A bilingual (Arabic + English) plaintext corpus of Wikipedia articles about Oman, built from the official MediaWiki APIs of ar.wikipedia.org and en.wikipedia.org. 1.
Dataset Card
By Jaifar Alshizawi, published under cc-by-sa-4.0, revision 86f6f7e4f821.
A bilingual (Arabic + English) plaintext corpus of Wikipedia articles about Oman,
built from the official MediaWiki APIs of ar.wikipedia.org and en.wikipedia.org.
مجموعة نصوص ثنائية اللغة (العربية والإنجليزية) من مقالات ويكيبيديا المتعلقة بسلطنة عُمان، مبنية من واجهات ميدياويكي الرسمية.
Dataset Summary
| Arabic (عربي) | English | |
|---|---|---|
| Articles | 1,205 | 522 |
| Total words | 702,327 | 400,257 |
| Categories walked | 258 | 147 |
| Harvest date | 2026-09-20 | 2026-09-20 |
Total: 1,727 articles, 1,102,584 words.
Methodology
- Walked the Oman root category tree (depth ≤ 2), pruning maintenance and stub categories
(
Category:Omanon enwiki;تصنيف:سلطنة عمانon arwiki). - Ranked candidate articles by page length and kept the top ones.
- Fetched clean plaintext extracts via the official
prop=extractsAPI (one article per request, globally rate-limited to ~100 requests/minute). - Dropped articles with fewer than 200 characters of text.
Data Fields
| Field | Type | Description |
|---|---|---|
id |
string | Stable ID, e.g. wiki-en-25554 |
pageid |
int | Wikipedia page ID |
lang |
string | en or ar |
title |
string | Article title |
url |
string | Canonical article URL |
categories |
list[string] | Oman category tree path(s) the article was found under |
text |
string | Full plaintext (section headings preserved as lines) |
word_count / char_count |
int | Article size |
source |
string | Origin wiki |
license |
string | License of the underlying text |
fetched |
string | Harvest date (ISO) |
Licensing and Attribution
Wikipedia text is CC BY-SA 4.0. This corpus is a derivative collection: any public use must credit Wikipedia contributors and link the articles (URLs included per row), and derived works must remain share-alike. See Wikipedia: Copyrights.
Intended Uses & Limitations
- Pretraining/fine-tuning corpora for Oman-focused language models; RAG knowledge bases; Arabic NLP research (Omani content is underrepresented in existing corpora).
- Wikipedia reflects contributor biases; verify facts against primary sources before high-stakes use.
- Not a substitute for official government statistics.
Maintenance
Maintained by Menasaat | Virtual Platforms LLC (menasaat). Rebuild script: category-tree walk + official API extracts; planned refresh quarterly.
Structure
default 1,727 rows
| Split | Rows | Size |
|---|---|---|
| train | 1,727 | 10.5 MB |
Details
- Repository
- jaifar/oman-wikipedia-corpus
- Publisher
- Jaifar Alshizawi
- Task category
- Text generation
- Tags
- oman, arabic, wikipedia
- Size category
- 1K<n<10K
- Languages
- en, ar
- Revision
- 86f6f7e4f8217291829190cec76655d8fd21e566
- Last updated
- 2026-09-21
Files
4 files, 3.5 MB in total.
Every file
| File | Type | Size | SHA-256 |
|---|---|---|---|
| data/oman_wiki_ar.parquet | Data | 2.4 MB | a508b9529256 |
| data/oman_wiki_en.parquet | Data | 1.1 MB | bfc4f25a9645 |
| README.md | Documentation | 3.0 KB | — |
| .gitattributes | Repository | 2.5 KB | — |
License and Download
- License
- cc-by-sa-4.0
- Access
- No access gate
Released by Jaifar Alshizawi through its official repository on Hugging Face. Read the license.