Skip to content

Releases: llnl/trata

1.2.1 Release

Choose a tag to compare

@olson59 olson59 released this 16 Oct 21:15

Description

This is an update to an existing sampler providing different but more robust results.

Changed

The QuasiRandomNumberSampler has been updated to use scipy's implementation for Sobol and Halton sequences.

Benefits include:

  • Continue the sequence from a certain point with the sequence_offset option.
  • A scramble option for better randomization
  • The directional numbers for generating a Sobol sequence are updated to allow for higher dimensions. Joe and Kuo 2008 This gives different but still valid results compared to the previous version.

1.2.0 Release

Choose a tag to compare

@olson59 olson59 released this 17 Sep 04:42

Description

Added

  • BestCandidateSampler: A new option in trata.adaptive_sampler that chooses new samples that add most to diversity of existing samples.
  • All adaptive samples can be passed a random seed for reproducability.
  • The sampling documentation notebook has been updated to include adaptive_sampler examples.
  • The ProbabilityDensityFunctionSampler will automatically scale values based on ranges passed with the "box" parameter.

Changed

  • Adaptive samplers use the same variable names as the rest of the modules. Only "values" for the inputs/X, and "output" for response/Y.

Bug fixes

  • Updated some indexing and variable assignment methods to be compatible with numpy > 2.0
  • One-at-a-time sampler has input validation and avoids duplicating default values.
  • Other fixes and updates to the testing suite

Release 1.1.1

Choose a tag to compare

@olson59 olson59 released this 31 Mar 21:42

1.1.1 Release

Description

This release includes a minor bug fix.

Bug fixes

  • In the sampler module, checking for empty lists was fixed.

Release 1.1.0

Choose a tag to compare

@olson59 olson59 released this 12 Aug 22:19

1.1.0 Release

Description

This release introduces a new kosh_sampler module

New in this release

  • The kosh_sampler module wraps the adaptive sampling functions in a Kosh operator. Users with existing Kosh datasets can easily find the next best set of samples based on their model's error or sensitivity.
  • A Sobol index sampler has been added to the sampler module. It creates samples to be used in
    IBIS's Sobol indices function in the sensitivity module.

Release 1.0.3

Choose a tag to compare

@olson59 olson59 released this 05 Jun 00:23

1.0.3

Release Notes

1.0.3 Release

This is a minor release with a few bug fixes.

Bug fixes

  • fixed adaptive sampling tests to be more robust to numpy versions.
  • fixed hanging problem in composite samples.
  • A few other minor changes related to testing

1.0.2 Release

Description

This release is a minor release with a few bux fixes and new features. We encourage users to upgrade.

New in this release

  • Added **kwargs to Morris sampler to catch extra arguments.

Improvements

  • Added code of conduct contributing documents

Bug fixes

  • No bug fixes

1.0.1 Release

Description

This release is a minor release with a few bux fixes and new features. We encourage users to upgrade.

Release 1.0.2

Choose a tag to compare

@doutriaux1 doutriaux1 released this 23 Feb 20:34
8eac165

1.0.2

Release Notes

1.0.2 Release

Description

This release is a minor release with a few bux fixes and new features. We encourage users to upgrade.

New in this release

  • Added **kwargs to Morris sampler to catch extra arguments.

Improvements

  • Added code of conduct contributing documents

Bug fixes

  • No bug fixes

1.0.1 Release

Description

This release is a minor release with a few bux fixes and new features. We encourage users to upgrade.