scikit-fingerprints 2.0.0

This is a major release to signify how scikit-fingerprints grew in scope and functionality since 1.0.0. It now covers not only molecular fingerprints, but also molecular filters, datasets, distances & similarities, applicability domain checkers, and much more. While this v2.0.0 does not bring significant breaking changes overall, it does change many smaller functionalities, in particular MAP4 fingerprint. Further, we want to underscore the wider scope of the library now, with many planned extensions down the line in 2.x versions.

New features

• ensure_mols now supports InChI strings, which also means all fingerprints etc. safely take InChI input in addition to SMILES and Mol objects
• TDC clearance_hepatocyte_az dataset removed due to low data quality
• First clustering method: MaxMin clustering
• Filtering only valid sequences in LRGB
• Multithreading for MolFromSDFTransformer for new RDKit versions

Breaking changes

• MAP4 fingerprint debugged, results will strongly differ from previous ones
• random_state removed from classes that didn't use it

Fixes, small modifications, internal changes

• A lot of small fixes, with code and docs style unification
• Faster tests and CI
• Many new benchmarks

New Contributors

• @LiudengZhang made their first contribution in #524
• @florian-huber made their first contribution in #528
• @abbadonaz made their first contribution in #512
• @FloudMe77 made their first contribution in #526
• @deepakdamera made their first contribution in #532

Full Changelog: v1.18.0...v2.0.0