This would be useful for:
- providing working examples for users to look at
- checking that SW dependency upgrades don't break the tool
- if/when bugs are found, new tests can be added to prevent regressions
Once these pytests are in, we can hook up GitHub actions to run them on PRs.