dependency-management.md

Dependency Management

We use pip-tools to manage our python dependencies and create reproducible environments.

Tools

• pip-tools for dependency management

Workflow

1. Our requirements are tracked in ../backend/requirements
2. We use pip-compile to generate a requirements.txt file with hashes
3. We track both in version control
4. We use requirements.txt to install dependencies locally as well as in deployed environments.
5. To avoid architecture-related complications, it's recommended to first exec into a Docker container before running the compilation:

• docker compose exec web /bin/bash
• python -m piptools compile --allow-unsafe --generate-hashes --output-file=requirements.txt ./requirements/requirements.in

Installing dependencies

If you are running the Docker containers, they will automatically install.

If you are running locally, you can run:

cd backend
pip install -r requirements.txt

for the dependencies for the project. Developer dependencies (like linters) can be installed with:

cd backend
pip install -r dev-requirements.txt