• Improvements to the 2D backend calculations
• Added array size thresholds for handing off to serialized fill methods

No real changes; minor release to fix conda-forge build on macOS

No real changes; minor release to fix MANIFEST.in for conda-forge builds.

• OpenMP is now required to build from source (not a difficult dependency to satisfy)
  • omp function argument deprecated
  • omp_max_threads renamed to omp_get_max_threads to mirror OpenMP C API
  • omp_available function removed
• Backend C++ code for 1D histograms rewritten (use Python and NumPy C APIs directly, no pybind11), more types supported (avoid conversions)
• Rewrite multiweight histograms (still via pybind11) to support more types (avoid conversions)
• Python code moved to the top level module (__init__.py) (without changing public API)

• OpenMP inspection improved; new functions replace pygram11.OPENMP:
  • pygram11.omp_available() -> bool checks for availability
  • pygram11.omp_max_threads() -> int checks for the maximum available threads
• some documentation improvements
• bump pybind11 version to 2.4.3
• pygram11.OPENMP will be removed in a future release

This is the first release dropping Python 2 support

• Add type hints
• Improve documentation

• In the backend we replace std::unique_ptr<T[]> with std::vector<T> for on-thread arrays.
• Improved documentation.

This is the last release supporting Python 2

• Fix setup.py such that setuptools doesn't try to install numpy via easy_install (remove unnecessary setup_requires argument).
• Add _max_threads attribute to pygram11._core.
• Fixed macOS wheels (use delocate to ensure OpenMP symbols are bundled).

New features

• histogramming multiple weight variations in a single function call is now possible via fix1dmw and var1dmw
• In the NumPy like API, passing a 2 dimensional array to the weights argument will use this feature as well.
• Binary wheels for Linux and macOS are provided on PyPI (conda-forge binaries are of course still available as well)

Other updates

• All functions now return the sqrt(sum-of-weights-squared) instead of sum-of-weights; before v0.5.x the 2D functions returned the sum of weights squared.

Second 0.5 series alpha.

• Added fixed bin one dimensional multiweight histogramming to front end Python API.
• Updated documentation

First alpha for 0.5.x release series. This is meant to test manylinux2010 wheels uploaded to PyPI.

A breaking change to the API w.r.t. returns. All functions now return sum of weights squared, even if it's None. The histogramming functions previously had a single array returned if weights were not used. Now None is returned if no weights are used

>>> x = np.random.randn(1000)
>>> w = np.random.uniform(0.5, 0.8, x.shape[0])
>>> pygram11.histogram(x, bins=5, range=(-3, 3))
(array([ 30., 249., 450., 227.,  41.]), None)
>>> pygram11.histogram(x, bins=5, range=(-3, 3), weights=w)
(array([ 19.97159325, 162.31069262, 291.30576246, 147.4478516 ,
        27.12625336]), array([ 3.66815712, 10.36897935, 13.85860416,  9.86909866,  4.27544407]))

before this release the first histogram function call would not return a two entry tuple.

Tagging minor release for documentation updates

1D histograms now have a flow argument, when True the under and overflow contents are included in the first and last bins, respectively.

Backend maintenance

• cleaning up some use of std::move
• cleaning up multiple returns code

• For 1D histogramming, allow the omp argument to be "auto".
  • For fixed width histograms OpenMP will be used for len(data) > 1e4, for variable width histograms its used for len(data) > 1e3. This is the new default.
  • If OpenMP is unavailable this is obviously ignored.
• Fixes some warnings that showed up with the Apple Clang update (related to std::move).

• New density argument for one dimensional histograms
• New pygram11.utils module, houses a densify1d function used internally by the histogramming functions if the density argument is True.

First release to work with conda-forge.

Updated linker flags in setup.py for conda-forge OpenMP on macOS attempt.

Release with a cleaner setup.py OpenMP discovery workflow.