Skip to content

Latest commit

 

History

History
110 lines (90 loc) · 5.03 KB

CHANGELOG.rst

File metadata and controls

110 lines (90 loc) · 5.03 KB
Raw

Changelog

v1.9.1 (October 24, 2020)

  • Fixed #79 - Chainstats display now auto updates formatting for really large or small numbers.
  • Fixed #82 - Initial value check routine in Parallel MCMC now correctly removes extra columns.

v1.9.0 (July 19, 2019)

  • Added feature to results structure to store all parameters names, regardless of whether or not they are included in sampling.
  • Added mcmcplot package to requirements. MCMCPlotting module is noted as deprecated.
  • Added new module for uncertainty propagation. Aims to provide more flexible API for user to plot different combinations of credible and prediction intervals.
  • Added a plotting routine so that you can plot a 2-D interval in 3-D space.

v1.8.0 (June 28, 2019)

  • Added acceptance rate display feature when calling chain statistics
  • User can specify skip or maxpoints in pairwise correlation and chain panel plots in order to thin chain.
  • User can request item definitions when calling chainstats.
  • Resolved #51 by adding method to check if limits are incompatible. User is provided with more descriptive error message.
  • Updated documentation for Bayesian components.
  • Added saving routines to accomodate post-processing. A lighter version of the results dictionary is saved to json when log routines are used for chains. This should address #55.
  • Option to return KDE objects when using plot_density_panel. User can then evaluate pdf directly. This should address #66.

v1.7.1 (May 3, 2019)

  • Fixed bug #41.
  • Updated how version attribute is defined in package and referenced in setup.py and documentation.

v1.7.0 (April 5, 2019)

  • Added custom sampler feature. Interface requires creation of a class, but allows for inclusion of things like Gibbs samplers. User can explicitly pass in features to the sum-of-squares function.
  • Added check routines for reading in parallel MCMC results to ensure only directories with name 'chain_' are queried.
  • All files were updated to comply with formatting standards using flake8.
  • Added coverage report check to ensure minimal testing requirements are met.
  • Added original covariance matrix to results structure for reference.

v1.6.0 (August 31, 2018)

  • Added optimal handling of prior function used numpy array multiplication.
  • Fixed bug with saving results to json file - output directory now automatically made.
  • Updated PI plotting to allow any valid matplotlib input.
  • Added Gelman-Rubin diagnostics to ChainStatistics module, with display feature.
  • Added numpy error settings option to MCMC initialization.
  • Setup no adaptation feature. Can sample parameters without adapting them.
  • Added restart routine to ParallelMCMC. Can restart from json files.
  • Added parallel chain organization routines to the chain subpackage.

v1.5.0 (July 27, 2018)

  • Final release of 1.5.0.
  • First release with complete test coverage.
  • Updated analysis options and diagnostics.
  • Improved workflow on remote OS.
  • Added link to Zenodo DOI.

v1.4.0 (July 11, 2018)

  • Added extensive unit tests.

v1.3.1 (April 20, 2018)

  • Fixed sample feature (allows for fixed variables).
  • automatic naming (and expansion) for plot labels.
  • added covariance matrix to log file output.

v1.3.0 (April 3, 2018)

  • All features updated to work with Python 3.6
  • This was the last version developed using Python 2.7. It may still work on Python 2.7, but it will not be included as a part of testing.

v1.2.0 (April 3, 2018)

  • Added ellipse contour routine for pairwise plots.
  • Added progress bar to prediction interval generation (as it can be time consuming).
  • Added log file saving so you can periodically dump the latest set of samples to file.

v1.1.1 (February 26, 2018)

  • Fixed bug associated with plotting credible intervals without prediction intervals.

v1.1.0 (February 19, 2018)

  • Added feature for generating and plotting prediction/credible intervals.

v1.0.0 (February 15, 2018)

  • First official release.
  • Incorporated class structures into entire architecture.