What’s New

v2.19.1.1 (25. Oct 2019)

  • Update Stan source to v2.19.1 (release notes)
  • Fix order of per-chain values returned by Fit summary (GH646)
  • Inverse mass matrix can be provided (GH505)

v2.19.0.0 (31. Mar 2019)

  • Update Stan source to v2.19.0 (release notes)
  • Stan 2.19 requires a compiler which supports C++14. For example, if you use gcc as your compiler you will need to use gcc version 4.9.3 or higher.
  • New sampler parameters for VB mode: log_p__ and log_g__ values are added to sample file (csv) if defined with sample_file.
  • Reminder: Python 2 is not supported after 2019-12-31. PyStan will phase out testing of Python 2 after this point.

v2.18.1.0 (22. Dec 2018)

  • Update Stan source to v2.18.1 (release notes)
  • Performance improvements for extract method.

v2.18.0.0 (16. Aug 2018)

  • Update Stan source to v2.18.0 (release notes)
  • Fit method to_dataframe organizes draws into a Pandas DataFrame. Only works if Pandas is installed. Thanks to Liam Brannigan and Ari Hartikainen.
  • Improved effective sample size calculation. Thanks to Aki Vehtari.
  • Implemented pystan.diagnostics. Thanks to @jjramsey.
  • Several minor bug fixes.

v2.17.1.0 (16. Jan 2018)

  • Update Stan source to v2.17.1 (release notes) (bugfix release)
  • Remove deprecation warnings (Thanks to Alexander Rudiuk)
  • Drop testing of Python 3.4.

v2.17.0.0 (6. Oct 2017)

  • Update Stan source to v2.17.0 (release notes),
  • Added pystan.stansummary function. Patch by @ahartikainen.
  • Marked pystan.stan as deprecated. It will be removed in version 3.0. Please compile and use a Stan program as two separate steps.
  • Reminder: Consider using Python 3.5 or higher. In a future release, Python 3.4 wheels will no longer be built.
  • Reminder: Stan v2.16.0 is the final release which will not require a C++11 compatible compiler. Future releases will require a C++11 compatible compiler. (This does not affect most users.)

v2.16.0.0 (22. June 2017)

  • Update Stan source to v2.16.0 (release notes),
  • Ari Hartikainen (Aalto University) @ahartikainen joins the Stan development team. 🎉
  • Added pystan.lookup (contributed by Marco Inacio, @randommm)
  • NOTE: Stan v2.16.0 is the final release which will not require a C++11 compatible compiler. Future releases will require a C+11 compatible compiler. The vast majority of users have a compatible compiler.

v2.15.0.1 (2. May 2017)

  • Python 2.7 compatibility fix (#332). Thanks to @monga for the report.

v2.15.0.0 (21. Apr 2017)

  • Update Stan source to v2.15.0 (release notes)
  • Allow UTF-8 comments in Stan Program code. Thanks to @ahartikainen
  • Expose constrain_pars method, thanks to Lars Mescheder.

V2.14.0.0 (1. Jan 2017)

  • Update Stan source to v2.14.0 (release notes), includes important fix to the default sampling algorithm (NUTS). All users are encouraged to upgrade.
  • Several documentation and minor bug fixes (thanks @ahartikainen, @jrings, @nesanders)
  • New OpenPGP signing key for use with PyPI. Key fingerprint is C3542448245BEC68F43070E4CCB669D9761F0CAC.

V2.12.0.0 (15. Sept 2016)

  • Update Stan source to v2.12.0 (release notes)
  • #239 Fix bug in array indexing (thanks @stephen-hoover)
  • #254 FIx off-by-one error in estimated sample size calculation

V2.11.0.0 (28. July 2016)

V2.10.0.0 (18. July 2016)

  • Update Stan source to v2.10.0 (release notes)
  • Sampling in Fixed_param mode now works. Thanks to @luac for the fix and @axch for the original report.
  • Detailed installation instructions from @chendaniely added to the documentation.

v2.9.0.0 (7. Jan 2016)

  • Update Stan source to v2.9.0 (release notes)
  • Bugs fixed in _chains.pyx and model.py (thanks to @stephen-hoover, Paul Kernfeld)

v2.8.0.2 (6. Nov 2015)

  • Cython 0.22 or higher requirement included on PyPI

v2.8.0.1 (5. Nov 2015)

  • Python 3.5 support added
  • Cython 0.22 or higher now required
  • Compiler optimization (-O2) turned on for model compilation. This should increase sampling speed.
  • Significant bug fixes (pickling, pars keyword)

v2.8.0.0 (1. Oct 2015)

v2.7.0.1 (22. August 2015)

  • Minor Cython 0.23.1 compatibility fixes
  • Bug preventing mean_pars from being recorded

v2.7.0.0 (21. July 2015)

v2.6.3.0 (21. Mar 2015)

v2.6.0.0 (9. Feb 2015)

v2.5.0.2 (30. Jan 2015)

  • Fix bug in rdump (for >1 dimensional arrays)

v2.5.0.1 (14. Nov 2014)

  • Support for pickling fit objects (experimental)
  • Fix bug that occurs when printing fit summary

v2.5.0.0 (21. Oct 2014)

  • Update Stan source to v2.5.0
  • Fix several significant bugs in the extract method

v2.4.0.3 (9. Sept 2014)

  • Performance improvements for the printed summary of a fit.

v2.4.0.2 (6. Sept 2014)

  • Performance improvements for the extract method (5-10 times faster)
  • Performance improvements for the printed summary of a fit. Printing a summary of a model with more than a hundred parameters is not recommended. Consider using extract and calculating summary statistics for the parameters of interest.

v2.4.0.1 (31. July 2014)

  • Sets LBFGS as default optimizer.
  • Adds preliminary support for Python binary wheels on OS X and Windows.
  • Fixes bug in edge case in new summary code.

v2.4.0.0 (26. July 2014)

  • Stan 2.4 (LBFGS optimizer added, Nesterov removed)
  • Improve display of fit summaries

v2.3.0.0 (26. June 2014)

  • Stan 2.3 (includes user-defined functions, among other improvements).
  • Optimizing returns a vector (array) by default instead of a dictionary.

v2.2.0.1 (30. April 2014)

  • Add support for reading Stan’s R dump files.
  • Add support for specifying parameters of interest in stan.
  • Add Windows installation instructions. Thanks to @patricksnape.
  • Lighten source distribution.

v2.2.0.0 (16. February 2014)

  • Updates Stan to v2.2.0.

v2.1.0.1 (27. January 2014)

  • Implement model name obfuscation. Thanks to @karnold
  • Improve documentation of StanFit objects

v2.1.0.0 (26. December 2013)

  • Updates Stan code to v2.1.0.

v2.0.1.3 (18. December 2013)

  • Sampling is parallel by default.
  • grad_log_prob method of fit objects is available.

v2.0.1.2 (1. December 2013)

  • Improves setuptools support.
  • Allows sampling chains in parallel using multiprocessing. See the n_jobs parameter for stan() and the sampling method.
  • Allows users to specify initial values for chains.

v2.0.1.1 (18. November 2013)

  • Clean up random_seed handling (Stephan Hoyer).
  • Add fit methods get_seed, get_inits, and get_stancode.

v2.0.1.0 (24. October 2013)

  • Updated to Stan 2.0.1.
  • Specifying sample_file now works as expected.

v2.0.0.1 (23. October 2013)

  • Stan array parameters are now handled correctly.
  • Ancillary methods added to fit instances.
  • Fixed bug that caused parameters in control dict to be ignored.

v2.0.0.0 (21. October 2013)

  • Stan source updated to to 2.0.0.
  • PyStan version now mirrors Stan version.
  • Rudimentary plot and traceplot methods have been added to fit instances.
  • Warmup and sampling progress now visible.

v.0.2.2 (28. September 2013)

  • log_prob method available from StanFit instances.
  • Estimated sample size and Rhat included in summary.

v.0.2.1 (17. September 2013)

  • StanModel instances can now be pickled.
  • Adds basic support for saving output to sample_file.

v.0.2.0 (25. August 2013)

  • optimizing method working for scalar, vector, and matrix parameters
  • stanfit objects now have summary and __str__ methods à la RStan
  • stan source updated to commit cc82d51d492d26f754fd56efe22a99191c80217b (July 26, 2013)
  • IPython-relevant bug fixes

v.0.1.1 (19. July 2013)

  • Support for Python 2.7 and Python 3.3
  • stan and stanc working with common arguments