Local development
Create a conda virtual environment and activate it:
conda env create --file environment.yml
conda activate rhyme-with-ai
Install the rhyme_with_ai package and all its dependencies:
pip install --editable .
Download the weights of the models (if you get any errors, make sure these align with those specified in app/app.py)):
make download-data
Run the app:
make streamlit
Test the Docker containers by running:
make docker-build
make docker-serve
This project uses black for code formatting. To incorporate it in your version control follow the instructions below (copied from black's own readme):
Use pre-commit. Once you have it installed, add this to the .pre-commit-config.yaml in your repository:
repos:
- repo: https://github.com/ambv/black
rev: stable
hooks:
- id: black
language_version: python3.7
Then run pre-commit install and you're ready to go.
Deploy to App Engine
Follow Google's documentation to set up Custom Runtimes in the App Engine Flexible Environment. Deploy the app:
gcloud app deploy
And you're done!
Todo
- Integrate TokenWeighter in the RhymeGenerator.
- Don't block on model loading or rhyme mutations (use API?).
