v0.1.13

Minor Changes:

• Mitigate pymatgen's incompatibility with numpy>=2 by setting the maximum version
• Minor updates to the testing routines and versions.

Full Changelog: v0.1.12...v0.1.13

v0.1.12

Minor Changes:

• Increase pymatgen version to one that patches their critical vulnerability in CIF-parser

Full Changelog: v0.1.11...v0.1.12

v0.1.11

Minor Updates:

• Added official Python 3.12 support to the model.
• Uses newer rpy2 to take advantage of (1) improved debugging and (2) better compatibility with newer Python versions.
• README updates

Full Changelog: v0.1.10...v0.1.11

v0.1.10

Hot Fix: Added upper limit on pymatgen version because v2023.9.25 fails to install alongside required rpy2=>3.5.0 (pymatgen/issues/3348)

v0.1.9

Minor Changes:

• Added a JSON file with an explicit machine-readable output definition in the form of a list of name-unit-definition dictionaries.

v0.1.8

Minor Changes:

• Added Model() class which initializes its own R model that may help with conflicts in model access. It should be more stable, but if initialized every time, it will be significantly slower than one initialization per package (at least in series).

v0.1.7

Minor Changes:

• Added an explicit context for Python-R object conversions that should help with (1) multithreaded works and (2) compatibility with future rpy2 versions

Full Changelog: v0.1.6...v0.1.7

v0.1.6

Major Changes

• Improved handling of the model, allowing return of both minimalistic ordered array of values (default) or dictionary of labeled values, by specifying outputType.
• Automated testing of the model runtime across Python 3.9, 3.10, 3.11 and across Linux and Mac OS. It works on Windows as well, but installation of R is more challenging and rpy2 is more likely to break on it, so we do not test nor recommend it.
• Added script-like run ability to the model (when installed in development style by cloning)

Full Changelog: v0.1.5...v0.1.6

v0.1.5

The first stable version published on PyPI and meant for use by other people :)

General Change Log:

• v0.1.5 - much, much faster prediction throughput through reorganizing R code to efficiently initialize and use ML models
• v0.1.4 - automating R dependencies installation and documentation improvements
• v0.1.0 - v.0.1.3 - improving flexibility and quality in general
• v0.0.1 - v0.1.0 - got everything to just work and reorganized the project into a Python package

Full Commit Log: https://github.com/amkrajewski/pqam-dparamhu2021/commits/v0.1.5