- Linux or MacOS
- python 3.10 installation (recommend using pyenv)
Set up virtual environment with make install, and activate it with source .venv/bin/activate
We use poetry for virtual env creation and dependency management.
We use black and ruff for python code formatting and linting.
We use mypy for python type checking.
We use pytest for unit and integration tests.
Manually download the reference file and save it to a directory of this repo named data/. The file is ~100MB.
You can take a subset of this data or run the whole thing.
Can use scripts/generate_predictions.py, or just use the ersilia CLI.
Requirements for the model hub (ersilia CLI)
- git lfs
- docker
python scripts/generate_predictions.py <path_to_input> <path_of_output>