v0.1.1 (2019-08-17)

Added

• resume_dyn_run parameter to allow the dyPolyChord sampling process to be saved and resumed from cached files. This parameter only allows the process to be resumed after the initial exploratory run is finished and the second dynamic run is started

Changed

• Fixed for error arising from edge case when dynamic run starts from very close to the start of the initial run (0dd8636).

v0.1.0 (2019-01-20)

Changed

• Now compatible with PolyChord Python wrapper when it is installed under both its new name ("pypolychord" and its old name "PyPolyChord").
• Updated references to dynamic nested sampling paper (now published in Statistics and Computing).
• Minor changes to MIT License (now includes "Edward Higson and contributors" and is dated "2018-Present").

v0.0.6 (2018-10-07)

Added

• Badges and BibTeX for JOSS paper.

v0.0.5 (2018-09-30)

Changed

• Improvements to documentation.
• Improvements to JOSS paper ready for publication.
• Bug fix for python2 point allocation (forced floating point not integer division).
• Improved test robustness.
• Added python2-specific requirements to setup.py.

v0.0.4 (2018-09-16)

Added

• JOSS paper.

Changed

• Minor changes and improvements to docs and error messages.

v0.0.3 (2018-09-12)

Added

• Single-source version stored in _version.py and accessible through __version__ attribute.
• More Python priors.
• Function for mapping python priors to PolyChord ini-format strings.
• Forced abort if a rank == 0 MPI process fails (to avoid hanging).
• Config file writing.
• Error handling for errors reading PolyChord .stats files.

v0.0.2 (2018-08-04)

Added:

• Duplicate logl values are now allowed (these are permitted by nestcheck >= v0.1.4).
• LogGamma likelihood to python_likelihoods.py.

Changed

• Updated tests for compatibility with nestcheck >= v0.1.4.
• Minor tidying and corrections to documentation.

v0.0.1 (2018-05-18)

Added:

• UserWarning when random seeding is used when running with MPI and more than 1 processes noting that, despite the seeding, the results will not be reproducible due to MPI.

Changed:

• fixed bugs occurring when running with MPI in python using mpi4py.

v0.0.0 (2018-05-15)

Initial release.