Skip to content

A simple Python library developed by Johnson & Johnson's Advanced Analytics team that combines existing libraries for common Natural Language Processing tasks.

License

Notifications You must be signed in to change notification settings

johnsonandjohnson/nlprov

Repository files navigation

NLProv: Natural Language Processing Tool

NLProv is a Python library developed by Johnson & Johnson's Advanced Analytics team that combines existing libraries for common Natural Language Processing tasks. It combines several existing open-source libraries such as pandas, spaCy, and scikit-learn to make a pipeline that is ready to process text data. There are many user defined parameters depending on your type of project such as the ability to choose stemming or lemmatization. Or, you might want to define explicitly what to substitute with NaN text fields. Overall, it is a way to get you started in your NLP task, no matter what you need.

A tutorial on how to use this package can be found here.

Installation Instructions

  • Using pip with Python version 3.7 or higher:
    pip install nlprov
  • For more information on installing packages using pip, click here.

Contributing

  • To help develop this package, you'll need to install a conda virtual environment defined by our dev_environment.yml file using the below command.

    conda env create -f dev_environment.yml
    • Then, just activate the environment when attempting to develop or run tests using the below command.

      conda activate nlp_env
    • When you're all done developing or testing, just deactivate the environment with the below command.

      conda deactivate

Docker Configuration

  • This codebase is dockerized to build, run all of the unit tests using pytest, and perform pip packaging.
    • In order to run the docker container, ensure you have Docker installed and running on your local machine.
    • To start the docker container locally, simply navigate to the root of the project directory and type:
    docker-compose up --build
    • Note: docker-compose is included in the Docker desktop installation link above for MacOS and Windows based systems. If you have issues executing docker-compose, Navigate Here to ensure docker-compose is supported on your system.
    • A Notey-er note: You can use docker-compose up --build during development to quickly run the tests after code changes without setting up/running a local conda environment.

GitHub Action CI Configuration

  • Every commit to this repository will trigger a build in GitHub Actions following the .github/workflows/pythonapp.yml located in the root of this project.
    • GitHub Actions is used to build and lint the NLProv package, run the tests, and perform pip packaging.
    • If the environment name or version changes, the pythonapp.yml file will need to be updated to follow the new pattern.

Our Workflow

Upcoming Features

Here is a roadmap of features to be implemented in this package. If you have any ideas for additional features, please let us know!

  • Preprocessing
    • Ability to use custom stop words
    • Incorporation of bi-grams
    • Ability for user to chose which langauge detection package to use
  • Vectorization
    • spaCy pre-trained models
    • spaCy custom models
  • Similarity Metrics
    • Additional pairwise distances
    • Levenshtein Distance
    • Word Mover's Distance
  • Visualizations
    • TF-IDF
    • Jaccard

About

A simple Python library developed by Johnson & Johnson's Advanced Analytics team that combines existing libraries for common Natural Language Processing tasks.

Resources

License

Stars

Watchers

Forks

Packages

No packages published