Changelog for python312-iminuit-2.28.0-69.5.x86_64.rpm :

* Mon Sep 23 2024 Ben Greiner - Fix numba/numpy resolve blocker in obs: Test in multibuild
* -test: no numba, numpy >= 2.1
* -test-numba: numpy < 2.1 for numba 0.60
* Remove entry_points empty file workaround: fixed in toolchain
* Fri Aug 02 2024 Atri Bhattacharya - Update to version 2.28.0:
* drop support for python-3.8 (gh#scikit-hep/iminuit#1015).
* Fix LeastSquares for functions with more than two arguments (gh#scikit-hep/iminuit#1016).
* Add name argument to all cost functions (gh#scikit-hep/iminuit#1017).- Bump minimum required python-devel to 3.9.
* Fri Mar 15 2024 Ben Greiner - Skip testing with ipywidgets on python39: no longer supported since ipython 8.19- Clean dependencies- Drop rpmlintrc
* Tue Feb 20 2024 Atri Bhattacharya - Update to version 2.25.2:
* update to latest root (gh#scikit-hep/iminuit#970).- Changes from version 2.25.1:
* improve docs for make_with_signature (gh#scikit-hep/iminuit#963).
* Update benchmark to ROOT-6.30 (gh#scikit-hep/iminuit#967).
* fix LeastSquares.visualize for models that accept parameter array (gh#scikit-hep/iminuit#968).- Changes from version 2.25.0:
* Some updates to the Ruff config (gh#scikit-hep/iminuit#933).
* Add extra config to pytest (gh#scikit-hep/iminuit#932).
* Cost gradient support (gh#scikit-hep/iminuit#936).
* Warn on errordef override (gh#scikit-hep/iminuit#937).
* fix: include debug info on failures (gh#scikit-hep/iminuit#946).
* Benchmark update to ROOT 6.30 (gh#scikit-hep/iminuit#951).
* Fix use of removed array rules in test (gh#scikit-hep/iminuit#952).
* Fix docstring parsing (gh#scikit-hep/iminuit#953).
* fix test that requires scipy and raise error on invalid value for use_pdf (gh#scikit-hep/iminuit#962).
* Approximate CDF from PDF (gh#scikit-hep/iminuit#950).- Do not disable any tests, fixed upstream.
* Thu Sep 07 2023 Markéta Machová - Update to 2.24.0
* Fix CostSum.visualize bug
* Fix _safe_log on systems which use 32 bit floats
* Iteration limit in smart sampling to fix behavior for step functions
* Clarify meaning of 2d contours in minuit.draw_mnmatrix
* Support interval type and check compatibility with pydantic
* Fri Jun 30 2023 Atri Bhattacharya - Update to version 2.22.0:
* add basic latex display support if unicodeit is installed [gh#scikit-help/iminuit#858].
* replace flake8 with ruff [gh#scikit-help/iminuit#859].
* Integer as variable [gh#scikit-help/iminuit#860].
* experimental mncontour algorithm [gh#scikit-help/iminuit#861].
* improve error message for cost function [gh#scikit-help/iminuit#863].
* Add RooFit tutorial [gh#scikit-help/iminuit#867].
* Use unicodeitplus instead of unicodeit to render LaTeX as unicode [gh#scikit-help/iminuit#868].
* Better ruff settings and adjustments, improvements to README [gh#scikit-help/iminuit#871].
* Update Progressbar [gh#scikit-help/iminuit#872].
* Add hide_modules and deprecated_parameters [gh#scikit-help/iminuit#873].
* Forward kwargs in Minuit.visualize to plotting function [gh#scikit-help/iminuit#874].
* Ignore missing matplotlib when calling Minuit.repr_html() [gh#scikit-help/iminuit#875].
* Rename keyword nbins to bins in UnbinnedCost.visualize [gh#scikit-help/iminuit#876].
* Roofit tutorials [gh#scikit-help/iminuit#877].
* better log-spacing detection [gh#scikit-help/iminuit#878].
* Add LeastSquares.pulls and LeastSquares.prediction [gh#scikit-help/iminuit#880].
* fix-typo-in-basic [gh#scikit-help/iminuit#888].
* Update fcn.hpp [gh#scikit-help/iminuit#889].
* Improve docs [gh#scikit-help/iminuit#890].
* Make covariance fields in display easier to understand [gh#scikit-help/iminuit#891].
* New benchmarks [gh#scikit-help/iminuit#893].
* Add Minuit.fixto [gh#scikit-help/iminuit#894].
* fix: make nograd not use grad at all in automatic diff doc [gh#scikit-help/iminuit#895].
* Use scikit-build-core [gh#scikit-help/iminuit#812].
* Update conclusions after the fix from Jonas [gh#scikit-help/iminuit#899].
* Hide confusing notes in docs: \"Not to be initialized by users.\" [gh#scikit-help/iminuit#906].- Drop missing-headers.patch: Incorporated upstream.- Add BuildRequires: python-scikit-build-core > 0.3.0.- Add %{name}.rpmlintc to suppress a false positive about a missing python module (only required for python < 3.9).
* Tue Jun 20 2023 Andreas Schwab - missing-headers.patch: add missing headers- Use the default compiler- Build with %{optflags}
* Mon May 29 2023 Dirk Müller - renenable build for python311 (numba available)
* Thu May 04 2023 Atri Bhattacharya - Update to version 2.21.3:
* Better docs for limits from annotated model parameters (gh#scikit-hep/scikit-hep-testdata#853).
* Bump pypa/cibuildwheel from 2.12.0 to 2.12.1 (gh#scikit-hep/scikit-hep-testdata#855).
* fix template input modification bug in Template class (gh#scikit-hep/scikit-hep-testdata#856).
* Fri Mar 24 2023 Atri Bhattacharya - Update to version 2.21.2:
* Fix CITATION.CFF.- Changes from version 2.21.1:
* Fix string annotations (gh#scikit-hep/iminuit#849).
* Specifiy minimum required numpy version (gh#scikit-hep/iminuit#848).- Changes from version 2.21.0:
* Fix of matrix_format (gh#scikit-hep/iminuit#843).
* Support annotated model parameters (gh#scikit-hep/iminuit#839).
* Visualize fit in minuit._repr_html_ (gh#scikit-hep/iminuit#838).- Disable builds for python 3.11 as python311-numba is missing.- Use GCC 12 for openSUSE >= 1550 as GCC 13 (now the default) causes build failures.- Update minimum version for (Build)Requires: python-numpy in keeping with upstream.- New BuildRequires: python-typing_extensions (for tests).
* Sat Jan 07 2023 Dirk Müller - update to v2.18.0:
* Bump actions/checkout from 2 to 3
* ci: update to Python 3.11 final release
* move tutorials
* added visualize function to Minuit
* Bump pypa/cibuildwheel from 2.10.2 to 2.11.2
* add more checks for gradients
* Add Python 3.11, drop 3.6
* Wed Aug 17 2022 Atri Bhattacharya - Update to version 2.16.0:
* Make FixedView act as mask for other views [gh#scikit-hep/iminuit#781].
* Fix typing issues and enable mypy in pre-commit [gh#scikit-hep/iminuit#783].
* Comparison with broadcast [gh#scikit-hep/iminuit#784].
* Fix corner case treatment of linear constraint [gh#scikit-hep/iminuit#785].
* Root update [gh#scikit-hep/iminuit#786].- Disable \'test_matrix\' test in the check section for 32-bit as the test fails for these archs due to a tolerance issue.
* Mon Jul 04 2022 Atri Bhattacharya - Update to version 2.12.1:
* cost.BarlowBeestonLite: method “hpd” has been modified to fix performance in cases where bins are not dominated by a single template.
* Wed Jun 22 2022 Atri Bhattacharya - Update to version 2.12.0:
* New - New cost function cost.BarlowBeestonLite for template fits with templates obtained from simulation or sWeighted data - Formerly private chi2 utility cost functions (cost.poisson_chi2, etc.), are now part of public API - Support custom grid in Minuit.profile, iminuit.mncontour, iminuit.contour - Handle common CL values in Minuit.mnprofile and Minuit.mncontour without scipy
* Fixes - Skip tests that use np.float128 on platforms where this type is not supported - Minuit.valid now returns False if EDM is NaN - subtract_min setting is no longer ignored by Minuit.draw_contour
* Documentation: New study about template fits
* Minuit no longer warns when a function is used that has no errordef attribute and Minuit.errordef is not explicitly set. The function is assumed to be chi-square-like up to an arbitrary constant, unless errordef is explicitly set to something else.
* More type correctness in API, better hiding of private objects in library
* Add option to use external pybind11- Pass cmake option to build against external pybind11; add BuildRequires: python-pybind11 and python-pybind11-devel and remove bundled pybind11 from source dir.
* Sun May 01 2022 Atri Bhattacharya - Update to version 2.11.2:
* Fixed wording in cost function tutorial- Changes from version 2.11.1:
* Fixed a failure of util.make_with_signature in some situations.
* Raise numpy.VisibleDeprecationWarning instead of warnings.DeprecationWarning.
* util.propagate is deprecated in favour of jacobi.propagate from the jacobi library.- Changes from version 2.11.0:
* All builtin cost functions now support multidimensional data
* Matrix.to_dict was added for symmetry with BasicValueView.to_dict
* For long-running fits, total runtime is now shown in FMin display and total runtime can be accessed via property FMin.time
* In binned fits when ndof is zero, show reduced chi2 = nan in the FMin display instead of raising a ZeroDivisionError
* Tutorials and studies are now listed separately
* Tutorial for fits of multivariate data were added
* The cost function tutorial was improved
* Studies in regard to performance were added, including a comparison with RooFit
* Sat Mar 05 2022 Atri Bhattacharya - Update to version 2.10.0:
* Minuit.draw_mncontour now works with matplotlib >= 3.5.
* Builtin cost functions now work correctly when the mask is set and data is updated on the existing cost function.
* Builtin cost functions are now more performant when used with weighted binned data.
* Thu Jan 20 2022 Atri Bhattacharya - Update to version 2.9.0:
* Minuit.draw_mncontour now works with matplotlib >= 3.5.
* Builtin cost functions now work correctly when the mask is set and data is updated on the existing cost function.
* Builtin cost functions are now more performant when used with weighted binned data.
* Wed Dec 08 2021 pgajdosAATTsuse.com- version update to 2.8.4 Fixes
* Pickling of util.Matrix resulted in incomplete state after unpickling, which would cause an exception when you tried to print the matrix Documentation
* New tutorial on fitting PDFs that depend on a conditional variable
* Fixed JAX tutorial, adapting to change in their interface
* Extended documentation of cost functions
* Thu Sep 16 2021 Atri Bhattacharya - Update to version 2.8.3:
* util.propagate now discriminates between diverging derivates (using the value NaN for the derivate) and non-converging derivatives (using the best value computed so far for the derivative).
* Update documentation.- Changes from version 2.8.2:
* Minuit.draw_mncontour can now be used by passing a single float to keyword cl, in addition to passing a list of floats.
* Use pybind11::ssize_t everywhere instead of non-standard ssize_t to fix compilation against Python-3.10 on Windows.
* Update documentation.
* Mon Aug 09 2021 Atri Bhattacharya - Update to version 2.8.1:
* Undocumented changes.
* Wed Jul 28 2021 Atri Bhattacharya - Update to version 2.8.0:
* Minor API change: Minuit.mncontour now raises RuntimeError instead of ValueError if it is not called at a valid minimum point.
* New features: Minuit.mncontour can now be called at any point without running a minimiser before, similar to Minuit.minos.
* Fixes: Minuit.mncontour used to fail if called twice in a row.
* Thu Jun 17 2021 Atri Bhattacharya - Update to version 2.6.1:
* Calling Minuit.fixed[...] = False on parameter that was not fixed before lead to undefined behaviour in Minuit2 C++ code (patch submitted to ROOT).
* Upgrade Minuit2 C++ code to latest ROOT master with simplified internal class structure and class tags replaced with enums.
* Wed May 05 2021 Atri Bhattacharya - Update to version 2.6.0:
* Builtin cost functions now report the number of data points with the attribute Cost.ndata.
* New attribute Minuit.ndof returns the degrees of freedom if the cost function reports it or NaN.
* New attribute FMin.reduced_chi2 to report the reduced chi2 of the fit; returns NaN if the reduced chi2 cannot be computed for the cost function, in case of unbinned maximum-likelihood or when the attribute Cost.ndata is missing.
* Thu Apr 22 2021 Steve Kowalik - Do not build for Python 3.6, due to no NumPy.
* Mon Feb 15 2021 Atri Bhattacharya - Update to version 2.4.0:
* New features: - iminuit.minimize: Keyword method now accepts \"migrad\" and \"simplex\". - iminuit.minimize: Keyword option now supports keyword \"stra\" to set Minuit.strategy. - iminuit.minimize: OptimizeResult.message now states if errors are not reliable.
* Minuit now supports functions wrapped with functools.partial.
* Upgrade Minuit2 C++ code in ROOT to latest master with following improvements: - improvement of seed when using an analytical gradient. - fix of last minimum state added twice to vector of minimum states in some cases (saves a bit of memory).
* Documentation improvements.
* Updated tutorial about automatic differentiation, added comparison of numba.njit and jax.jit.- Changes from version 2.3.0:
* New features: iminuit.cost.BinnedNLL and iminuit.cost.ExtendedBinnedNLL now support weighted binned data.
* Bug-fixes: FMin.edm_goal now remains unchanged if Minuit.hesse is run after Minuit.migrad.
* Other: Update to cibuildwheels-1.8.0 and workflow simplification.