v1.2.2

What's Changed

• add config file to readthedocs by @luigibonati in #167

Full Changelog: v1.2.1...v1.2.2

v1.2.1

What's Changed

• Remove torch-scatter dependency by @luigibonati in #166

Full Changelog: v1.2.0...v1.2.1

v1.2.0

This version includes several bugfixes, improvements as well as new features:

• Interpretability module
• Committor CV via Kolmogorov’s variational principle

What's Changed

• Fix DeepLDA regularization by @luigibonati in #114
• Fix bug in lr scheduler configure by @EnricoTrizio in #116
• Fix device issue checkpoint test by @EnricoTrizio in #117
• Bug fix in nn.feedforward by @EnricoTrizio in #119
• Remove scipy version requirement from pyproject by @andrrizzi in #121
• Fix issues in CI by @EnricoTrizio in #129
• Extension of transform module by @EnricoTrizio in #124
• Add committor code by @EnricoTrizio in #132
• Bug fix transform and committor by @EnricoTrizio in #133
• Make last layer activation of committor an explicit block by @EnricoTrizio in #135
• Interpretability module: sparse linear models via LASSO by @luigibonati in #120
• Register TDA target parameters as loss module buffers by @recisic in #142
• Improvement committor module by @EnricoTrizio in #145
• Test PyPI version via CI by @luigibonati in #146
• Add SequentialTransform by @EnricoTrizio in #147
• Numpy 2 update temporary fixes by @EnricoTrizio in #152
• Committor updates by @EnricoTrizio in #151
• Fixed deltaF bug by @EnricoTrizio in #156
• Doc minor fixes by @EnricoTrizio in #157
• Improve transform tutorial coverage by @EnricoTrizio in #158
• Committor tutorials update by @EnricoTrizio in #159
• Fixed docs requirements by @EnricoTrizio in #162
• Fixed markdown headers by @EnricoTrizio in #163
• Fixed colab links by @EnricoTrizio in #164
• Add cutoff on distances lists by @EnricoTrizio in #165

New Contributors

• @recisic made their first contribution in #142

Full Changelog: v1.1.0...v1.2.0

v1.1.1

Bug fixes to v1.1 (see #144)

v1.1.0

Bug fixes + sensitivity analysis

What's Changed

• Fix mamba commands migration issue by @EnricoTrizio in #98
• Fix bug in CI badge by @EnricoTrizio in #99
• general split_dataset functions by @EnricoTrizio in #100
• Add installation info for main page by @EnricoTrizio in #102
• change register_cmap to register by @luigibonati in #107
• remove version check on scipy by @luigibonati in #109
• Add sensitivity analysis by @luigibonati in #108
• add sensitivity tutorial to doc by @luigibonati in #112
• Fix bug unrepresented state. by @andrrizzi in #110
• Fix #103: CVs cannot be loaded from checkpoint by @andrrizzi in #113
• Add lr_scheduler options by @EnricoTrizio in #111

Full Changelog: v1.0.1...v1.1.0

v1.0.1

Bug fixes and doc improvements

What's Changed

• Fixed path to state file timelagged experiment by @EnricoTrizio in #82
• added scipy version requirement by @luigibonati in #85
• Update doc and readme by @luigibonati in #86
• fix citation by @luigibonati in #87
• specify dependency version requirement by @luigibonati in #89
• Minor fixes by @EnricoTrizio in #92
• Allow training without validation dataset by @luigibonati in #88
• Fix doc by @luigibonati in #93

Full Changelog: v1.0.0...v1.0.1

v1.0.0

First stable release of mlcolvar package. Most of the changes with respect to 0.3.0 are related to documentation and notebooks illustrating tutorials and examples on how to use and expand the library, together with small fixes and enhancements.

What's Changed

• Support for multiple datasets of different size by @andrrizzi in #59
• NormalizationInverse class by @EnricoTrizio in #61
• Codecov_debug by @EnricoTrizio in #64
• Codecov_improve by @EnricoTrizio in #65
• Include notebook testing to codecov by @EnricoTrizio in #70
• Fix AutoEncoder calling pre/postprocessing during training by @andrrizzi in #69
• Colab interface for notebooks by @EnricoTrizio in #71
• Tutorials_fix by @EnricoTrizio in #62
• Moved notebook tests only to Ubuntu by @EnricoTrizio in #76
• Chingolin example with TPI-Deep-TDA + minor fix in create_dataset_from_file by @EnricoTrizio in #77
• Update documentation with contributions and references by @luigibonati in #78
• Add section for examples notebooks by @luigibonati in #79
• Prepare for pypi release by @luigibonati in #80
• fix dependencies in pyproject by @luigibonati in #81

Full Changelog: v0.3.0...v1.0.0

v0.3.0 (mlcolvar)

First mlcolvar release

The library went through a complete refactoring (including being renamed from mlcvs to mlcolvar) with new structure and APIs.

Main changes

• Pytorch lightning support
• Modular CV structure
• Dict-like datasets
• Many new CVs methods

Pull requests

Only the recent changes are shown in this list:

• First implementation of a Variation Autoencoder CV by @andrrizzi in #27
• Change base cv attributes by @luigibonati in #32
• change MSE_loss signature from diff to input, target by @luigibonati in #37
• add option in ae/vae to compare output of the decoder with a different target by @luigibonati in #38
• Improve load_dataframe by @luigibonati in #41
• Implement Loss classes, some PEP8 fixes by @andrrizzi in #42
• Switch from functional to class losses by @andrrizzi in #43
• Add PCA method by @luigibonati in #44
• Support for multiple datasets by @andrrizzi in #45
• Added reduced_rank to TICA by @pietronvll in #36
• Multi-task CV and various fixes by @andrrizzi in #46
• rename library v2 by @luigibonati in #49
• Fix device loss by @luigibonati in #50
• Rename data objects by @luigibonati in #51
• Support for task-specific layers by @andrrizzi in #53
• Merge lightning into main by @luigibonati in #55
• Add paper experiments by @luigibonati in #56
• Code fixes by @luigibonati in #58

New Contributors

• @andrrizzi made their first contribution in #27
• @pietronvll made their first contribution in #36

Full Changelog: v0.2.2...v0.3.0

v0.2.2 (mlcvs)

What's Changed

• Fix bug in the logging for the tutorials (see #25)

Note that this is the last version that does not use the PyTorch Lightning high-level framework, and in which the module is named mlcvs.

Full Changelog: v0.2.1...v0.2.2

v0.2.1

What's Changed

• Prune the tree: move general functions to base class NNCV by @luigibonati in #18
• add autoencoder cv by @luigibonati in #22
• Add CodeQL workflow for GitHub code scanning by @lgtm-com in #21

New Contributors

• @lgtm-com made their first contribution in #21

Full Changelog: v0.2.0...v0.2.1