pp_vectorizer

This package provides a vectorizer (inherits from sklearn.feature_extraction.text.TfidfVectorizer) that uses a PoolParty instance to add a semantic layer to the extraction. Namely, PoolParty enriches the resulting vector with extracted concepts (contained in the specified PoolParty project) and with additional information about those concepts.

To use:

1. Clone

2. pip3 install -r requirements.txt

3. Use .env file to create a file called .localenv containing your credentials and configurations

   • PP_USER, PP_PASSWORD and PP_SERVER refer to your PoolParty credentials
   • PP_PID Is the Project ID of the PoolParty project you wish to use for extraction
   • CACHE_PATH Is a path where you want your local extraction temporary cache saved. /tmp would work
   • STORE_PATH The extraction cache also has a long-term version, it will be saved here.
   • DOCS_PATH Is a path where you have directories, each containing documents. Used for tests

4. Set your configuration by: export $(cat .localenv | xargs)

5. See examples directory and play around