TrueCase

A language independent, statistical, language modeling based tool in Python that restores case information for text.

The model was inspired by the paper of Lucian Vlad Lita et al., tRuEcasIng but with some simplifications.

A model trained on NLTK English corpus comes with the package by default, and for other languages, a script is provided to create the model. This model is not perfect, train the system on a large and recent dataset to achieve the best results (e.g. on a recent dump of Wikipedia).

Prerequisites

Python 3

The project uses NLTK. Find install instructions here.

Installing

pip3 install truecase

Usage

Simple usecase:

>>> import truecase
>>> truecase.get_true_case('hey, what is the weather in new york?')
'Hey, what is the weather in New York?''

Training your own model

TODO. For now refer to Trainer.py

Contributing

I see a lot of space for improvement. Feel free to fork and improve. Do sent a pull request.

Authors

Dalton Fury - Initial work - daltonfury42

License

This project is licensed under the MIT License - see the LICENSE.md file for details

Acknowledgments

Lucian Vlad Lita et al., tRuEcasIng
Borrowed a lot of code, and the idea from truecaser by nreimers

makai281/truecase