Le’s Lesk

A fast Python 3 Word-Sense Disambiguation package (WSD) using the extended LESK algorithm

Install

lelesk is available on PyPI and can be installed using pip

pip install lelesk

Lelesk uses NLTK lemmatizer and yawlib wordnet API. To install NLTK data, start a Python prompt, import nltk and then download the required data

$ python3
Python 3.6.9 (default, Jan 26 2021, 15:33:00) 
[GCC 8.4.0] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import nltk
>>> nltk.download(['stopwords', 'punkt', 'averaged_perceptron_tagger', 'wordnet'])

Download and extract yawlib pre-built databases to ~/wordnet.

For more information:

Installing NLTK data: https://www.nltk.org/data.html
Installing Yawlib wordnets: https://pypi.org/project/yawlib/

Command-line tools

To disambiguate a sentence, run this command on the terminal:

python3 -m lelesk wsd "I go to the bank to get money."

To perform word-sense disambiguation on a text file, prepare a text file with each line is a sentence.

For example here is the content of the file demo.txt

I go to the bank to withdraw money.
I sat at the river bank.

you then can run the following command

# output to TTL/JSON (a single file)
python3 -m lelesk file demo.txt demo_wsd_output.json --ttl json

# output to TTL/TSV (multiple TSV files)
python3 -m lelesk file demo.txt demo_wsd_output.json --ttl tsv

Issues

If you have any issue, please report at https://github.com/letuananh/lelesk/issues