Release Notes

Version 0.5.0 (2018-11-18)

New Features:

  • Adds unfolding matrix to output of iterative_unfold (See PR #88).
  • Adds CallbackList container for storing a collection of Callback objects (See PR #101).

Changes:

  • Adds a check to Mixer to ensure response matrices are two-dimensional. An error is raised if a non-2D response matrix is input (See PR #90).
  • Removes unnecessary NumPy array copies in Mixer and CovarianceMatrix object (See PR #99).
  • Vectorized some operations in Mixer and CovarianceMatrix objects (See PR #102).

Bug Fixes:

  • Fixed bug in Bayes factor test statistic calc method (See PR #93).

Documentation:

  • Add nbsphinx prolog to examples with link to interactive Binder notebook (See PR #96).
  • Corrected the accepted ts parameters in the iterative_unfold docstring (See PR #91).

Version 0.4 (2018-06-03)

Changes:

  • Adds check to iterative_unfold for negative values in input distributions. An error is raised if negative efficiencies, responses, or observed data distributions are input. (See PR #73)
  • Adds several documentation updates in response to JOSS reviewer comments (See PR #83, PR #82, PR #77, PR #76, PR #71, PR #70, and PR #69)

Bug Fixes:

  • Fixes environment used for running the “Tutorial” and “Advanced Techniques” example notebooks on binder. (See PR #68)

Version 0.3 (2018-05-15)

New Features:

  • Adds cause bin group support to SplineRegularizer for independent regularization. (See PR #47)
  • Adds cov_type parameter to the iterative_unfold function. This adds the option to choose between using the Multinomial or Poisson form of the covariance matrix. (See PR #50)
  • Adds check for negative values in prior input to iterative_unfold. (See PR #63)

Changes:

  • Makes data, response, efficiencies, and associated uncertainties parameters in iterative_unfold keyword arguments. Now they can be input in any order. (See PR #46)
  • Changes default prior in iterative_unfold from Jeffreys prior to a uniform prior. (See PR #52)

Bug Fixes:

  • Fixes bug in calculation of Jeffreys prior. (See PR #52)

Version 0.2.2 (2018-05-04)

Changes:

  • Adds checks for optional PyTables testing dependency. (See PR #43)

Version 0.2.1 (2018-05-04)

Changes:

  • Adds MANIFEST.in file to package additional files with source distribution.

Version 0.2 (2018-05-04)

New Features:

  • Added SplineRegularizer callback to perform smoothing at each unfolding iteration. (See PR #27)
  • Added Logger callback to write test statistic information to stdout. (See PR #18)
  • Added contributing guide to the documentation page. (See PR #31)
  • Added setup_prior convenience function to help validate user prior input (See PR #35)

Changes:

  • Removed datadist.py module. (See PR #19)
  • Vectorized covmatrix.py module. (See PR #24)
  • Removed covmatrix and loadstats modules. (See PR #38)

Bug Fixes:

  • Fixed typos in the online documentation (See PR #26)

Version 0.1 (2018-04-05)

New Features:

  • Initial release that includes iterative_unfold function.