Contributing

Developer Installation¤

Clone the git repository and navigate to that directory
Install virtual environment (the below example assumes conda)
```
conda create --name equine python==3.11
conda activate equine
```
We currently support python versions >= 3.9
Install the code with the extra tests dependencies
```
pip install -e .'[tests]'
```
Activate the pre-commit hooks
```
pre-commit install
pre-commit run
```

We prefer that any contributed code be outfitted with contracts from icontract and tested with hypothesis. This combination frequently means that the tests require few (if any) actual post-checks -- if the contracts are well-written, then hypothesis can generate reasonable tests that will explore the bounds of the contracts for each method. Assuming that tests pass, then make a pull request.

Documentation¤

In the MIT-LL Responsible AI GitHub organization, we use the numpy/scipy format for docstrings.

We use mkdocs to build our documentation and autodocumented API. Most of these files can be found in the docs folder, and the dependencies to generate your own version of the documentation can be installed via:

pip install -e .'[docs]'

You can test new documentation by issuing mkdocs serve from the root directory. Once it has built, you can check out the new documentation by connecting your browser to localhost:8000.