v2.4.3

What's Changed

• adding rho jump proposals for free spectra by @astrolamb in #189
• update version to increment Zenodo by @AaronDJohnson in #210
• Bugfix draw_from_gwb_log_uniform_distribution by @Wang-weiYu in #221
• Bump scikit-learn from 0.24 to 1.0.1 by @dependabot in #224

New Contributors

• @Wang-weiYu made their first contribution in #221
• @dependabot made their first contribution in #224

Full Changelog: v2.4.2...v2.4.3

Update to support new versions of numpy

What's Changed

• Add Python 3.10, Remove 3.6 from CI by @AaronDJohnson in #197
• Attack of the clones: Empirical distribution fix for duplicate setup_sampler by @AaronDJohnson in #183
• Cover more priors with empirical distribution extension function by @AaronDJohnson in #184
• remove print statement by @AaronDJohnson in #198
• added ptmcmc version info by @astrolamb in #192
• Fix memory issues in Fp statistic by @AaronDJohnson in #188
• reformat .flake8 file by @AaronDJohnson in #202
• Fix HyperModel to work with newer versions of numpy.
• Allow manual specification of Tspan for models 1/2/3 in model.py by @vallis in #172
• increment version to 2.4.2 by @AaronDJohnson in #203
• update CI versions by @AaronDJohnson in #204

Full Changelog: v2.4.1...v2.4.2

Wideband Analysis Update

Release includes a minor change to enable the new WidebandTimingModel noise definitions

What's Changed

• New keyword arguments have been added to the models to support the timing group's efforts in noise analyses

NG15yr Analysis, EQUAD Redefinition and Empirical Distribution Tweak

This version makes a major change to the way that the EQUAD parameter is modeled. Enterprise has officially broken backwards compatibility, renaming EQUAD parameters to either:

• t2equad to follow the timing package (Tempo, Tempo2, Pint) definition of EQUAD, i.e. sigma^2_T = EFAC^2*(sigma^2_TOA + EQUAD^2)
• tnequad to follow the TempoNest definition of EQUAD, i.e. sigma^2_T = EFAC^2*(sigma^2_TOA) + EQUAD^2

In addition, a number of fixes to the empirical distribution submodule were made as well as the full complement of solar wind models added. See below for all of the details.

What's Changed

• Fix empirical distributions and hardcoded gwb prior draws by @AaronDJohnson in #156
• fixed simulated psr bugs, added params to specify spectrum priors by @astrolamb in #158
• More changes to empirical distributions by @AaronDJohnson in #157
• Use svd by @AaronDJohnson in #160
• KDE empirical distributions by @meyers-academic in #159
• fix white noise in the models for non-NANOGrav PTA analysis by @siyuan-chen in #164
• add fix for timing model by @siyuan-chen in #162
• add chi square functionality and tests by @Hazboun6 in #166
• Solar wind models by @Hazboun6 in #167
• New Equad by @Hazboun6 in #169
• added kwargs for logL, logp in sampler by @astrolamb in #170
• Version2.4.0 by @Hazboun6 in #171

New Contributors

• @meyers-academic made their first contribution in #159

Full Changelog: v2.3.4...v2.4.0

NANOGrav 15-year Analysis Feature Freeze with Phase Shift Seed Fix

This keeps the same feature freeze as previous minor versions while fixing a bug in the phase shift seed functionality for the optimal statistic. We also have started adding documentation.

What's Changed

• Documentation by @Hazboun6 in #146
• Fixes to Autodocs by @Hazboun6 in #148
• Updating Dev by @Hazboun6 in #149
• Finish 1st Round of Documentation by @Hazboun6 in #150
• Revise caching so that optimal statistic works with variable pseed by @vallis in #151
• Added test for OS phase shift by @Hazboun6 in #152

Full Changelog: v2.3.3...v2.3.4

NANOGrav 15-year Analysis Feature Freeze with Package Data Included

This release includes the feature freeze code planned for use in the NANOGrav 15-year GW analyses.
This also includes a bug fix to include needed data files in the compressed package files.

What's Changed

• make changes to include package data by @Hazboun6 in #145

Full Changelog: v2.3.2...v2.3.3

NANOGrav 15-year Analysis Feature Freeze with Runtime Info Bug Fix

This release includes the feature freeze code planned for use in the NANOGrav 15-year GW analyses.
This includes a bug fix to now write the model information from all models in a HyperModel analysis.

NANOGrav 15-year Analysis Feature Freeze with Bug fixes and Linting

Merge pull request #140 from Hazboun6/linting

Merging these changes which fixed @paulthebaker concerns.

NANOGrav 15yr Analysis Feature Freeze

This release encompasses a number of new analysis tools, including alternative polarization searches, burst with memory searches, and code improvements for the optimal statistic and sampling code.

This version also constitutes the feature freeze for NANOGrav 15 year dataset analyses.

Software Version with Outlier Analysis

An outlier analysis using enterprise to detect time-of-arrival outliers was included in the code, but is now being removed into a separate module, see enterprise_outlier. This tag keep an archive of this version since it has been used extensively in various timing analyses.