atnlp

A tool for natural language processing in python based on scipy and sklearn.

Current focus is topic labelling. Read the docs for more information.

Technologies

atnlp uses the following technologies:

• conda - datascience platform
• scipy - datascience tools
• sphinx - documentaion
• bumpversion - semantic versioning
• github - software development platform
• travis - continuous integration

Quick start

More details in quickstart docs

Prerequisites

• conda (anaconda / miniconda) - follow the installation instructions for your platform and select python 3 version.

After installing update conda from the conda-forge repo:

conda update conda -c conda-forge 

Install (production)

Install atnlp (with pip) and its dependencies (with conda):

wget https://raw.githubusercontent.com/wedavey/atnlp/master/envprod.yml
conda env create -f envprod.yml -n atnlp
conda activate atnlp

Install (development)

Fork wedave/atnlp then install from github:

git clone git@github.com:<your-user-name>/atnlp.git
conda env create -f atnlp/envdev.yml -n atnlp-dev
conda activate atnlp-dev
cd atnlp; python setup.py develop

Running tests

TODO...

Deployment

Start training topic models straight away using scripts, or open a notebook and start hacking.

Versioning

We use SemVer for versioning, implemented through bumpversion. For the versions available, see the tags on this repository.

Authors

• Will Davey - main developer

License

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