README.md

metadata

license: cc-by-nc-sa-4.0
task_categories:
  - translation
  - summarization
  - question-answering
  - text-classification
language:
  - ar
  - en
tags:
  - islam
  - hadith
  - religion
  - arabic
  - multi-lingual
pretty_name: Sunnah.com Comprehensive Dataset (freococo)
size_categories:
  - 10K<n<100K

Sunnah.com Comprehensive Dataset (freococo)

Dataset Summary

This dataset is a high-quality, professional-grade collection of 39,318 Hadiths meticulously scraped and structured from Sunnah.com. It represents one of the most complete and multi-faceted Hadith databases available in the open-data community, covering the primary sources of Islamic tradition (The Kutub al-Sittah) as well as several essential secondary compilations.

Creator: freococo
Source Data: Sunnah.com
Total Rows: 39,318

Books Included

The dataset contains the following 11 collections, ordered by their traditional importance:

1. Sahih al-Bukhari (7,563 Hadiths)
2. Sahih Muslim (3,033+ narrations)
3. Sunan an-Nasa'i (5,758 Hadiths)
4. Sunan Abi Dawud (5,274 Hadiths)
5. Jami` at-Tirmidhi (3,956 Hadiths)
6. Sunan Ibn Majah (4,341 Hadiths)
7. Mishkat al-Masabih (5,945+ Hadiths)
8. Riyad as-Salihin (1,896 Hadiths)
9. Al-Adab Al-Mufrad (1,322 Hadiths)
10. Ash-Shama'il Al-Muhammadiyah (417 Hadiths)
11. Bulugh al-Maram (1,596+ Hadiths)

Field Definitions

Field Name	Description
freococo_id	Primary Key: A unique unique incrementing ID for this dataset version.
arabic_full	The complete Arabic text including narrators (Sanad) and the statement (Matn).
english_full	The complete English text (Narrator + Body text) normalized for readability.
gradings	JSON formatted string containing authenticity verdicts (Sahih, Hasan, Da'if) and the specific scholar who provided the grading (e.g., Al-Albani, Darussalam).
collection	The English name of the source book.
hadith_id	The global unique serial number used in the source URL (Web ID).
hadith_no_in_book	The canonical number as found in physical, printed editions.
book_no	The book/volume index within the collection.
book_en / book_ar	English and Arabic names of the "Book" (e.g., "The Book of Faith").
chapter_no	The specific chapter index within the book.
chapter_en / chapter_ar	English and Arabic names of the chapter.
url	Direct URL to verify the source at Sunnah.com.
sunnah_id	The hidden internal database ID used by the Sunnah.com platform.
narrator_en / text_en	Raw split components of the English translation.
arabic_sanad / arabic_matn	Raw split components of the Arabic chain and core text.
trans_vol / trans_book_no	Legacy numbering for compatibility with USC-MSA web references.

How to Load in Python

You can load this dataset easily using the datasets library or pandas:

Using datasets (Recommended)

from datasets import load_dataset

dataset = load_dataset("freococo/sunnah_dataset")
print(dataset['train'][0])

Credits and Attribution

This dataset would not be possible without the massive efforts of:

• Source Platform: Sunnah.com, who have provided a vital service to the Muslim Ummah by digitizing these texts.
• Scholars & Gradings: Sheikh Muhammad Nasiruddin al-Albani, the team at Darussalam, Sheikh Zubair Ali Za'i, and many others who provided the Takhrij (verdicts).
• Translators: Dr. Muhammad Muhsin Khan (Bukhari), Siddiqui (Muslim), and the Darussalam team for their tireless work in translating these narrations into English.

License & Usage

This dataset is shared under the Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International (CC BY-NC-SA 4.0).

• Attribution: You must credit Sunnah.com but NO need to give CREDIT to the dataset creator freococo.
• Non-Commercial: This data should not be sold or used for commercial profit.
• Share Alike: If you remix or build upon this data, you must distribute your contributions under the same license.