Changelog¶

Version 0.6.1 (April 2019)¶

Genetic algorithm feature selection can parallelize over population within each generation.
Default fingerprinter function sets accessible using catlearn.fingerprint.setup.default_fingerprinters
New surrogate model utility
New utility for evaluating cutoff radii for connectivity based fingerprinting.
default_catlearn_radius improved.

AtoML renamed to CatLearn and moved to Github.
Adsorbate fingerprinting again parallelizable.
Adsorbate fingerprinting use atoms.tags to get layers if present.
Adsorbate fingerprinting relies on connectivity matrix before neighborlist.
New bond-electronegativity centered fingerprints for adsorbates.
Fixed a bug that caused the negative log marginal likelihood to be attached to the gp class.
Small speed improvement for initialize and updates to GaussianProcess.

Added autogen_info function for list of atoms objects representing adsorbates.
- This can auto-generate all atomic group information and attach it to atoms.info.
- Parallelized fingerprinting is not yet supported for output from autogen_info.
Added database_to_list for import of atoms objects from ase.db with formatted metadata.
Added function to translate a connection matrix to a formatted neighborlist dict.
periodic_table_data.list_mendeleev_params now returns a numpy array.
Magpie api added, allows for Voronoi and prototype feature generation.
A genetic algorithm added for feature optimization.
Parallelism updated to be compatable with Python2.
Added in better neighborlist generation.
- Updated wrapper for ase neighborlist.
- Updated CatLearn neighborlist generator.
- Defaults cutoffs changed to atomic_radius plus a relative tolerance.
Added basic NetworkX api.
Added some general functions to clean data and build a GP.
Added a test for dependencies. Will raise a warning in the CI if things get out of date.
Added a custom docker image for the tests. This is compiled in the setup/ directory in root.
Modified uncertainty output. The user can ask for the uncertainty with and without adding noise parameter (regularization).
Clean up some bits of code, fix some bugs.

Added a parallel version of the greedy feature selection. Python3 only!
Updated the k-fold cross-validation function to handle features and targets explicitly.
Added some basic read/write functionality to the k-fold CV.
A number of minor bugs have been fixed.

Update the fingerprint generator functions so there is now a FeatureGenerator class that wraps round all type specific generators.
Feature generation can now be performed in parallel, setting nprocs variable in the FeatureGenerator class. Python3 only!
Add better handling when passing variable length/composition data objects to the feature generators.
More acquisition functions added.
Penalty functions added.
Started adding a general api for ASE.
Added some more test and changed the way test are called/handled.
A number of minor bugs have been fixed.

Update functions to compile features allowing for variable length of atoms objects.
Added some tutorials for hierarchy cross-validation and prediction on organic molecules.

Gradients added to hyperparameter optimization.
More features added to the adsorbate fingerprint generator.
Acquisition function structure updated. Added new functions.
Add some standardized input/output functions to save and load models.
The kernel setup has been made more modular.
Better test coverage, the tests have also been optimized for speed.
Better CI configuration. The new method is much faster and more flexible.
Added Dockerfile and appropriate documentation in the README and CONTRIBUTING guidelines.
A number of minor bugs have been fixed.

The first stable version of the code base!
For those that used the precious development version, there are many big changes in the way the code is structured. Most scripts will need to be rewritten.
A number of minor bugs have been fixed.