Changelog¶
Version 0.2.4¶
- Fix #2: a regression when using the polars backend: missing uncertainty data is handled properly again
- Fix #3: Relaxed version pin of numpy from
>=2.0to>=1.20.3to maintain support for older environments - Support Python 3.9 again, reverting type hints etc. to what is supported by this version.
- Add support for Python 3.14
- Increase default cache time-to-live to two weeks
- Update test suite
- Minor refactoring and doc updates
Version 0.2.3¶
- Fix issue #1: update the set of queried parameters so the ASD replies with data again; likely some small internal changes to the ASD were implemented.
- Update unit tests and test cleanup
polarsis now a mandatory dependency for unit tests- Cleanup polars expressions for parsing ASD data
- Added support for new handling of strings as datatype in pandas 3.0; supporting both pandas 2 and 3
- Test matrix tests against both versions of pandas.
Version 0.2.*¶
Public release version
- Release to PyPI
- Update documentation only on new release
Version 0.1.*¶
Initial development release.
- Support fetching and caching from the NIST ASD
- Cache-TTL defaults to 1 week for now
- Caching hash is computed on all query parameters by default (
strict=True)- Can be cached on only elements, ionization and wavelength range, with
strict= False
- Can be cached on only elements, ionization and wavelength range, with
- Enable retrieving all cached species/queries
- Give user the option to use
polarsinstead ofpandasas Dataframe backend if installed - Start writing documentation
- Auto-generated code reference included
- Example usage using a jupyter notebook included
- Add citing information
- Add test using
pytest - Also checks for consistentcy of results between
pandasandpolarsbackends. - Set up CI/CD
- Documentation automatic generation and deploy of latest version
- Automated run of test suite
- Add bibliography lookups
- Archive releases to Zenodo
- Renamed project to
ASDCache