license: other
license_name: waqf
license_link: LICENSE
language:
- ar
configs:
- config_name: pipe
data_files: quran-simple-ak.csv
sep: '|'
Qur'an Simple Text Dataset
This dataset is licensed as Waqf (see license and credits). It contains the text of the Qur'an without and with diacritics.
It is based on files downloaded from Tanzil.net where you can read/listen online, view translation and tafsir (explanation) in different languages.
The available columns are:
surah number, surah name, no. of ayas, class (Makki/Madani), order of revelation, aya number, aya text (no diacritics), aya text with diacritics.
Scroll the viewer to the right to see the last column(s).
There are various applications based on this dataset, including educational and NLP. Make sure you use the text column for NLP (d_text will not work).
There are many Qur'an related datasets on this platform (text, tafsir, recitation ... etc.). This is just a simple and small one with some metadata.
Usage
from datasets import load_dataset
import pandas as pd
ds = load_dataset('akhooli/quran-simple-text', split="train", trust_remote_code=True)
df = ds.to_pandas()
df.head()
A full colab notebook example using Arabic-ColBERT-100k model is here.