license: cc-by-4.0
task_categories:
- text-generation
language:
- fa
tags:
- persian
- corpus
- poem
- text
pretty_name: Ganjoor - Persian Poem Corpus
size_categories:
- 100K<n<1M
Dataset Card for Dataset Name
This is the csv format of the Ganjoor Database that is published in their github
Dataset Details
- Curated by: Navid Abbaspoor
- Language(s) (NLP): Persian (Farsi)
- License: Creative Commons Attribution 4.0 International (cc-by-4.0)
Dataset Description
This dataset contains almost all of poems by Iran's great poets through many many past years till now. The original database was tabular, that I convert it to a csv format that contains these columns:
- id: id of the poem in the original database
- poem: name of the poem
- poet: name of the poet
- cat: category of the poem
- text: text and verses of the poem
Dataset Sources [optional]
- Repository: [More Information Needed]
Usage
from datasets import load_dataset
dataset = load_dataset("mabidan/ganjoor")
Dataset Creation
Persian language is one of the low-resource languages in NLP tasks. I'm trying to collect persian datasets to help the community to build helpful tools for this language. So I'm usually collect persian data from web and transform it to a structure that is convinient for my tasks in NLP.
Bias, Risks, and Limitations
I didnt do any edit or modification in text and verses. So all the data is comming from the original Ganjoor database.
Recommendations
I double checked the data and found that some of poems have no text, I guess there might be a problem in the preprocessing steps or in the actual database. You might want to drop that rows from data if you want to use the dataset.
Users should be made aware of the risks, biases and limitations of the dataset. More information needed for further recommendations.
Dataset Card Contact
For any question or colaboration you can contact me: [email protected]