Thanks for contributing to this project.
The goal of this repository is to keep the codebase production-style, easy to understand, and easy to maintain, so small, clean improvements are preferred over large mixed changes.
-
Create and activate a virtual environment:
python3 -m venv .venv source .venv/bin/activate -
Install the project with development dependencies:
pip install '.[dev]' -
Place the M5 source files in
data/raw/m5if you want to run the full real-data workflow.
Run the full pipeline:
retail-forecast run-allStart the API:
retail-forecast serve-apiStart the dashboard:
streamlit run src/retail_forecasting/dashboard/app.pyRun tests:
pytest- Keep changes focused and easy to review.
- Prefer separate commits for docs, tooling, and feature work.
- Update the README when the project behavior or setup changes.
- Do not commit raw M5 CSV files, local
.envfiles, or generated warehouse artifacts. - Before pushing, run
pytestand make sure the app still starts locally.
Examples:
docs: clarify local setup and configurationci: add github actions test workflowfeat: improve forecast explorer filteringfix: correct reorder point calculation edge case