/paper2code

Primary LanguagePython

paper2code

*Warning Use at your own risk. paper2code uses Restricted Python to define a subset of Python that is allowed to be executed. However, there could still be potential vulnerabilities.

Prep

  1. Obtain Google Application Key and download to a local path
  2. Rename file to p2p.json
  3. Export path with keys (needs to be done every shell session, consider adding to ~/.bashrc): export GOOGLE_APPLICATION_CREDENTIALS="path/to/p2p.json"

Building and testing using Google Cloud Run

  • Build: gcloud builds submit --tag gcr.io/[PROJECT-ID]/[IMAGE]
  • Local testing:
PORT=8080 && docker run -p 8080:${PORT} \
--name my-p2p \
-e PORT=${PORT} \
-e GOOGLE_APPLICATION_CREDENTIALS=/tmp/keys/p2p.json \
-v $GOOGLE_APPLICATION_CREDENTIALS:/tmp/keys/p2p.json:ro \
-v </path/to/src>:/app \
gcr.io/[PROJECT-ID]/[IMAGE]

Add package

  1. Have container up and running
  2. Access docker container shell: docker exec -it my-p2p sh
  3. Exit app folder cd ..
  4. Install package: pipenv install <package-name> *Note use --dev for dev packages
  5. Copy Pipfile and Pipfile.lock to local directory: docker cp my-p2p:/Pipfile . && docker cp my-p2p:/Pipfile.lock .

Linting

docker exec -it my-p2p pipenv run pylint app

  • Add lint pre-commit hook: cp pre-commit.sh .git/hooks/pre-commit. Can also use a symlink here

Deployment to Google Cloud Run

  1. Ensure you are using Dockerfile.prod
  2. Generate requirements.txt using pipenv (need to do this within local container): pipenv lock -r > requirements.txt
  3. Deploy: gcloud builds submit --tag gcr.io/[PROJECT-ID]/[IMAGE]
  4. Ensure Cloud Run service is selected to newest image.