Get related content recommendations from Proposer, inside Composer
In the suggested-content directory is a Python file that will run a server at https:://localhost:5000 which will accept terms and phrases and find related terms.
The service is written in Python. This is because the gensim library is far more sophisticated than the equivalent in Scala or Java.
You will need nginx installed and running as well as dev-nginx to setup the correct domain for the tool.
All scripts are written using Python 3 so this is a hard dependency. You will also need pip installed in order to fetch dependencies.
It's highly recommended you use a virtual environment when using the service as this will save you a lot of pain. I'd also recommend using the virtualenvwrapper as this will make your life even easier.
The instructions below assume you are using both virtualenv and virtualenvwrapper.
cd
into the cloned repo.mkvirtualenv <name-of-virtualenv> --python=<path-to-python3>
pip install -r requirements.txt
word-2-vec-service.py
and the service will be at https://suggested-content.local.dev-gutools.co.uk