Powered by FastAPI + pipreqs. Specifically, FastAPI runs the API server for receiving requests and pipreqs does the code requirements assessing.

Basic API to power bots / helpers to rid the world of Python projects without correct requirements!

Attempts to shallow git clone a given repo then run pipreqs on the codebase. Returns the resulting requirements.txt contents!

NOTE: Caches responses for a given url for 5 minutes or until it is evicted by LRU policy

• Use the pyscript frontend on github pages
• Use the Streamlit frontend 🎈
• Use the fastapi generated swagger docs
• Use curl or other http client

curl "https://pipreqs-api.herokuapp.com/pipreqs?code_url=https://github.com/gerardrbentley/pipreqs-api"

Response

cachetools==5.0.0
fastapi==0.75.2
httpx==0.22.0
pydantic==1.9.0
pytest==7.1.2
streamlit==1.8.1
uvicorn==0.17.6

• 🧪 Generation Options:
  • Strict / Unpinned / Compatible / Greater Than options
  • Alternate PyPi server
  • Use requirements data to make conda environment.yml or pyproject.toml
• 🤖 API Options:
  • Caching
  • Cache Busting
  • Auto open Github Pull Request with updated requirements
  • Allow partial repo / branch urls instead of full git url
  • Safe way to allow zip fetch / analysis?
• 📺 Frontend:
  • Streamlit app for requesting repo(s)
  • Deployed on separate branch streamlit_deploy to contain requirements

• Copy or Rename .env.example as .env.dev

mv .env.example .env.dev

Requires docker-compose to be installed (this comes with Docker Desktop).

make run
# Open localhost:8000/health

# Linting
make lint
# Unit Testing
make test

# Display coverage report after test (uses local python, not docker)
make coverage

For code completion / linting / developing / etc.

make install
. ./venv/bin/activate
cd backend

• Development Containerization with Docker
• Dependency installation with Pip
• Code formatting with Black and isort
• Testing with pytest