (full text in LICENSE-CC-BY-4.0).

Wikisource and Wikidata are trademarks of the Wikimedia Foundation. This dataset is not affiliated with or endorsed by the Wikimedia Foundation.