Arabic OCR Books Dataset (ocr_arabic_books)

This repository is a structurally aligned Arabic Optical Character Recognition (OCR) dataset. It hosts a collection of classical Islamic books, organized by separate configurations (subsets) to support modular loading without filename collisions.

---

📚 Master Book Inventory

Total running pages in repository: 181,427

#	Book Name (English)	Book Name (Arabic)	Subset / Config Name	Page Count	Image Index Range
1	Taqreeb Fatawa wa Rasa'il Ibn Taymiyyah	تقريب فتاوى ورسائل شيخ الإسلام ابن تيمية	default	3,537	0001.jpg – 3537.jpg
2	Sunan Ibn Majah	سنن ابن ماجه	sunan_ibn_majah	3,012	3538.jpg – 6549.jpg
3	Al-Musannaf by Ibn Abi Shaibah	المصنف لابن أبي شيبة	musannaf_ibn_abi_shaibah	10,475	6550.jpg – 17024.jpg
4	Tafsir al-Tabari (Jami' al-Bayan)	جامع البيان عن تأويل آي القرآن (تفسير الطبري)	tafsir_al_tabari	16,700	17025.jpg – 33724.jpg
5	Al-Mabsut by al-Sarakhsi	المبسوط للسرخسي	al_mabsut	6,277	33725.jpg – 40001.jpg
6	Al-Ghayah fi Ikhtisar al-Nihayah	الغاية في اختصار النهاية	al_ghayah	3,302	40002.jpg – 43328.jpg*
7	Musnad Ahmad ibn Hanbal	مسند الإمام أحمد بن حنبل	musnad_ahmad	23,338	43304.jpg – 66641.jpg
8	Sahih al-Bukhari (Dar al-Taseel)	صحيح البخاري (طبعة دار التأصيل)	sahih_al_bukhari	4,719	66642.jpg – 71360.jpg
9	Al-Musnad al-Musannaf al-Mu'allal	المسند المصنف المعلل	al_musnad_al_muallal	22,631	71361.jpg – 93991.jpg
10	Kanz al-Ummal fi Sunan al-Aqwal wa al-Af'al	كنز العمال في سنن الأقوال والأفعال	kanz_al_ummal	11,533	93992.jpg – 105524.jpg
11	Al-Jami' al-Kamil fi al-Hadith al-Sahih al-Shamil	الجامع الكامل في الحديث الصحيح الشامل	al_jami_al_kamil	8,497	105525.jpg – 114021.jpg
12	Al-Ta'liqat al-Hissan 'ala Sahih Ibn Hibban	التعليقات الحسان على صحيح ابن حبان	al_taliqat_al_hissan	4,396	114022.jpg – 118417.jpg
13	Jam' al-Jawami' / Al-Jami' al-Kabir	جمع الجوامع المعروف بـ الجامع الكبير	jam_al_jawami	18,715	118418.jpg – 137132.jpg
14	Al-Matalib al-Aliyah bi-Zawa'id al-Masanid al-Thamaniyah	المطالب العالية بزوائد المسانيد الثمانية	al_matalib_al_aliyah	13,191	137133.jpg – 150323.jpg
15	Majma' al-Zawa'id wa-Manba' al-Fawa'id	مجمع الزوائد ومنبع الفوائد	majma_al_zawaid	3,478	150324.jpg – 153801.jpg
16	Jami' al-Masanid wa-al-Sunan	جامع المسانيد والسنن	jami_al_masanid	6,314	153802.jpg – 160115.jpg
17	Kitab al-Ahadith al-Mukhtarah	الأحاديث المختارة (المستخرج من الأحاديث المختارة)	kitab_ahadith_mukhtharah	3,882	160116.jpg – 163997.jpg
18	Sharh al-Sunnah by al-Baghawi	شرح السنة للبغوي	sharh_al_sunnah	6,133	163998.jpg – 170130.jpg
19	Ma'rifat al-Sunan wa al-Athar by al-Bayhaqi	معرفة السنن والآثار للبيهقي	marifat_al_sunan_wa_al_athar	4,204	170131.jpg – 174334.jpg
20	Shu'ab al-Iman by al-Bayhaqi	الجامع لشعب الإيمان للبيهقي	shuab_al_iman	7,093	174335.jpg – 181427.jpg
Total				181,427

*Note: Specific pages representing raw cover layout variations and visual manuscript photo appendices have been intentionally skipped to ensure high-fidelity print OCR training signals.

---

🛠️ Dataset Structure

Every book subset inside this repository uses a standardized, clean schema:

Field Name	Type	Description
file_name	string	Relative path to the image on disk (present in certain sub-configurations).
image	Image	Scanned page image (JPEG format, 150 DPI)
page_header	string	Spatially aligned running headers containing page numbers and titles (or null if the book lacks running headers or has had headers cropped out to avoid layout mismatches).
body	string	Clean Arabic main body text (fully stripped of HTML tags, normalized to Unix newlines, stripped of tashkeel/diacritics where appropriate to match printed page layouts, and free of database anchors).
footnotes	string	Clean footnotes and/or bottom-centered Arabic page numbers matching the printed layout (null if none exist).

---

💻 How to Use

Because the repository is formatted in native Parquet shards, you can load or stream any individual book subset directly using its config_name:

from datasets import load_dataset

# 1. Load Taqreeb Fatawa wa Rasa'il Ibn Taymiyyah (default subset)
dataset_taqreeb = load_dataset("freococo/ocr_arabic_books", split="train")

# 2. Load Sunan Ibn Majah subset
dataset_majah = load_dataset("freococo/ocr_arabic_books", "sunan_ibn_majah", split="train")

# 3. Load Al-Musannaf by Ibn Abi Shaibah subset
dataset_musannaf = load_dataset("freococo/ocr_arabic_books", "musannaf_ibn_abi_shaibah", split="train")

# 4. Load Tafsir al-Tabari subset
dataset_tabari = load_dataset("freococo/ocr_arabic_books", "tafsir_al_tabari", split="train")

# 5. Load Al-Mabsut by al-Sarakhsi subset
dataset_mabsut = load_dataset("freococo/ocr_arabic_books", "al_mabsut", split="train")

# 6. Load Al-Ghayah fi Ikhtisar al-Nihayah subset
dataset_ghayah = load_dataset("freococo/ocr_arabic_books", "al_ghayah", split="train")

# 7. Load Musnad Ahmad subset
dataset_ahmad = load_dataset("freococo/ocr_arabic_books", "musnad_ahmad", split="train")

# 8. Load Sahih al-Bukhari subset
dataset_bukhari = load_dataset("freococo/ocr_arabic_books", "sahih_al_bukhari", split="train")

# 9. Load Al-Musnad al-Musannaf al-Mu'allal subset
dataset_muallal = load_dataset("freococo/ocr_arabic_books", "al_musnad_al_muallal", split="train")

# 10. Load Kanz al-Ummal subset
dataset_kanz = load_dataset("freococo/ocr_arabic_books", "kanz_al_ummal", split="train")

# 11. Load Al-Jami' al-Kamil subset
dataset_kamil = load_dataset("freococo/ocr_arabic_books", "al_jami_al_kamil", split="train")

# 12. Load Al-Ta'liqat al-Hissan subset
dataset_hissan = load_dataset("freococo/ocr_arabic_books", "al_taliqat_al_hissan", split="train")

# 13. Load Jam' al-Jawami' subset
dataset_jawami = load_dataset("freococo/ocr_arabic_books", "jam_al_jawami", split="train")

# 14. Load Al-Matalib al-Aliyah subset
dataset_matalib = load_dataset("freococo/ocr_arabic_books", "al_matalib_al_aliyah", split="train")

# 15. Load Majma' al-Zawa'id subset
dataset_zawaid = load_dataset("freococo/ocr_arabic_books", "majma_al_zawaid", split="train")

# 16. Load Jami' al-Masanid subset
dataset_masanid = load_dataset("freococo/ocr_arabic_books", "jami_al_masanid", split="train")

# 17. Load Kitab al-Ahadith al-Mukhtarah subset
dataset_mukhtarah = load_dataset("freococo/ocr_arabic_books", "kitab_ahadith_mukhtharah", split="train")

# 18. Load Sharh al-Sunnah subset
dataset_sunnah = load_dataset("freococo/ocr_arabic_books", "sharh_al_sunnah", split="train")

# 19. Load Ma'rifat al-Sunan wa al-Athar subset
dataset_sunan_athar = load_dataset("freococo/ocr_arabic_books", "marifat_al_sunan_wa_al_athar", split="train")

# 20. Load Shu'ab al-Iman subset
dataset_shuab = load_dataset("freococo/ocr_arabic_books", "shuab_al_iman", split="train")

# Access and preview
first_page = dataset_shuab[0]
print(first_page["body"])         # Displays the cleaned main body text
print(first_page["footnotes"])    # Displays the clean footnotes text (or None if empty)
first_page["image"].show()        # Automatically opens the visual page JPEG