Datasetpublic

MoMonir/shamela_books_text_full

Shamela_Books_Text_Full This dataset contains the full text content of Islamic Arabic books from the Shamela Library, organized by category, book, volume, and page, with footnotes stored separately. It is designed to support Arabic NLP, digital humanities, and bibliographic analysis. 🔗 This dataset is linked to the companion metadata dataset: 👉 Shamela_Books_info via the book_id field. Update : The dataset includes the original raw files as well as a single… See the full description on the dataset page: https://huggingface.co/datasets/MoMonir/shamela_books_text_full.

sourceHugging Faceapache-2.0updated 4mo agoView on Hugging Face

2likes182downloads

Dataset Card

Shamela\Books\Text\_Full

This dataset contains the full text content of Islamic Arabic books from the Shamela Library, organized by category, book, volume, and page, with footnotes stored separately. It is designed to support Arabic NLP, digital humanities, and bibliographic analysis.

🔗 This dataset is linked to the companion metadata dataset: 👉 Shamela\_Books\_info via the book_id field.

Update :

The dataset includes the original raw files as well as a single merged file.

📊 Dataset Summary

—Total Categories: 40
—Total Books: 8,538
—Total Records (Pages): 7,552,019
—Granularity: Each row = a page of a book
—Footnotes: Stored in a separate column for clean processing

📁 Dataset Fields

Column	Description
`serial_number`	Unique row ID for each page
`category_id`	Numeric identifier for book category
`category`	Name of the category (e.g., Fiqh, Tafsir, etc.)
`book_title`	Title of the book
`book_id`	Unique ID used by Shamela.ws
`edition`	Edition info (if available)
`publisher`	Publisher name (if available)
`page_number`	Page number within the volume
`volume_number`	volume number (supports sub-parts and prefixes)
`text`	Main content text of the page
`foot_note`	Footnote text from the same page

🔗 Link to Metadata

To access full author info, publishing details, and classification, use the book_id to join this dataset with: 📘 Shamela\_Books\_info

python

# Example join (Pandas):
import pandas as pd

text_df = pd.read_csv('shamela_books_text_full.csv')
info_df = pd.read_csv('shamela_books_info.csv')

merged_df = text_df.merge(info_df, on='book_id')

📚 Volume Numbering Reference

🔹 Main Content

Pattern Example	Description	Volume Number
`5 أ`, and `5 ب`	Arabic letter multi-parts (أ = 1)	`5.2` , `5.1`
`2- 1`	Hyphenated parts	`2.1`
`47 - 48`	Multi-volume, first number used	`47`
`001`, `1`	Standard numeric volume	`1`

🧠 Usage Ideas

—Train/fine-tune Arabic language models
—Information retrieval / document segmentation
—Digital Islamic studies / literary analysis
—Linking text to authors, death years, categories, and editors

Acknowledgment and Gratitude

We extend our sincere gratitude to everyone behind the development of المكتبة الشاملة (shamela.ws), including the project maintainers, contributors, transcribers, supporters, and sponsors who have worked over many years to make this remarkable resource available. May Allah reward them abundantly and accept their efforts.

📅 Disclaimer

This dataset is based on exported data from Shamela Library software synced withShamela.ws in June 2025. Please note that the Shamela Library may add or remove entries at any time, and this dataset may not reflect future updates.

📖 Citation

If you use this dataset, please cite it as:

@dataset{shamela_books_text,
  title = {Shamela Books Text},
  author = {MoMonir},
  year = {2025},
  url = {https://huggingface.co/datasets/MoMonir/shamela_books_text}
}