This version is the version used for all plots in Murray, Robotham, Power (2018), and is released along with that paper. There are many changes in the code from previous versions, the result of a couple of years of sporadic work.
- New
PerObjFitclass supersedesget_fit_perobjfunction, providing more coherent fitting capabilities. - Added heaps of "real-world" examples (used in MRP paper):
- https://github/steven-murray/mrpy/docs/examples/fit_curve_against_analytic.ipynb
- https://github/steven-murray/mrpy/docs/examples/fit_simulation_suite.ipynb
- https://github/steven-murray/mrpy/docs/examples/heirarchical_model_stan.ipynb
- https://github/steven-murray/mrpy/docs/examples/explore_analytic_model.ipynb
- https://github/steven-murray/mrpy/docs/examples/mmin_dependence.ipynb
- https://github/steven-murray/mrpy/docs/examples/physical_dependence.ipynb
- https://github/steven-murray/mrpy/docs/examples/parameterization_performance.ipynb
- https://github/steven-murray/mrpy/docs/examples/SMHM.ipynb
- Added
modelargument tofit_perobj_stanto facilitate pickling of multiple fits. - Added ability to send keyword arguments to priors in
PerObjFitclass - Added a
normal_priorfunction for simple normal priors.
- Changed default weighting from 1 to 0 in
get_fit_curve. - Added tests for the
PerObjLikeWeightsclass. - Added tests for
nbarandrhobarfor generalmin ``MRP` subclasses. - Changed imports so that they wouldn't show up in docs
- Many improvements to documentation (including this file!)
- Fixed issue setting
log_mmininIdealAnalytic - Fixed issue in which
nbarandrhobarare wrong ifmminis notm.min()inMRPsubclasses.