Datasetpublic

MohamedRashad/Al-Maktabah-Al-Shamilah

Al-Maktabah-Al-Shamilah Dataset (Dense Islamic collection of books) About the Dataset This dataset is drawn from Al-Maktabah Al-Shamilah, one of the most widely used digital libraries of Arabic and Islamic heritage. It brings together thousands of books in a single place — a resource many scholars, students, and readers have relied on for years. Here, the collection has been transformed into a structured, machine-readable format to support… See the full description on the dataset page: https://huggingface.co/datasets/MohamedRashad/Al-Maktabah-Al-Shamilah.

sourceHugging Faceupdated 1y agoView on Hugging Face

14likes148downloads

Dataset Card

Al-Maktabah-Al-Shamilah Dataset (Dense Islamic collection of books)

About the Dataset

This dataset is drawn from Al-Maktabah Al-Shamilah, one of the most widely used digital libraries of Arabic and Islamic heritage. It brings together thousands of books in a single place — a resource many scholars, students, and readers have relied on for years.

Here, the collection has been transformed into a structured, machine-readable format to support research in Arabic Natural Language Processing (NLP), linguistics, and digital humanities.

—📚 Books: 6,521 unique works
—📝 Segments: 5,9M text rows (each one a passage from a book)
—🌍 Language: Arabic (mostly Classical Arabic, some Modern Standard Arabic)

Think of it as a way to bridge centuries of scholarship with modern computational tools.

What You’ll Find

The dataset is very simple: just two columns.

json

{
  "Book_name": "تفسير ابن كثير",
  "text": "الْحَمْدُ لِلَّهِ الَّذِي أَنْزَلَ عَلَى عَبْدِهِ الْكِتَابَ..."
}

—Book\_name: The title of the book
—text: A segment of that book

That’s it — clean and straightforward.

How You Might Use It

The data is as clean and ready as it was in the shamela website. A few things you could do:

—Train a language model on Classical Arabic
—Build a search engine across centuries of Islamic scholarship
—Experiment with topic modeling or semantic clustering
—Explore the evolution of style and vocabulary over time

Things to Keep in Mind

Like any dataset, this one comes with context:

—🕌 Nature of the content: These are mostly religious and scholarly texts. They deserve respectful and careful use.
—⚖️ Licensing: Al-Maktabah Al-Shamilah contains many works that are public domain, but not all. Please check the legal status before redistribution or commercial use.
—🔎 Structure: Segments are based on how the website organizes the text. They may not always align perfectly with semantic or thematic breaks.

Why This Matters

For too long, the Arabic language — with all its richness and history — has been underrepresented in the digital and computational world. While other languages have seen vast investment and technological support, Arabic has often been left behind, its texts scattered, unstructured, or inaccessible to modern tools.

Preserving and organizing these works is not just a technical achievement; it is an act of care and responsibility. It is a way to honor the legacy of a language that has carried knowledge, culture, and faith for centuries, and to ensure it continues to thrive in the age of data and AI.

My wish is that this dataset contributes, even in a small way, to strengthening Arabic, empowering its speakers, and ensuring that our language is not only preserved but made vibrant in the age of AI and data-driven discovery.

Citation

If you use this dataset, please cite it as:

bibtex

@dataset{almaktabah_alshamilah_2025,
  title        = {Al-Maktabah Al-Shamilah Dataset},
  year         = {2025},
  url          = {https://huggingface.co/datasets/MohamedRashad/Al-Maktabah-Al-Shamilah},
  note         = {Scraped from https://shamela.ws}
}

✨ This dataset is meant to be a bridge — between the richness of Arabic scholarly heritage and the curiosity of today’s researchers. Explore it with openness, use it with respect, and share what you learn.