The goal of the Indic NLP Library is to build Python based libraries for common text processing and Natural Language Processing in Indian languages. Indian languages share a lot of similarity in terms of script, phonology, language syntax, etc. and this library is an attempt to provide a general solution to very commonly required toolsets for Indian language text.
The library provides the following functionalities:
- Text Normalization
- Script Information
- Word Tokenization and Detokenization
- Sentence Splitting
- Word Segmentation
- Syllabification
- Script Conversion
- Romanization
- Indicization
Note: Shatanuvadak translation and BrahmiNet transliteration APIs are no longer supported. You can use newer IndicTrans translation and IndicXlit transliteration models we developed at AI4Bharat. In fact, you can find many state-of-the-art datasets and models on the AI4Bharat homepage.
The data resources required by the Indic NLP Library are hosted in a different repository. These resources are required for some modules. You can download from the Indic NLP Resources project.
If you are interested in Indian language NLP resources, you should check the Indic NLP Catalog for pointers.
- Python 3.x
- (For Python 2.x version check the tag
PYTHON_2.7_FINAL_JAN_2019. Not actively supporting Python 2.x anymore, but will try to maintain as much compatibility as possible)
- (For Python 2.x version check the tag
- Indic NLP Resources
- Urduhack: Needed only if Urdu normalization is required. It has other dependencies like Tensorflow.
- Other dependencies are listed in setup.py
-
Installation from pip:
pip install indic-nlp-library -
If you want to use the project from the github repo, add the project to the Python Path:
- Clone this repository
- Install dependencies:
pip install -r requirements.txt - Run:
export PYTHONPATH=$PYTHONPATH:<project base directory>
-
In either case, export the path to the Indic NLP Resources directory
Run:
export INDIC_RESOURCES_PATH=<path to Indic NLP resources>
You can use the Python API to access all the features of the library. Many of the most common operations are also accessible via a unified commandline API.
Check this IPython Notebook for examples to use the Python API.
- You can find the Python 2.x Notebook here
You can find detailed documentation HERE
This documents the Python API as well as the commandline reference.
If you use this library, please include the following citation:
@misc{kunchukuttan2020indicnlp,
author = "Anoop Kunchukuttan",
title = "{The IndicNLP Library}",
year = "2020",
howpublished={\url{https://github.com/anoopkunchukuttan/indic_nlp_library/blob/master/docs/indicnlp.pdf}}
}
You can find the document HERE
http://anoopkunchukuttan.github.io/indic_nlp_library
Anoop Kunchukuttan (anoop.kunchukuttan@gmail.com)
0.81 : 26 May 2021
- Bug fix in version number extraction
0.80 : 24 May 2021
- Improved sentence splitting
- Bug fixes
- Support for Urdu Normalizer
0.71 : 03 Sep 2020
- Improved documentation
- Bug fixes
0.7 : 02 Apr 2020:
- Unified commandline
- Improved documentation
- Added setup.py
0.6 : 16 Dec 2019:
- New romanizer and indicizer
- Script Unifiers
- Improved script normalizers
- Added contrib directory for sample uses
- changed to MIT license
0.5 : 03 Jun 2019:
- Improved word tokenizer to handle dates and numbers.
- Added sentence splitter that can handle common prefixes/honorofics and uses some heuristics.
- Added detokenizer
- Added acronym transliterator that can convert English acronyms to Brahmi-derived scripts
0.4 : 28 Jan 2019: Ported to Python 3, and lots of feature additions since last release; primarily around script information, script similarity and syllabification.
0.3 : 21 Oct 2014: Supports morph-analysis between Indian languages
0.2 : 13 Jun 2014: Supports transliteration between Indian languages and tokenization of Indian languages
0.1 : 12 Mar 2014: Initial version. Supports text normalization.
Indic NLP Library is released under the MIT license