Thanks for taking the time to contribute! 👍
In general, we would like to
- keep the repository as small as possible but still produce (more or less) physically meaningful results (e.g., roughly correct band structures). Of course, for testing datasets, no such requirement is needed.
- keep the original DFT input files and scripts for generating the datasets, to ensure reproducibility. This also allows us to regenerate the datasets if needed.
Pull requests are the best way to propose changes to the codebase. We actively welcome your pull requests:
- Fork the repo and create your branch from
main
. - Add dataset(s), ideally small (in terms of file size) but still gives
physical results. The
prefix
of the files (i.e.prefix.amn
/prefix.mmn
) should be the same as the folder name, to allow easy loading of the dataset byWannier.jl
(the functionWannier.Datasets.load_dataset
) - Make sure the original input files are committed in a
generator
subfolder, so we can always reproduce the results on different machines. - (Optional) put useful reference files in a
reference
subfolder - Always
git rebase
your branch on top of the latestmain
branch, so that we can merge your PR without ugly 😟 merge commits. Otherwise, we will rebase explicitly your branch when merging. - Create that pull request!