diff --git a/.all-contributorsrc b/.all-contributorsrc index 678e7dc2..1179f316 100644 --- a/.all-contributorsrc +++ b/.all-contributorsrc @@ -158,6 +158,16 @@ "contributions": [ "doc" ] + }, + { + "login": "amangoel185", + "name": "Aman Goel", + "avatar_url": "https://avatars.githubusercontent.com/u/10528392?v=4", + "profile": "http://amangoel.me", + "contributions": [ + "doc", + "code" + ] } ], "contributorsPerLine": 7, diff --git a/.github/workflows/examples.yml b/.github/workflows/examples.yml index f42b3e36..b0c7bb21 100644 --- a/.github/workflows/examples.yml +++ b/.github/workflows/examples.yml @@ -20,7 +20,7 @@ jobs: - uses: actions/setup-python@v1 with: python-version: '3.8' - - name: Pre-install Numpy + - name: Pre-install NumPy run: python -m pip install -r dev-requirements.txt nbconvert ipykernel - name: Install kernel run: python -m ipykernel install --user --name boost-hist diff --git a/.github/workflows/tests.yml b/.github/workflows/tests.yml index fbc6d417..a5eb70f2 100644 --- a/.github/workflows/tests.yml +++ b/.github/workflows/tests.yml @@ -49,11 +49,10 @@ jobs: strategy: fail-fast: false matrix: - python-version: ["3.6", "3.9", "pypy3"] + python-version: ["3.6", "3.9", "3.10-dev", "pypy3"] include: - python-version: "3.8" cmake-extras: "-DCMAKE_CXX_STANDARD=17" - - python-version: "3.10-dev" name: CMake Python ${{ matrix.python-version }} @@ -83,6 +82,7 @@ jobs: name: Wheels on ${{ matrix.os }} runs-on: ${{ matrix.os }} strategy: + fail-fast: false matrix: os: [ubuntu-latest, windows-latest, macos-latest] @@ -91,11 +91,10 @@ jobs: with: submodules: true - - uses: pypa/cibuildwheel@v2.0.0a4 + - uses: pypa/cibuildwheel@v2.1.2 env: CIBW_BUILD: cp38-win_amd64 cp310-manylinux_i686 cp37-macosx_x86_64 CIBW_BUILD_VERBOSITY: 1 - CIBW_PRERELEASE_PYTHONS: 1 - uses: actions/upload-artifact@v2 with: diff --git a/.github/workflows/wheels.yml b/.github/workflows/wheels.yml index a9bfcc88..543ec20e 100644 --- a/.github/workflows/wheels.yml +++ b/.github/workflows/wheels.yml @@ -17,9 +17,13 @@ on: types: - published +concurrency: + group: wheels-${{ github.head_ref }} + cancel-in-progress: true + env: SETUPTOOLS_SCM_PRETEND_VERSION: ${{ github.event.inputs.overrideVersion }} - CIBW_ENVIRONMENT: "PIP_ONLY_BINARY=numpy SETUPTOOLS_SCM_PRETEND_VERSION=${{ github.event.inputs.overrideVersion }}" + CIBW_ENVIRONMENT: "PIP_PREFER_BINARY=1 SETUPTOOLS_SCM_PRETEND_VERSION=${{ github.event.inputs.overrideVersion }}" jobs: build_sdist: @@ -34,7 +38,7 @@ jobs: run: pipx run build --sdist - name: Check metadata - run: pipx run twine check dist/* + run: pipx run twine check --strict dist/* - uses: actions/upload-artifact@v2 with: @@ -46,7 +50,7 @@ jobs: runs-on: ubuntu-20.04 strategy: matrix: - python: [36, 37, 38, 39] + python: [36, 37, 38, 39, 310] arch: [aarch64] steps: @@ -58,7 +62,7 @@ jobs: with: platforms: all - - uses: pypa/cibuildwheel@v2.0.0a4 + - uses: pypa/cibuildwheel@v2.1.2 env: CIBW_BUILD: cp${{ matrix.python }}-* CIBW_ARCHS: ${{ matrix.arch }} @@ -79,14 +83,16 @@ jobs: strategy: fail-fast: false matrix: - os: [windows-latest, macos-latest] + os: [windows-latest, macos-latest, ubuntu-latest] arch: [auto64] - build: ["*"] + build: ["?p3?-* *macos* *win*"] + CIBW_MANYLINUX_X86_64_IMAGE: [manylinux2010] + CIBW_MANYLINUX_I686_IMAGE: [manylinux2010] include: - os: ubuntu-latest - arch: auto - build: "pp*" + arch: auto32 + build: "?p3?-*" CIBW_MANYLINUX_X86_64_IMAGE: manylinux2010 CIBW_MANYLINUX_I686_IMAGE: manylinux2010 @@ -97,6 +103,13 @@ jobs: CIBW_MANYLINUX_X86_64_IMAGE: skhep/manylinuxgcc-x86_64 CIBW_MANYLINUX_I686_IMAGE: skhep/manylinuxgcc-i686 + - os: ubuntu-latest + type: ManyLinux2014 + arch: auto + build: "cp310-*" + CIBW_MANYLINUX_X86_64_IMAGE: manylinux2014 + CIBW_MANYLINUX_I686_IMAGE: manylinux2014 + - os: macos-latest arch: universal2 build: "*" @@ -110,7 +123,7 @@ jobs: with: submodules: true - - uses: pypa/cibuildwheel@v2.0.0a4 + - uses: pypa/cibuildwheel@v2.1.2 env: CIBW_BUILD: ${{ matrix.build }} CIBW_MANYLINUX_I686_IMAGE: ${{ matrix.CIBW_MANYLINUX_I686_IMAGE }} diff --git a/.gitignore b/.gitignore index cd34cf7d..b1769507 100644 --- a/.gitignore +++ b/.gitignore @@ -34,6 +34,7 @@ # Build directories /*build* /docs/_build/* +/examples/*/*build* /pip-wheel-metadata/* # Outputs @@ -56,6 +57,7 @@ *.swp *~ *.clangd +/.cache/* compile_commands.json # Venvs such as /.env @@ -68,12 +70,14 @@ compile_commands.json /node_modules/* /yarn.lock /package.json +/package-lock.json # Testing /.benchmarks/* +/.hypothesis/* # PyCharm -.idea +/.idea/* -# All-contributors -package-lock.json +# VSCode +/.vscode/* diff --git a/.pre-commit-config.yaml b/.pre-commit-config.yaml index bd1f66a4..a830f23d 100644 --- a/.pre-commit-config.yaml +++ b/.pre-commit-config.yaml @@ -4,26 +4,9 @@ ci: repos: - repo: https://github.com/psf/black - rev: 21.6b0 + rev: 21.8b0 hooks: - - id: black - types: [file] - types_or: [python, pyi] - -- repo: https://github.com/nbQA-dev/nbQA - rev: 0.13.1 - hooks: - - id: nbqa-black - additional_dependencies: [black==20.8b1] - args: ["--nbqa-mutate"] - -# - repo: https://github.com/asottile/blacken-docs -# rev: v1.10.0 -# hooks: -# - id: blacken-docs -# additional_dependencies: [black==20.8b1] -# args: ["--target-version=py36"] -# stages: ["manual"] + - id: black-jupyter - repo: https://github.com/pre-commit/pre-commit-hooks rev: v4.0.1 @@ -46,13 +29,13 @@ repos: additional_dependencies: [pyyaml] - repo: https://github.com/asottile/pyupgrade - rev: v2.20.0 + rev: v2.26.0 hooks: - id: pyupgrade args: [--py36-plus] - repo: https://github.com/PyCQA/isort - rev: 5.9.1 + rev: 5.9.3 hooks: - id: isort @@ -60,6 +43,7 @@ repos: rev: v1.17.0 hooks: - id: setup-cfg-fmt + args: [--max-py-version=3.10] - repo: https://github.com/pycqa/flake8 rev: 3.9.2 @@ -74,7 +58,7 @@ repos: - id: mypy files: src args: [] - additional_dependencies: [numpy==1.20.*, uhi, types-dataclasses] + additional_dependencies: [numpy==1.21.*, uhi, types-dataclasses] - repo: https://github.com/mgedmin/check-manifest rev: "0.46" @@ -82,13 +66,20 @@ repos: - id: check-manifest stages: [manual] +- repo: https://github.com/codespell-project/codespell + rev: v2.1.0 + hooks: + - id: codespell + args: ["-L", "hist,nd,circularly,ba"] + exclude: ^(notebooks/xarray.ipynb|notebooks/BoostHistogramHandsOn.ipynb)$ + - repo: local hooks: - - id: disallow-pybind - name: Disallow PyBind (dual caps) + - id: disallow-caps + name: Disallow improper capitalization language: pygrep - entry: PyBind - exclude: .pre-commit.*.yaml + entry: PyBind|Numpy|Cmake|CCache|Github|PyTest + exclude: .pre-commit-config.yaml - repo: local hooks: diff --git a/README.md b/README.md index c761569b..aeda83f5 100644 --- a/README.md +++ b/README.md @@ -48,7 +48,7 @@ hist.fill( [0.1, 0.4, 0.9], ) -# Numpy array view into histogram counts, no overflow bins +# NumPy array view into histogram counts, no overflow bins values = hist.values() # Make a new histogram with just the second axis, summing over the first, and @@ -142,7 +142,7 @@ histograms can be plotted via any compatible library, such as [mplhep][]. * `v = h[{0:b}]`: All actions can be represented by `axis:item` dictionary instead of by position (mostly useful for slicing) * Slicing to get histogram or set array of values * `h2 = h[a:b]`: Access a slice of a histogram, cut portions go to flow bins if present - * `h2 = h[:, ...]`: Using `:` and `...` supported just like Numpy + * `h2 = h[:, ...]`: Using `:` and `...` supported just like NumPy * `h2 = h[::sum]`: Third item in slice is the "action" * `h[...] = array`: Set the bin contents, either include or omit flow bins * Special accessors @@ -186,12 +186,12 @@ platforms have wheels provided in boost-histogram: | System | Arch | Python versions | PyPy versions | |---------|-----|------------------|--------------| -| ManyLinux1 (custom GCC 9.2) | 32 & 64-bit | 3.7, 3.8 | | -| ManyLinux2010 | 32 & 64-bit | 3.6, 3.7, 3.8, 3.9 | (64-bit) 7.3: 3.7 | -| ManyLinux2014 | ARM64 | 3.6, 3.7, 3.8, 3.9 | | -| macOS 10.9+ | 64-bit | 3.6, 3.7, 3.8, 3.9 | 7.3: 3.7 | -| macOS Universal2 | Arm64 | 3.8, 3.9 | | -| Windows | 32 & 64-bit | 3.6, 3.7, 3.8, 3.9 | (64-bit) 7.3: 3.7 | +| ManyLinux1 (custom GCC 9.2) | 32 & 64-bit | 3.6, 3.7, 3.8 | | +| ManyLinux2010 | 32 & 64-bit | 3.6, 3.7, 3.8, 3.9, 3.10 | (64-bit) 7.3: 3.7 | +| ManyLinux2014 | ARM64 | 3.6, 3.7, 3.8, 3.9, 3.10 | | +| macOS 10.9+ | 64-bit | 3.6, 3.7, 3.8, 3.9, 3.10 | 7.3: 3.7 | +| macOS Universal2 | Arm64 | 3.8, 3.9, 3.10 | | +| Windows | 32 & 64-bit | 3.6, 3.7, 3.8, 3.9, 3.10 | (64-bit) 7.3: 3.7 | * manylinux1: Using a custom docker container with GCC 9 to produce. Anything running Python 3.9 should be compatible with manylinux2010, so manylinux1 not provided for Python 3.9 (like NumPy). @@ -255,6 +255,7 @@ We would like to acknowledge the contributors that made this project possible ([
Konstantin Gizdov
๐Ÿ“ฆ ๐Ÿ›
Kyle Cranmer
๐Ÿ“– +
Aman Goel
๐Ÿ“– ๐Ÿ’ป diff --git a/docs/CHANGELOG.md b/docs/CHANGELOG.md index 8ec040cd..ce16ad9f 100644 --- a/docs/CHANGELOG.md +++ b/docs/CHANGELOG.md @@ -1,7 +1,40 @@ # What's new in boost-histogram +## Version 1.2 + +### Version 1.2.0 + +#### User changes +* Python 3.10 officially supported, with wheels. +* Support subtraction on histograms [#636][] +* Integer histograms are now signed [#636][] + +#### Bug fixes +* Support custom setters on AxesTuple subclasses. [#627][] +* Faster picking if slices are not also used [#645][] or if they are [#648][] (1000x or more in some cases) +* Throw an error when an AxesTuple setter is the wrong length (inspired by zip strict in Python 3.10) [#627][] +* Fix error thrown on comparison with axis and non-axis object [#631][] +* Static typing no longer thinks `storage=` is required [#604][] + +#### Developer changes +* Support NumPy 1.21 for static type checking [#625][] +* Use newer Boost 1.77 and Boost.Histogram 1.77+1 [#594][] +* Provide nox support [#647][] + +[#594]: https://github.com/scikit-hep/boost-histogram/pull/594 +[#604]: https://github.com/scikit-hep/boost-histogram/pull/604 +[#625]: https://github.com/scikit-hep/boost-histogram/pull/625 +[#627]: https://github.com/scikit-hep/boost-histogram/pull/627 +[#631]: https://github.com/scikit-hep/boost-histogram/pull/631 +[#636]: https://github.com/scikit-hep/boost-histogram/pull/636 +[#645]: https://github.com/scikit-hep/boost-histogram/pull/645 +[#647]: https://github.com/scikit-hep/boost-histogram/pull/647 +[#648]: https://github.com/scikit-hep/boost-histogram/pull/648 + ## Version 1.1 +### Version 1.1.0 + #### User changes * Experimentally support list selection on categorical axes [#577][] * Support Python 3.8 on Apple Silicon [#600][] @@ -15,7 +48,7 @@ #### Developer changes * Test on Python 3.10 beta releases [#600][] * Provide a CMakeLists for quick standalone Boost.Histogram C++ experiments [#591][] -* Adding logging with PyTest failure output [#575][] +* Adding logging with pytest failure output [#575][] [#575]: https://github.com/scikit-hep/boost-histogram/pull/575 [#576]: https://github.com/scikit-hep/boost-histogram/pull/576 @@ -130,7 +163,7 @@ Python 2, and mostly equivalent in API to 1.0. * Supports converting user histogram objects that provide a `_to_boost_histogram_` method. [#483][] * A `view=True` parameter must now be passed to get a View instead of a standard - NumPy values arrray from `to_numpy()`. [#498][] + NumPy values array from `to_numpy()`. [#498][] #### Bug fixes @@ -484,7 +517,7 @@ library. Using Boost 1.72 release. * Properties on accumulator views now resolve correctly [#273][] * Division of a histogram by a number is supported again [#278][] * Setting a histogram with length one slice fixed [#279][] -* Numpy functions now work with Numpy ints in `bins=` [#282][] +* NumPy functions now work with NumPy ints in `bins=` [#282][] * In-place addition avoids a copy [#284][] [#273]: https://github.com/scikit-hep/boost-histogram/pull/273 @@ -524,7 +557,7 @@ transition existing 0.5.x code to the new API. * `h.axes` now has the functions from axis as well. [#183][] * `bh.project` has become `bh.sum` [#185][] * `.reduce(...)` and the reducers in `bh.algorithm` have been removed in favor of dictionary based UHI slicing [#259][] -* `bh.numpy` module interface updates, `histogram=bh.Histogram` replaces cryptic `bh=True`, and `density=True` is now supported in Numpy mode [#256][] +* `bh.numpy` module interface updates, `histogram=bh.Histogram` replaces cryptic `bh=True`, and `density=True` is now supported in NumPy mode [#256][] * Added `hist.copy()` [#218][] and `hist.shape` [#264][] * Signatures are much nicer in Python 3 [#188][] * Reprs are better, various properties like `__module__` are now set correctly [#200][] diff --git a/docs/index.rst b/docs/index.rst index 14b9392f..8a7d2f1e 100644 --- a/docs/index.rst +++ b/docs/index.rst @@ -11,8 +11,8 @@ Welcome to boost-histogram's documentation! =========================================== -|Gitter| |Build Status| |Actions Status| |Documentation Status| |DOI| -|Code style: black| |PyPI version| |Conda-Forge| |Scikit-HEP| +|Actions Status| |Documentation Status| |Code style: black| |PyPI version| |Conda-Forge| +|PyPI platforms| |DOI| |GitHub Discussion| |Gitter| |Scikit-HEP| Boost-histogram (`source `__) is a Python package providing Python bindings for Boost.Histogram_ (`source @@ -43,7 +43,7 @@ virtual environment, etc. See :ref:`usage-installation` for more details. An exa hist.fill([.3, .5, .2], [.1, .4, .9]) - # Numpy array view into histogram counts, no overflow bins + # NumPy array view into histogram counts, no overflow bins counts = hist.view() See :ref:`usage-quickstart` for more. @@ -106,21 +106,23 @@ Indices and tables * :ref:`search` .. _Boost.Histogram: https://www.boost.org/doc/libs/release/libs/histogram/doc/html/index.html -.. |Gitter| image:: https://badges.gitter.im/HSF/PyHEP-histogramming.svg - :target: https://gitter.im/HSF/PyHEP-histogramming?utm_source=badge&utm_medium=badge&utm_campaign=pr-badge -.. |Build Status| image:: https://dev.azure.com/scikit-hep/boost-histogram/_apis/build/status/bh-tests?branchName=develop - :target: https://dev.azure.com/scikit-hep/boost-histogram/_build/latest?definitionId=2&branchName=develop .. |Actions Status| image:: https://github.com/scikit-hep/boost-histogram/workflows/Tests/badge.svg :target: https://github.com/scikit-hep/boost-histogram/actions .. |Documentation Status| image:: https://readthedocs.org/projects/boost-histogram/badge/?version=latest :target: https://boost-histogram.readthedocs.io/en/latest/?badge=latest -.. |DOI| image:: https://zenodo.org/badge/148885351.svg - :target: https://zenodo.org/badge/latestdoi/148885351 .. |Code style: black| image:: https://img.shields.io/badge/code%20style-black-000000.svg :target: https://github.com/ambv/black .. |PyPI version| image:: https://badge.fury.io/py/boost-histogram.svg - :target: https://pypi.org/project/boost-histogram/ + :target: https://pypi.org/project/boost-histogram .. |Conda-Forge| image:: https://img.shields.io/conda/vn/conda-forge/boost-histogram :target: https://github.com/conda-forge/boost-histogram-feedstock +.. |PyPI platforms| image:: https://img.shields.io/pypi/pyversions/boost-histogram + :target: https://pypi.org/project/boost-histogram/ +.. |DOI| image:: https://zenodo.org/badge/148885351.svg + :target: https://zenodo.org/badge/latestdoi/148885351 +.. |GitHub Discussion| image:: https://img.shields.io/static/v1?label=Discussions&message=Ask&color=blue&logo=github + :target: https://github.com/scikit-hep/boost-histogram/discussions +.. |Gitter| image:: https://badges.gitter.im/HSF/PyHEP-histogramming.svg + :target: https://gitter.im/HSF/PyHEP-histogramming?utm_source=badge&utm_medium=badge&utm_campaign=pr-badge .. |Scikit-HEP| image:: https://scikit-hep.org/assets/images/Scikit--HEP-Project-blue.svg :target: https://scikit-hep.org/ diff --git a/docs/usage/accumulators.rst b/docs/usage/accumulators.rst index ab182ff7..98e3b676 100644 --- a/docs/usage/accumulators.rst +++ b/docs/usage/accumulators.rst @@ -37,14 +37,14 @@ showing how non-accurate sums fail to produce the obvious answer, 2.0:: print(f"{sum(values) = } (simple)") print(f"{math.fsum(values) = }") print(f"{np.sum(values) = } (pairwise)") - print(f{bh.accumulators.Sum().fill(values) = }") + print(f"{bh.accumulators.Sum().fill(values) = }") .. code:: text - sum(values) = 0.0 + sum(values) = 0.0 (simple) math.fsum(values) = 2.0 - np.sum(values) = 0.0 - bh.accumulators.Sum().fill(values) = Sum(2) + np.sum(values) = 0.0 (pairwise) + bh.accumulators.Sum().fill(values) = Sum(0 + 2) Note that this is still intended for performance and does not guarantee @@ -53,12 +53,12 @@ orders of values:: values = [1., 1e100, 1e50, 1., -1e50, -1e100] print(f"{math.fsum(values) = }") - print(f{bh.accumulators.Sum().fill(values) = }") + print(f"{bh.accumulators.Sum().fill(values) = }") .. code:: text math.fsum(values) = 2.0 - bh.accumulators.Sum().fill(values) = Sum(0) + bh.accumulators.Sum().fill(values) = Sum(0 + 0) You should note that this is a highly contrived example and the Sum accumulator should still outperform simple and pairwise summation methods for a minimal @@ -156,7 +156,11 @@ Views ----- Most of the accumulators (except Sum) support a View. This is what is returned from -a histogram when ``.view()`` is requested. This is a structured Numpy ndarray, with a few small -additions to make them easier to work with. Like a Numpy recarray, you can access the fields with +a histogram when ``.view()`` is requested. This is a structured NumPy ndarray, with a few small +additions to make them easier to work with. Like a NumPy recarray, you can access the fields with attributes; you can even access (but not set) computed attributes like ``.variance``. A view will -also return an accumulator instance if you select a single item. +also return an accumulator instance if you select a single item. You can set a view's contents +with a stacked array, and each item in the stack will be used for the (computed) values that a +normal constructor would take. For example, WeighedMean can take an array with a final +dimension four long, with ``sum_of_weights``, ``sum_of_weights_squared``, ``value``, and ``variance`` +elements, even though several of these values are computed from the internal representation. diff --git a/docs/usage/analyses.rst b/docs/usage/analyses.rst index e6c3811a..305f4ffd 100644 --- a/docs/usage/analyses.rst +++ b/docs/usage/analyses.rst @@ -36,7 +36,7 @@ True ``is_valid`` selection, you can use a ``sum``: .. code:: python3 - h1 = hist[:, True, :: bh.sum] + h1 = hist[:, True, sum] You can expand this example to any number of dimensions, boolean flags, and categories. diff --git a/docs/usage/histogram.rst b/docs/usage/histogram.rst index 4f6369a1..ed7d6e6a 100644 --- a/docs/usage/histogram.rst +++ b/docs/usage/histogram.rst @@ -19,13 +19,25 @@ The summing accumulators (not ``Mean()`` and ``WeightedMean())``) support thread Data ^^^^ -The primary value from a histogram is always available as ``.value()``. The variance is available as ``.variances()``, unless you fill an unweighed histogram with weights, which will cause this to be return None, since the variance is no longer computable (use a weighted storage instead if you need the variances). The counts are available as ``.counts()``. If the histogram is weighted, ``.counts()`` returns the effective counts; see `UHI `_ for details. +The primary values from a histogram are always available as ``.values()``. The variances are available as ``.variances()``, unless you fill an unweighed histogram with weights, which will cause this to return None, since the variances are no longer computable (use a weighted storage instead if you need the variances). The counts are available as ``.counts()``. If the histogram is weighted, ``.counts()`` returns the effective counts; see `UHI `_ for details. Views ^^^^^ -While Histograms do conform to the Python buffer protocol, the best way to get access to the raw contents of a histogram as a NumPy array is with ``.view()``. This way you can optionally pass ``flow=True`` to get the flow bins, and if you have an accumulator storage, you will get a View, which is a slightly augmented ndarrray subclass (see :ref:`usage-accumulators`). +While Histograms do conform to the Python buffer protocol, the best way to get access to the raw contents of a histogram as a NumPy array is with ``.view()``. This way you can optionally pass ``flow=True`` to get the flow bins, and if you have an accumulator storage, you will get a View, which is a slightly augmented ndarrray subclass (see :ref:`usage-accumulators`). Views support setting as well for non-computed properties; you can use an expression like this to set the values of an accumulator storage: +.. code:: python3 + + h.view().value = values + + +You can also used stacked arrays (N+1 dimensional) to set a histogram's contents. This is especially useful if you need to set a computed value, like variance on a Mean/WeightedMean storage, which cannot be set using the above method: + +.. code:: python3 + + h[...] = np.stack([values, variances], axis=-1) + +If you leave endpoints off (such as with ``...`` above), then you can match the size with or without flow bins. Operations ^^^^^^^^^^ diff --git a/docs/usage/installation.rst b/docs/usage/installation.rst index c44e067a..2f2da569 100644 --- a/docs/usage/installation.rst +++ b/docs/usage/installation.rst @@ -79,7 +79,7 @@ are dictated by Boost.Histogramโ€™s C++ requirements: gcc >= 5.5, clang >= 3.8, msvc >= 14.1. You should have a version of pip less than 2-3 years old (10+). -Numpy is downloaded during the build (enables multithreaded builds). +NumPy is downloaded during the build (enables multithreaded builds). Boost is not required or needed (this only depends on included header-only dependencies).This library is under active development; you can install directly from GitHub if you would like. diff --git a/docs/usage/quickstart.rst b/docs/usage/quickstart.rst index df804776..23578ecf 100644 --- a/docs/usage/quickstart.rst +++ b/docs/usage/quickstart.rst @@ -64,10 +64,10 @@ should give arrays, but single values work as well: Slicing and rebinning --------------------- -You can slice into a histogram using bin coordinates -or data coordinates using ``bh.loc(v)``. You can also -rebin with ``bh.rebin(n)`` or remove an entire axis -using ``bh.sum`` as the third slice argument: +You can slice into a histogram using bin coordinates or data coordinates using +``bh.loc(v)``. You can also rebin with ``bh.rebin(n)`` or remove an entire axis +using ``sum`` (technically as the third slice argument, though it is allowed by +itself as well): .. code:: python3 @@ -76,7 +76,7 @@ using ``bh.sum`` as the third slice argument: bh.axis.Regular(10, 0, 1), bh.axis.Regular(10, 0, 1), ) - mini = hist[1:5, bh.loc(0.2) : bh.loc(0.9), :: bh.sum] + mini = hist[1:5, bh.loc(0.2) : bh.loc(0.9), sum] # Will be 4 bins x 7 bins See :ref:`usage-indexing`. @@ -84,7 +84,7 @@ See :ref:`usage-indexing`. Accessing the contents ---------------------- -You can use ``hist.values()`` to get a Numpy array from any histogram. You can +You can use ``hist.values()`` to get a NumPy array from any histogram. You can get the variances with ``hist.variances()``, though if you fill an unweighted storage with weights, this will return None, as you no longer can compute the variances correctly (please use a weighted storage if you need to). You can diff --git a/docs/usage/storage.rst b/docs/usage/storage.rst index 97e5ad64..4ca0e660 100644 --- a/docs/usage/storage.rst +++ b/docs/usage/storage.rst @@ -3,7 +3,7 @@ Storages ======== -There are several storage to choose from. To select a storage, pass the +There are several storages to choose from. To select a storage, pass the ``storage=bh.storage.`` argument when making a histogram. Simple storages diff --git a/docs/usage/transforms.rst b/docs/usage/transforms.rst index e7ff8f1c..cb871dda 100644 --- a/docs/usage/transforms.rst +++ b/docs/usage/transforms.rst @@ -38,7 +38,7 @@ transformed axis, and this will be 15-90 times slower than a compiled method, li 10, 1, 4, transform=bh.axis.transform.Function(ftype(math.log), ftype(math.exp)) ) - # Pure Python: Numpy (90x slower) + # Pure Python: NumPy (90x slower) bh.axis.Regular( 10, 1, 4, transform=bh.axis.transform.Function(ftype(np.log), ftype(np.exp)) ) @@ -152,7 +152,7 @@ That's it. Using Numba ^^^^^^^^^^^ -The same procedure works for numba decorators. Numpy only supports functions, not builtins like ``math.log``, +The same procedure works for numba decorators. NumPy only supports functions, not builtins like ``math.log``, so if you want to pass those, you'll need to wrap them in a lambda function or add a bit of logic to the convert function. Here are your options: diff --git a/examples/simple_2d.py b/examples/simple_2d.py index e6372c63..931a9da6 100755 --- a/examples/simple_2d.py +++ b/examples/simple_2d.py @@ -10,12 +10,12 @@ bh.axis.Regular(20, -3, 3, metadata="x"), bh.axis.Regular(20, -3, 3, metadata="y") ) -# Generate some Numpy arrays with data to fill into histogram, +# Generate some NumPy arrays with data to fill into histogram, # in this case normal distributed random numbers in x and y x = np.random.randn(1_000) y = 0.5 * np.random.randn(1_000) -# Fill histogram with Numpy arrays, this is very fast +# Fill histogram with NumPy arrays, this is very fast h.fill(x, y) # Get numpy.histogram compatible representation of the histogram diff --git a/examples/simple_numpy.py b/examples/simple_numpy.py index 3cc41515..6e858f6e 100755 --- a/examples/simple_numpy.py +++ b/examples/simple_numpy.py @@ -10,7 +10,7 @@ bh.axis.Regular(10, -3, 3, metadata="x"), bh.axis.Regular(10, -3, 3, metadata="y") ) -# Generate some Numpy arrays with data to fill into histogram, +# Generate some NumPy arrays with data to fill into histogram, # in this case normal distributed random numbers in x and y x_data = np.random.randn(1000) y_data = 0.5 * np.random.randn(1000) @@ -18,7 +18,7 @@ # Fill histogram with numpy arrays, this is very fast h.fill(x_data, y_data) -# Get representations of the bin edges as Numpy arrays +# Get representations of the bin edges as NumPy arrays x = h.axes[0].edges y = h.axes[1].edges diff --git a/extern/assert b/extern/assert index 2cd99e24..6aabfeba 160000 --- a/extern/assert +++ b/extern/assert @@ -1 +1 @@ -Subproject commit 2cd99e24e834d600d51cf3b8cdc3df6ecfd8e9f1 +Subproject commit 6aabfebae6d4acf996fe711de4e9b444ad88c17d diff --git a/extern/config b/extern/config index fad75493..088b79a0 160000 --- a/extern/config +++ b/extern/config @@ -1 +1 @@ -Subproject commit fad75493bd2da03524854976f2d80dd7989c90fb +Subproject commit 088b79a0ca751932010f82d3f95457c8b483fb9b diff --git a/extern/core b/extern/core index ca489306..f4b3d5db 160000 --- a/extern/core +++ b/extern/core @@ -1 +1 @@ -Subproject commit ca4893063fa244a690b1f799328f16408b794c77 +Subproject commit f4b3d5dba6f86caaf96e45901655a954b2ff68b4 diff --git a/extern/histogram b/extern/histogram index 7c24d268..4a10c2c1 160000 --- a/extern/histogram +++ b/extern/histogram @@ -1 +1 @@ -Subproject commit 7c24d2683e6be78cda90dcc0026ceeb7b5131507 +Subproject commit 4a10c2c11b272f435f61233eb9f585604476b2f8 diff --git a/extern/mp11 b/extern/mp11 index 21cace4e..9d43d1f6 160000 --- a/extern/mp11 +++ b/extern/mp11 @@ -1 +1 @@ -Subproject commit 21cace4e574180ba64d9307a5e4ea9e5e94d3e8d +Subproject commit 9d43d1f69660617266c9168e6e121ab2b0ea2287 diff --git a/extern/throw_exception b/extern/throw_exception index da6f5420..95e02ea5 160000 --- a/extern/throw_exception +++ b/extern/throw_exception @@ -1 +1 @@ -Subproject commit da6f5420fe7e96f673ee185c2ff37cbaa3f95b84 +Subproject commit 95e02ea52b8525ecf34dbf1e7fae34af09986b8a diff --git a/extern/variant2 b/extern/variant2 index 1ebc29aa..4153a535 160000 --- a/extern/variant2 +++ b/extern/variant2 @@ -1 +1 @@ -Subproject commit 1ebc29aa0277cbb284a70327b5af14519f2702bc +Subproject commit 4153a535a0fa8eb4d18abc262fcf2ae834601261 diff --git a/include/bh_python/accumulators/ostream.hpp b/include/bh_python/accumulators/ostream.hpp index 03e25158..2140d5a2 100644 --- a/include/bh_python/accumulators/ostream.hpp +++ b/include/bh_python/accumulators/ostream.hpp @@ -79,7 +79,7 @@ std::basic_ostream& operator<<(std::basic_ostream& template std::basic_ostream& operator<<(std::basic_ostream& os, - const ::boost::histogram::accumulators::thread_safe& x) { + const ::boost::histogram::accumulators::count& x) { os << x.load(); return os; } diff --git a/include/bh_python/histogram.hpp b/include/bh_python/histogram.hpp index 0d26afb0..3199ff45 100644 --- a/include/bh_python/histogram.hpp +++ b/include/bh_python/histogram.hpp @@ -20,8 +20,8 @@ namespace pybind11 { /// The descriptor for atomic_* is the same as the descriptor for *, as long this uses /// standard layout template -struct format_descriptor> : format_descriptor { - static_assert(std::is_standard_layout>::value, ""); +struct format_descriptor> : format_descriptor { + static_assert(std::is_standard_layout>::value, ""); }; } // namespace pybind11 @@ -31,11 +31,11 @@ namespace detail { template py::buffer_info make_buffer_impl(const Axes& axes, bool flow, T* ptr) { // strides are in bytes - auto shape = bh::detail::make_stack_buffer(axes); - auto strides = bh::detail::make_stack_buffer(axes); - ssize_t stride = sizeof(T); - unsigned rank = 0; - char* start = reinterpret_cast(ptr); + auto shape = bh::detail::make_stack_buffer(axes); + auto strides = bh::detail::make_stack_buffer(axes); + py::ssize_t stride = sizeof(T); + unsigned rank = 0; + char* start = reinterpret_cast(ptr); bh::detail::for_each_axis(axes, [&](const auto& axis) { const bool underflow = bh::axis::traits::options(axis) & bh::axis::option::underflow; diff --git a/include/bh_python/pybind11.hpp b/include/bh_python/pybind11.hpp index db948049..c46e9be9 100644 --- a/include/bh_python/pybind11.hpp +++ b/include/bh_python/pybind11.hpp @@ -45,7 +45,7 @@ std::string shift_to_string(const T& x) { } template -void unchecked_set_impl(std::true_type, py::tuple& tup, ssize_t i, Obj&& obj) { +void unchecked_set_impl(std::true_type, py::tuple& tup, py::ssize_t i, Obj&& obj) { // PyTuple_SetItem steals a reference to 'val' if(PyTuple_SetItem(tup.ptr(), i, obj.release().ptr()) != 0) { throw py::error_already_set(); @@ -53,7 +53,7 @@ void unchecked_set_impl(std::true_type, py::tuple& tup, ssize_t i, Obj&& obj) { } template -void unchecked_set_impl(std::false_type, py::tuple& tup, ssize_t i, T&& t) { +void unchecked_set_impl(std::false_type, py::tuple& tup, py::ssize_t i, T&& t) { unchecked_set_impl(std::true_type{}, tup, i, py::cast(std::forward(t))); } @@ -62,6 +62,6 @@ template void unchecked_set(py::tuple& tup, std::size_t i, T&& t) { unchecked_set_impl(std::is_base_of>{}, tup, - static_cast(i), + static_cast(i), std::forward(t)); } diff --git a/include/bh_python/register_histogram.hpp b/include/bh_python/register_histogram.hpp index 68ab7a78..acd41b47 100644 --- a/include/bh_python/register_histogram.hpp +++ b/include/bh_python/register_histogram.hpp @@ -97,6 +97,9 @@ auto register_histogram(py::module& m, const char* name, const char* desc) { def_optionally(hist, bh::detail::has_operator_rmul{}, py::self *= py::self); + def_optionally(hist, + bh::detail::has_operator_rsub{}, + py::self -= py::self); #ifdef __clang__ #pragma GCC diagnostic pop #endif diff --git a/include/bh_python/storage.hpp b/include/bh_python/storage.hpp index bcbdd5bd..a4027994 100644 --- a/include/bh_python/storage.hpp +++ b/include/bh_python/storage.hpp @@ -11,7 +11,7 @@ #include #include -#include +#include #include #include @@ -22,8 +22,8 @@ namespace storage { // Names match Python names -using int64 = bh::dense_storage; -using atomic_int64 = bh::dense_storage>; +using int64 = bh::dense_storage; +using atomic_int64 = bh::dense_storage>; using double_ = bh::dense_storage; using unlimited = bh::unlimited_storage<>; using weight = bh::dense_storage>; @@ -81,7 +81,11 @@ void save(Archive& ar, const storage::atomic_int64& s, unsigned /* version */) { // We cannot view the memory as a numpy array, because the internal layout of // std::atomic is undefined. So no reinterpret_casts are allowed. py::array_t a(static_cast(s.size())); - std::copy(s.begin(), s.end(), a.mutable_data()); + + auto in_ptr = s.begin(); + auto out_ptr = a.mutable_data(); + for(; in_ptr != s.end(); ++in_ptr, ++out_ptr) + *out_ptr = in_ptr->value(); ar << a; } @@ -187,7 +191,7 @@ struct type_caster { auto ptr = PyNumber_Long(src.ptr()); if(!ptr) return false; - value.store(PyLong_AsUnsignedLongLong(ptr)); + value = PyLong_AsLongLong(ptr); Py_DECREF(ptr); return !PyErr_Occurred(); } @@ -195,7 +199,7 @@ struct type_caster { static handle cast(storage::atomic_int64::value_type src, return_value_policy /* policy */, handle /* parent */) { - return PyLong_FromUnsignedLongLong(src.load()); + return PyLong_FromLongLong(src.value()); } }; } // namespace detail diff --git a/include/bh_python/transform.hpp b/include/bh_python/transform.hpp index 654ca392..3c86c12e 100644 --- a/include/bh_python/transform.hpp +++ b/include/bh_python/transform.hpp @@ -25,7 +25,7 @@ struct func_transform { py::object _forward_converted; // Held for reference counting if conversion makes a // new object (ctypes does not bump the refcount) py::object _inverse_converted; - py::object _convert_ob; // Called before computing tranform if not None + py::object _convert_ob; // Called before computing transform if not None py::str _name; // Optional name (uses repr from objects otherwise) /// Convert an object into a std::function. Can handle ctypes @@ -84,7 +84,7 @@ struct func_transform { src); } - // Note that each error is slighly different just to help with debugging + // Note that each error is slightly different just to help with debugging throw py::type_error("Only ctypes double(double) and C++ functions allowed " "(must be stateless)"); } diff --git a/notebooks/BoostHistogramHandsOn.ipynb b/notebooks/BoostHistogramHandsOn.ipynb index e6c9acd2..62c9ebe0 100644 --- a/notebooks/BoostHistogramHandsOn.ipynb +++ b/notebooks/BoostHistogramHandsOn.ipynb @@ -231,7 +231,7 @@ "cell_type": "markdown", "metadata": {}, "source": [ - "## 2: Drop-in replacement for Numpy\n", + "## 2: Drop-in replacement for NumPy\n", "\n", "To start using this yourself, you don't even need to change your code. Let's try the numpy adapters." ] @@ -330,7 +330,7 @@ "cell_type": "markdown", "metadata": {}, "source": [ - "Now we can move over to boost-histogram one step at a time! Just to be complete, we can also go back to a Numpy tuple from a Histogram object:" + "Now we can move over to boost-histogram one step at a time! Just to be complete, we can also go back to a NumPy tuple from a Histogram object:" ] }, { @@ -948,7 +948,7 @@ "cell_type": "markdown", "metadata": {}, "source": [ - "Let's try to make a density histogram like Numpy's." + "Let's try to make a density histogram like NumPy's." ] }, { diff --git a/notebooks/PerformanceComparison.ipynb b/notebooks/PerformanceComparison.ipynb index 998c425c..361fb265 100644 --- a/notebooks/PerformanceComparison.ipynb +++ b/notebooks/PerformanceComparison.ipynb @@ -81,7 +81,7 @@ "cell_type": "markdown", "metadata": {}, "source": [ - "#### Traditional 1D Numpy Histogram\n", + "#### Traditional 1D NumPy Histogram\n", "\n", "This is reasonably optimized; it should provide good perforance." ] @@ -232,7 +232,7 @@ "cell_type": "markdown", "metadata": {}, "source": [ - "#### Traditional 2D Numpy histogram\n", + "#### Traditional 2D NumPy histogram\n", "\n", "Not as well optimized for regular filling." ] diff --git a/notebooks/ThreadedFills.ipynb b/notebooks/ThreadedFills.ipynb index d39fc625..745fd93c 100644 --- a/notebooks/ThreadedFills.ipynb +++ b/notebooks/ThreadedFills.ipynb @@ -136,7 +136,7 @@ "cell_type": "markdown", "metadata": {}, "source": [ - "This makes four seperate histograms, then fills them and adds at the end." + "This makes four separate histograms, then fills them and adds at the end." ] }, { diff --git a/notebooks/xarray.ipynb b/notebooks/xarray.ipynb index afef4ecb..9746a489 100644 --- a/notebooks/xarray.ipynb +++ b/notebooks/xarray.ipynb @@ -235,7 +235,7 @@ "### More features\n", "\n", "Let's add a few more features to our function defined above.\n", - "* Let's allow bins to be a list of axes or even a completly prepared histogram; this will allow us to take advantage of boost-histogram features later.\n", + "* Let's allow bins to be a list of axes or even a completely prepared histogram; this will allow us to take advantage of boost-histogram features later.\n", "* Let's add a weights keyword so we can do weighted histograms as well." ] }, diff --git a/noxfile.py b/noxfile.py new file mode 100644 index 00000000..31225c32 --- /dev/null +++ b/noxfile.py @@ -0,0 +1,50 @@ +import nox + +ALL_PYTHONS = ["3.6", "3.7", "3.8", "3.9"] + +nox.options.sessions = ["lint", "tests"] + + +@nox.session(python=ALL_PYTHONS) +def tests(session): + """ + Run the unit and regular tests. + """ + session.install(".[test]") + session.run("pytest") + + +@nox.session +def docs(session): + """ + Build the docs. Pass "serve" to serve. + """ + + session.chdir("docs") + session.install("-r", "requirements.txt") + session.run("sphinx-build", "-M", "html", ".", "_build") + + if session.posargs: + if "serve" in session.posargs: + session.log("Launching docs at http://localhost:8000/ - use Ctrl-C to quit") + session.run("python", "-m", "http.server", "8000", "-d", "_build/html") + else: + session.error("Unsupported argument to docs") + + +@nox.session +def lint(session): + """ + Run the linter. + """ + session.install("pre-commit") + session.run("pre-commit", "run", "--all-files") + + +@nox.session +def make_pickle(session): + """ + Make a pickle file for this version + """ + session.install(".[dev]") + session.run("python", "tests/pickles/make_pickle.py", *session.posargs) diff --git a/pyproject.toml b/pyproject.toml index 35b1759f..2dce28cf 100644 --- a/pyproject.toml +++ b/pyproject.toml @@ -1,9 +1,8 @@ [build-system] requires = [ - "setuptools>=42", + "setuptools>=45", "wheel", - "toml", - "setuptools_scm>=4.1.2" + "setuptools_scm[toml]>=4.1.2" ] build-backend = "setuptools.build_meta" @@ -12,7 +11,7 @@ write_to = "src/boost_histogram/version.py" [tool.pytest.ini_options] junit_family = "xunit2" -addopts = "--benchmark-disable -Wd --strict-markers" +addopts = "--benchmark-disable -Wd --strict-markers -ra" xfail_strict = true testpaths = ["tests"] required_plugins = ["pytest-benchmark"] @@ -85,5 +84,5 @@ ignore = [ [tool.cibuildwheel] test-extras = "test" test-command = "pytest {project}/tests" -test-skip = ["pp*macos*", "pp*win*", "*universal2:arm64"] +test-skip = ["pp*macos*", "pp*win*", "*universal2:arm64", "cp310*macos*"] skip = ["pp*-manylinux_i686"] # not supported by NumPy diff --git a/scripts/performance_report.py b/scripts/performance_report.py index 93877a01..93e869a8 100644 --- a/scripts/performance_report.py +++ b/scripts/performance_report.py @@ -72,7 +72,7 @@ def print_timer(setup, statement, name, storage, fill, flow, base=None, n=10): base = print_timer( setup_1d, "h, _ = np.histogram(vals, bins=bins, range=ranges)", - name="Numpy", + name="NumPy", storage="uint64", fill="", flow="no", @@ -163,7 +163,7 @@ def print_timer(setup, statement, name, storage, fill, flow, base=None, n=10): base = print_timer( setup_2d, "H, *ledges = np.histogram2d(*vals, bins=bins, range=ranges)", - name="Numpy", + name="NumPy", storage="uint64", fill="", flow="no", diff --git a/setup.cfg b/setup.cfg index 5e934a87..ae0b2dd0 100644 --- a/setup.cfg +++ b/setup.cfg @@ -28,6 +28,7 @@ classifiers = Programming Language :: Python :: 3.7 Programming Language :: Python :: 3.8 Programming Language :: Python :: 3.9 + Programming Language :: Python :: 3.10 Programming Language :: Python :: Implementation :: CPython Programming Language :: Python :: Implementation :: PyPy Topic :: Scientific/Engineering diff --git a/src/boost_histogram/_core/algorithm.pyi b/src/boost_histogram/_core/algorithm.pyi index c04ad15d..32fbadbb 100644 --- a/src/boost_histogram/_core/algorithm.pyi +++ b/src/boost_histogram/_core/algorithm.pyi @@ -2,6 +2,7 @@ import enum import typing class reduce_command: + iaxis: int def __repr__(self) -> str: ... class slice_mode(enum.Enum): diff --git a/src/boost_histogram/_core/axis/__init__.pyi b/src/boost_histogram/_core/axis/__init__.pyi index 5947ee9b..63f24a4f 100644 --- a/src/boost_histogram/_core/axis/__init__.pyi +++ b/src/boost_histogram/_core/axis/__init__.pyi @@ -34,13 +34,13 @@ class _BaseAxis: @property def extent(self) -> int: ... @property - def edges(self) -> np.ndarray: ... + def edges(self) -> "np.typing.NDArray[Any]": ... @property - def centers(self) -> np.ndarray: ... + def centers(self) -> "np.typing.NDArray[Any]": ... @property - def widths(self) -> np.ndarray: ... - def index(self, arg0: ArrayLike) -> int | np.ndarray: ... - def value(self, arg0: ArrayLike) -> float | np.ndarray: ... + def widths(self) -> "np.typing.NDArray[Any]": ... + def index(self, arg0: ArrayLike) -> int | "np.typing.NDArray[Any]": ... + def value(self, arg0: ArrayLike) -> float | "np.typing.NDArray[Any]": ... class _BaseRegular(_BaseAxis): def __init__(self, bins: int, start: float, stop: float) -> None: ... diff --git a/src/boost_histogram/_core/hist.pyi b/src/boost_histogram/_core/hist.pyi index 8fc5d48e..523f3741 100644 --- a/src/boost_histogram/_core/hist.pyi +++ b/src/boost_histogram/_core/hist.pyi @@ -22,8 +22,8 @@ class _BaseHistogram: def __copy__(self: T) -> T: ... def __deepcopy__(self: T, memo: Any) -> T: ... def __iadd__(self: T, other: _BaseHistogram) -> T: ... - def to_numpy(self, flow: bool = ...) -> Tuple[np.ndarray, ...]: ... - def view(self, flow: bool = ...) -> np.ndarray: ... + def to_numpy(self, flow: bool = ...) -> Tuple["np.typing.NDArray[Any]", ...]: ... + def view(self, flow: bool = ...) -> "np.typing.NDArray[Any]": ... def axis(self, i: int = ...) -> axis._BaseAxis: ... def fill(self, *args: ArrayLike, weight: ArrayLike | None = ...) -> None: ... def empty(self, flow: bool = ...) -> bool: ... diff --git a/src/boost_histogram/_internal/axestuple.py b/src/boost_histogram/_internal/axestuple.py index 549cdb39..d2fb31d4 100644 --- a/src/boost_histogram/_internal/axestuple.py +++ b/src/boost_histogram/_internal/axestuple.py @@ -4,7 +4,7 @@ import numpy as np from .axis import Axis -from .utils import set_module +from .utils import set_module, zip_strict A = TypeVar("A", bound="ArrayTuple") @@ -17,12 +17,12 @@ class ArrayTuple(tuple): # type: ignore def __getattr__(self, name: str) -> Any: if name in self._REDUCTIONS: - return partial(getattr(np, name), np.broadcast_arrays(*self)) + return partial(getattr(np, name), np.broadcast_arrays(*self)) # type: ignore else: return self.__class__(getattr(a, name) for a in self) def __dir__(self) -> List[str]: - names = dir(self.__class__) + dir(np.ndarray) + names = dir(self.__class__) + dir("np.typing.NDArray[Any]") return sorted(n for n in names if not n.startswith("_")) def __call__(self, *args: Any, **kwargs: Any) -> Any: @@ -34,7 +34,7 @@ def broadcast(self: A) -> A: Use this method to broadcast them out into their full memory representation. """ - return self.__class__(np.broadcast_arrays(*self)) + return self.__class__(np.broadcast_arrays(*self)) # type: ignore B = TypeVar("B", bound="AxesTuple") @@ -56,17 +56,17 @@ def extent(self) -> Tuple[int, ...]: @property def centers(self) -> ArrayTuple: gen = (s.centers for s in self) - return ArrayTuple(np.meshgrid(*gen, **self._MGRIDOPTS)) + return ArrayTuple(np.meshgrid(*gen, **self._MGRIDOPTS)) # type: ignore @property def edges(self) -> ArrayTuple: gen = (s.edges for s in self) - return ArrayTuple(np.meshgrid(*gen, **self._MGRIDOPTS)) + return ArrayTuple(np.meshgrid(*gen, **self._MGRIDOPTS)) # type: ignore @property def widths(self) -> ArrayTuple: gen = (s.widths for s in self) - return ArrayTuple(np.meshgrid(*gen, **self._MGRIDOPTS)) + return ArrayTuple(np.meshgrid(*gen, **self._MGRIDOPTS)) # type: ignore def value(self, *indexes: float) -> Tuple[float, ...]: if len(indexes) != len(self): @@ -97,7 +97,10 @@ def __getattr__(self, attr: str) -> Any: return self.__class__(getattr(s, attr) for s in self) def __setattr__(self, attr: str, values: Any) -> None: - self.__class__(s.__setattr__(attr, v) for s, v in zip(self, values)) + try: + return super().__setattr__(attr, values) + except AttributeError: + self.__class__(s.__setattr__(attr, v) for s, v in zip_strict(self, values)) value.__doc__ = Axis.value.__doc__ index.__doc__ = Axis.index.__doc__ diff --git a/src/boost_histogram/_internal/axis.py b/src/boost_histogram/_internal/axis.py index 1ff830b1..818ffd86 100644 --- a/src/boost_histogram/_internal/axis.py +++ b/src/boost_histogram/_internal/axis.py @@ -138,10 +138,10 @@ def bin(self, index: float) -> Union[int, str, Tuple[float, float]]: return self._ax.bin(index) # type: ignore def __eq__(self, other: Any) -> bool: - return self._ax == other._ax # type: ignore + return hasattr(other, "_ax") and self._ax == other._ax def __ne__(self, other: Any) -> bool: - return self._ax != other._ax # type: ignore + return (not hasattr(other, "_ax")) or self._ax != other._ax @classmethod def _convert_cpp(cls: Type[T], cpp_object: Any) -> T: @@ -248,18 +248,18 @@ def __getitem__(self, i: AxCallOrInt) -> Union[int, str, Tuple[float, float]]: return self.bin(i) @property - def edges(self) -> np.ndarray: + def edges(self) -> "np.typing.NDArray[Any]": return self._ax.edges # type: ignore @property - def centers(self) -> np.ndarray: + def centers(self) -> "np.typing.NDArray[Any]": """ An array of bin centers. """ return self._ax.centers # type: ignore @property - def widths(self) -> np.ndarray: + def widths(self) -> "np.typing.NDArray[Any]": """ An array of bin widths. """ diff --git a/src/boost_histogram/_internal/hist.py b/src/boost_histogram/_internal/hist.py index fa5ddb58..6e13b731 100644 --- a/src/boost_histogram/_internal/hist.py +++ b/src/boost_histogram/_internal/hist.py @@ -65,7 +65,9 @@ T = TypeVar("T") -def _fill_cast(value: T, *, inner: bool = False) -> Union[T, np.ndarray, Tuple[T, ...]]: +def _fill_cast( + value: T, *, inner: bool = False +) -> Union[T, "np.typing.NDArray[Any]", Tuple[T, ...]]: """ Convert to NumPy arrays. Some buffer objects do not get converted by forcecast. If not called by itself (inner=False), then will work through one level of tuple/list. @@ -150,8 +152,8 @@ def __init__(self, *args: CppHistogram, metadata: Any = ...) -> None: def __init__( self, *axes: Union[Axis, CppAxis], - storage: Storage, - metadata: Any = None, + storage: Storage = ..., + metadata: Any = ..., ) -> None: ... @@ -297,13 +299,13 @@ def ndim(self) -> int: def view( self, flow: bool = False - ) -> Union[np.ndarray, WeightedSumView, WeightedMeanView, MeanView]: + ) -> Union["np.typing.NDArray[Any]", WeightedSumView, WeightedMeanView, MeanView]: """ Return a view into the data, optionally with overflow turned on. """ return _to_view(self._hist.view(flow)) - def __array__(self) -> np.ndarray: + def __array__(self) -> "np.typing.NDArray[Any]": return self.view(False) def __eq__(self, other: Any) -> bool: @@ -312,11 +314,15 @@ def __eq__(self, other: Any) -> bool: def __ne__(self, other: Any) -> bool: return (not hasattr(other, "_hist")) or self._hist != other._hist - def __add__(self: H, other: Union["Histogram", np.ndarray, float]) -> H: + def __add__( + self: H, other: Union["Histogram", "np.typing.NDArray[Any]", float] + ) -> H: result = self.copy(deep=False) return result.__iadd__(other) - def __iadd__(self: H, other: Union["Histogram", np.ndarray, float]) -> H: + def __iadd__( + self: H, other: Union["Histogram", "np.typing.NDArray[Any]", float] + ) -> H: if isinstance(other, (int, float)) and other == 0: return self self._compute_inplace_op("__iadd__", other) @@ -326,36 +332,69 @@ def __iadd__(self: H, other: Union["Histogram", np.ndarray, float]) -> H: return self - def __radd__(self: H, other: Union["Histogram", np.ndarray, float]) -> H: + def __radd__( + self: H, other: Union["Histogram", "np.typing.NDArray[Any]", float] + ) -> H: return self + other + def __sub__( + self: H, other: Union["Histogram", "np.typing.NDArray[Any]", float] + ) -> H: + result = self.copy(deep=False) + return result.__isub__(other) + + def __isub__( + self: H, other: Union["Histogram", "np.typing.NDArray[Any]", float] + ) -> H: + if isinstance(other, (int, float)) and other == 0: + return self + self._compute_inplace_op("__isub__", other) + + self.axes = self._generate_axes_() + + return self + # If these fail, the underlying object throws the correct error - def __mul__(self: H, other: Union["Histogram", np.ndarray, float]) -> H: + def __mul__( + self: H, other: Union["Histogram", "np.typing.NDArray[Any]", float] + ) -> H: result = self.copy(deep=False) return result._compute_inplace_op("__imul__", other) - def __rmul__(self: H, other: Union["Histogram", np.ndarray, float]) -> H: + def __rmul__( + self: H, other: Union["Histogram", "np.typing.NDArray[Any]", float] + ) -> H: return self * other - def __truediv__(self: H, other: Union["Histogram", np.ndarray, float]) -> H: + def __truediv__( + self: H, other: Union["Histogram", "np.typing.NDArray[Any]", float] + ) -> H: result = self.copy(deep=False) return result._compute_inplace_op("__itruediv__", other) - def __div__(self: H, other: Union["Histogram", np.ndarray, float]) -> H: + def __div__( + self: H, other: Union["Histogram", "np.typing.NDArray[Any]", float] + ) -> H: result = self.copy(deep=False) return result._compute_inplace_op("__idiv__", other) - def __idiv__(self: H, other: Union["Histogram", np.ndarray, float]) -> H: + def __idiv__( + self: H, other: Union["Histogram", "np.typing.NDArray[Any]", float] + ) -> H: return self._compute_inplace_op("__idiv__", other) - def __itruediv__(self: H, other: Union["Histogram", np.ndarray, float]) -> H: + def __itruediv__( + self: H, other: Union["Histogram", "np.typing.NDArray[Any]", float] + ) -> H: return self._compute_inplace_op("__itruediv__", other) - def __imul__(self: H, other: Union["Histogram", np.ndarray, float]) -> H: + def __imul__( + self: H, other: Union["Histogram", "np.typing.NDArray[Any]", float] + ) -> H: return self._compute_inplace_op("__imul__", other) def _compute_inplace_op( - self: H, name: str, other: Union["Histogram", np.ndarray, float] + self: H, name: str, other: Union["Histogram", "np.typing.NDArray[Any]", float] ) -> H: # Also takes CppHistogram, but that confuses mypy because it's hard to pick out if isinstance(other, Histogram): @@ -442,26 +481,26 @@ def fill( }: raise RuntimeError("Mean histograms do not support threaded filling") - data = [np.array_split(a, threads) for a in args_ars] + data = [np.array_split(a, threads) for a in args_ars] # type: ignore if weight is None or np.isscalar(weight): assert threads is not None weights = [weight_ars] * threads else: - weights = np.array_split(weight_ars, threads) + weights = np.array_split(weight_ars, threads) # type: ignore if sample_ars is None or np.isscalar(sample_ars): assert threads is not None samples = [sample_ars] * threads else: - samples = np.array_split(sample_ars, threads) + samples = np.array_split(sample_ars, threads) # type: ignore if self._hist._storage_type is _core.storage.atomic_int64: def fun( weight: Optional[ArrayLike], sample: Optional[ArrayLike], - *args: np.ndarray, + *args: "np.typing.NDArray[Any]", ) -> None: self._hist.fill(*args, weight=weight, sample=sample) @@ -471,7 +510,7 @@ def fun( def fun( weight: Optional[ArrayLike], sample: Optional[ArrayLike], - *args: np.ndarray, + *args: "np.typing.NDArray[Any]", ) -> None: local_hist = self._hist.__copy__() local_hist.reset() @@ -646,9 +685,12 @@ def _compute_commonindex( def to_numpy( self, flow: bool = False, *, dd: bool = False, view: bool = False - ) -> Union[Tuple[np.ndarray, ...], Tuple[np.ndarray, Tuple[np.ndarray, ...]]]: + ) -> Union[ + Tuple["np.typing.NDArray[Any]", ...], + Tuple["np.typing.NDArray[Any]", Tuple["np.typing.NDArray[Any]", ...]], + ]: """ - Convert to a Numpy style tuple of return arrays. Edges are converted to + Convert to a NumPy style tuple of return arrays. Edges are converted to match NumPy standards, with upper edge inclusive, unlike boost-histogram, where upper edge is exclusive. @@ -799,8 +841,32 @@ def __getitem__( # noqa: C901 assert isinstance(stop, int) slices.append(_core.algorithm.slice_and_rebin(i, start, stop, merge)) - logger.debug("Reduce with %s", slices) - reduced = self._hist.reduce(*slices) + # Will be updated below + if slices or pick_set or pick_each or integrations: + reduced = self._hist + else: + logger.debug("Reduce actions are all empty, just making a copy") + reduced = copy.copy(self._hist) + + if pick_each: + tuple_slice = tuple( + pick_each.get(i, slice(None)) for i in range(reduced.rank()) + ) + logger.debug("Slices for pick each: %s", tuple_slice) + axes = [ + reduced.axis(i) for i in range(reduced.rank()) if i not in pick_each + ] + logger.debug("Axes: %s", axes) + new_reduced = reduced.__class__(axes) + new_reduced.view(flow=True)[...] = reduced.view(flow=True)[tuple_slice] + reduced = new_reduced + integrations = {i - sum(j <= i for j in pick_each) for i in integrations} + for slice_ in slices: + slice_.iaxis -= sum(j <= slice_.iaxis for j in pick_each) + + if slices: + logger.debug("Reduce with %s", slices) + reduced = reduced.reduce(*slices) if pick_set: warnings.warn( @@ -830,20 +896,6 @@ def __getitem__( # noqa: C901 new_reduced.view(flow=True)[...] = reduced_view reduced = new_reduced - if pick_each: - tuple_slice = tuple( - pick_each.get(i, slice(None)) for i in range(reduced.rank()) - ) - logger.debug("Slices for pick each: %s", tuple_slice) - axes = [ - reduced.axis(i) for i in range(reduced.rank()) if i not in pick_each - ] - logger.debug("Axes: %s", axes) - new_reduced = reduced.__class__(axes) - new_reduced.view(flow=True)[...] = reduced.view(flow=True)[tuple_slice] - reduced = new_reduced - integrations = {i - sum(j <= i for j in pick_each) for i in integrations} - if integrations: projections = [i for i in range(reduced.rank()) if i not in integrations] reduced = reduced.project(*projections) @@ -887,7 +939,7 @@ def __setitem__( if ( value.ndim > 0 and len(view.dtype) > 0 # type: ignore - and len(value.dtype) == 0 # type: ignore + and len(value.dtype) == 0 and len(view.dtype) == value.shape[-1] # type: ignore ): value_shape = value.shape[:-1] @@ -984,7 +1036,7 @@ def kind(self) -> Kind: else: return Kind.COUNT - def values(self, flow: bool = False) -> np.ndarray: + def values(self, flow: bool = False) -> "np.typing.NDArray[Any]": """ Returns the accumulated values. The counts for simple histograms, the sum of weights for weighted histograms, the mean for profiles, etc. @@ -995,7 +1047,7 @@ def values(self, flow: bool = False) -> np.ndarray: :param flow: Enable flow bins. Not part of PlottableHistogram, but included for consistency with other methods and flexibility. - :return: np.ndarray[np.float64] + :return: "np.typing.NDArray[Any]"[np.float64] """ view = self.view(flow) @@ -1005,7 +1057,7 @@ def values(self, flow: bool = False) -> np.ndarray: else: return view.value # type: ignore - def variances(self, flow: bool = False) -> Optional[np.ndarray]: + def variances(self, flow: bool = False) -> Optional["np.typing.NDArray[Any]"]: """ Returns the estimated variance of the accumulated values. The sum of squared weights for weighted histograms, the variance of samples for profiles, etc. @@ -1026,7 +1078,7 @@ def variances(self, flow: bool = False) -> Optional[np.ndarray]: :param flow: Enable flow bins. Not part of PlottableHistogram, but included for consistency with other methods and flexibility. - :return: np.ndarray[np.float64] + :return: "np.typing.NDArray[Any]"[np.float64] """ view = self.view(flow) @@ -1053,7 +1105,7 @@ def variances(self, flow: bool = False) -> Optional[np.ndarray]: else: return view.variance # type: ignore - def counts(self, flow: bool = False) -> np.ndarray: + def counts(self, flow: bool = False) -> "np.typing.NDArray[Any]": """ Returns the number of entries in each bin for an unweighted histogram or profile and an effective number of entries (defined below) @@ -1073,7 +1125,7 @@ def counts(self, flow: bool = False) -> np.ndarray: The larger the spread in weights, the smaller it is, but it is always 0 if filled 0 times, and 1 if filled once, and more than 1 otherwise. - :return: np.ndarray[np.float64] + :return: "np.typing.NDArray[Any]"[np.float64] """ view = self.view(flow) diff --git a/src/boost_histogram/_internal/utils.py b/src/boost_histogram/_internal/utils.py index 1d9fc1d6..6fe7abe8 100644 --- a/src/boost_histogram/_internal/utils.py +++ b/src/boost_histogram/_internal/utils.py @@ -1,5 +1,17 @@ +import itertools +import sys import typing -from typing import Callable, ClassVar, Iterator, Optional, Set, Type, TypeVar +from typing import ( + Any, + Callable, + ClassVar, + Iterator, + Optional, + Set, + Tuple, + Type, + TypeVar, +) import boost_histogram @@ -176,3 +188,16 @@ def _walk_subclasses(cls: Type[object]) -> Iterator[Type[object]]: # user subclasses to work yield from _walk_subclasses(base) yield base + + +def zip_strict(*args: Any) -> Iterator[Tuple[Any, ...]]: + if sys.version_info >= (3, 10): + yield from zip(*args, strict=True) + return + + marker = object() + for each in itertools.zip_longest(*args, fillvalue=marker): + for val in each: + if val is marker: + raise ValueError("zip() arguments are not the same length") + yield each diff --git a/src/boost_histogram/_internal/view.py b/src/boost_histogram/_internal/view.py index 6ebd9c51..b06ab245 100644 --- a/src/boost_histogram/_internal/view.py +++ b/src/boost_histogram/_internal/view.py @@ -6,11 +6,11 @@ from .typing import ArrayLike, StrIndex, Ufunc -class View(np.ndarray): +class View(np.ndarray): # type: ignore __slots__ = () _FIELDS: ClassVar[Tuple[str, ...]] - def __getitem__(self, ind: StrIndex) -> np.ndarray: + def __getitem__(self, ind: StrIndex) -> "np.typing.NDArray[Any]": sliced = super().__getitem__(ind) # If the shape is empty, return the parent type @@ -24,7 +24,7 @@ def __getitem__(self, ind: StrIndex) -> np.ndarray: return sliced # type: ignore def __repr__(self) -> str: - # Numpy starts the ndarray class name with "array", so we replace it + # NumPy starts the ndarray class name with "array", so we replace it # with our class name return f"{self.__class__.__name__}(\n " + repr(self.view(np.ndarray))[6:] @@ -97,13 +97,13 @@ class WeightedSumView(View): __slots__ = () _PARENT = WeightedSum - value: np.ndarray - variance: np.ndarray + value: "np.typing.NDArray[Any]" + variance: "np.typing.NDArray[Any]" # Could be implemented on master View def __array_ufunc__( self, ufunc: Ufunc, method: str, *inputs: Any, **kwargs: Any - ) -> np.ndarray: + ) -> "np.typing.NDArray[Any]": # This one is defined for record arrays, so just use it # (Doesn't get picked up the pass-through) @@ -130,14 +130,14 @@ def __array_ufunc__( # Addition of two views if input_0.dtype == input_1.dtype: - if ufunc in {np.add}: + if ufunc in {np.add, np.subtract}: ufunc( input_0["value"], input_1["value"], out=result["value"], **kwargs, ) - ufunc( + np.add( input_0["variance"], input_1["variance"], out=result["variance"], @@ -205,13 +205,13 @@ class WeightedMeanView(View): __slots__ = () _PARENT = WeightedMean - sum_of_weights: np.ndarray - sum_of_weights_squared: np.ndarray - value: np.ndarray - _sum_of_weighted_deltas_squared: np.ndarray + sum_of_weights: "np.typing.NDArray[Any]" + sum_of_weights_squared: "np.typing.NDArray[Any]" + value: "np.typing.NDArray[Any]" + _sum_of_weighted_deltas_squared: "np.typing.NDArray[Any]" @property - def variance(self) -> np.ndarray: + def variance(self) -> "np.typing.NDArray[Any]": with np.errstate(divide="ignore", invalid="ignore"): return self["_sum_of_weighted_deltas_squared"] / ( # type: ignore self["sum_of_weights"] @@ -224,20 +224,20 @@ class MeanView(View): __slots__ = () _PARENT = Mean - count: np.ndarray - value: np.ndarray - _sum_of_deltas_squared: np.ndarray + count: "np.typing.NDArray[Any]" + value: "np.typing.NDArray[Any]" + _sum_of_deltas_squared: "np.typing.NDArray[Any]" # Variance is a computation @property - def variance(self) -> np.ndarray: + def variance(self) -> "np.typing.NDArray[Any]": with np.errstate(divide="ignore", invalid="ignore"): - return self["_sum_of_deltas_squared"] / (self["count"] - 1) # type: ignore + return self["_sum_of_deltas_squared"] / (self["count"] - 1) def _to_view( - item: np.ndarray, value: bool = False -) -> Union[np.ndarray, WeightedSumView, WeightedMeanView, MeanView]: + item: "np.typing.NDArray[Any]", value: bool = False +) -> Union["np.typing.NDArray[Any]", WeightedSumView, WeightedMeanView, MeanView]: for cls in View.__subclasses__(): if cls._FIELDS == item.dtype.names: ret = item.view(cls) diff --git a/src/boost_histogram/numpy.py b/src/boost_histogram/numpy.py index 2488409c..a0eb8666 100644 --- a/src/boost_histogram/numpy.py +++ b/src/boost_histogram/numpy.py @@ -38,7 +38,7 @@ def histogramdd( if normed is not None: raise KeyError( - "normed=True is not recommended for use in Numpy, and is not supported in boost-histogram; use density=True instead" + "normed=True is not recommended for use in NumPy, and is not supported in boost-histogram; use density=True instead" ) if density and histogram is not None: raise KeyError( @@ -67,7 +67,7 @@ def histogramdd( if np.issubdtype(type(b), np.integer): if r is None: # Nextafter may affect bin edges slightly - r = (np.min(a[n]), np.max(a[n])) + r = (np.amin(a[n]), np.amax(a[n])) cpp_ax = _core.axis.regular_numpy(b, r[0], r[1]) new_ax = _cast(None, cpp_ax, _axis.Axis) axs.append(new_ax) @@ -183,6 +183,6 @@ def histogram( lets you set the number of threads to fill with (0 for auto, None for 1). """ - f.__doc__ = H.format(n.__name__) + n.__doc__ + f.__doc__ = H.format(n.__name__) + (n.__doc__ or "") del f, n, H diff --git a/src/register_accumulators.cpp b/src/register_accumulators.cpp index 5e2b51fd..77453d6c 100644 --- a/src/register_accumulators.cpp +++ b/src/register_accumulators.cpp @@ -59,7 +59,7 @@ decltype(auto) make_buffer() { void register_accumulators(py::module& accumulators) { // Naming convention: - // If a value is publically available in Boost.Histogram accumulators + // If a value is publicly available in Boost.Histogram accumulators // as a method, it has the same name in the numpy record array. // If it is not available except through a computation, it has // the same name as the private property without the trailing _. diff --git a/src/register_algorithm.cpp b/src/register_algorithm.cpp index fbf2e6d8..3d1aee03 100644 --- a/src/register_algorithm.cpp +++ b/src/register_algorithm.cpp @@ -10,6 +10,7 @@ void register_algorithms(py::module& algorithm) { py::class_(algorithm, "reduce_command") .def(py::init()) + .def_readwrite("iaxis", &bh::algorithm::reduce_command::iaxis) .def("__repr__", [](const bh::algorithm::reduce_command& self) { using range_t = bh::algorithm::reduce_command::range_t; diff --git a/src/register_axis.cpp b/src/register_axis.cpp index 88b68935..32687087 100644 --- a/src/register_axis.cpp +++ b/src/register_axis.cpp @@ -55,7 +55,7 @@ void register_axes(py::module& mod) { "bins"_a, "start"_a, "stop"_a, - "tranform"_a) + "transform"_a) .def_property_readonly("transform", [](const axis::regular_trans& self) { return self.transform(); }); diff --git a/tests/conftest.py b/tests/conftest.py index c43903a7..fe67f385 100644 --- a/tests/conftest.py +++ b/tests/conftest.py @@ -1,6 +1,6 @@ import pytest -import boost_histogram # noqa: F401 +import boost_histogram as bh @pytest.fixture(params=(False, True), ids=("no_growth", "growth")) @@ -29,3 +29,30 @@ def flow(request): ) def metadata(request): return request.param + + +@pytest.fixture( + params=( + bh.storage.Double, + bh.storage.Int64, + bh.storage.AtomicInt64, + bh.storage.Weight, + bh.storage.Unlimited, + ), + ids=("Double", "Int64", "AtomicInt64", "Weight", "Unlimited"), +) +def count_storage(request): + return request.param + + +@pytest.fixture( + params=( + bh.storage.Double, + bh.storage.Int64, + bh.storage.AtomicInt64, + bh.storage.Unlimited, + ), + ids=("Double", "Int64", "AtomicInt64", "Unlimited"), +) +def count_single_storage(request): + return request.param diff --git a/tests/pickles/bh_1.1.0.pkl b/tests/pickles/bh_1.1.0.pkl new file mode 100644 index 00000000..aaaab253 Binary files /dev/null and b/tests/pickles/bh_1.1.0.pkl differ diff --git a/tests/pickles/make_pickle.py b/tests/pickles/make_pickle.py index 49939c8a..97b4ba18 100755 --- a/tests/pickles/make_pickle.py +++ b/tests/pickles/make_pickle.py @@ -3,14 +3,19 @@ import pickle from pathlib import Path +from typing import Optional import typer import boost_histogram as bh +DIR = Path(__file__).parent.resolve() + def make_pickle( - output: Path = typer.Argument(..., exists=False), *, protocol: int = 2 # noqa: B008 + output: Optional[Path] = typer.Argument(None, exists=False), # noqa: B008 + *, + protocol: int = 2, ): """ Make a pickle file with the current boost-histogram for use in tests. @@ -18,7 +23,10 @@ def make_pickle( VER = tuple(map(int, bh.__version__.split("."))) - h1 = bh.Histogram(bh.axis.Regular(31, -15, 15)) + if output is None: + output = DIR / f"bh_{bh.__version__}.pkl" + + h1 = bh.Histogram(bh.axis.Regular(31, -15, 15), storage=bh.storage.Int64()) h2 = bh.Histogram( bh.axis.Integer(0, 5, metadata={"hello": "world"}), storage=bh.storage.Weight() ) diff --git a/tests/test_axis.py b/tests/test_axis.py index 2b89b0c9..ad38979e 100644 --- a/tests/test_axis.py +++ b/tests/test_axis.py @@ -88,7 +88,7 @@ def test_metadata(axis, args, opt, kwargs): # The point of this ABC is to force all the tests listed here to be implemented -# for each axis type. PyTest instantiates these test classes for us, so missing +# for each axis type. Pytest instantiates these test classes for us, so missing # one really does fail the test. class Axis(abc.ABC): @abc.abstractmethod @@ -197,6 +197,8 @@ def test_equal(self): assert a != bh.axis.Regular(3, 1.0, 2.0) assert a != bh.axis.Regular(4, 1.1, 2.0) assert a != bh.axis.Regular(4, 1.0, 2.1) + assert a != object() + assert not (a == object()) # __eq__ and __ne__ are separately implemented # metadata compare assert bh.axis.Regular(1, 2, 3, metadata=1) == bh.axis.Regular( @@ -325,12 +327,12 @@ def test_log_transform(self): assert a.index(-1) == 2 assert a.index(0.99) == -1 - assert a.index(1.0) == 0 + assert a.index(1.001) == 0 assert a.index(9.99) == 0 - assert a.index(10.0) == 1 + assert a.index(10.01) == 1 assert a.index(99.9) == 1 - assert a.index(100) == 2 - assert a.index(1000) == 2 + assert a.index(100.01) == 2 + assert a.index(1000.1) == 2 assert a.bin(0)[0] == approx(1e0) assert a.bin(1)[0] == approx(1e1) diff --git a/tests/test_benchmark_category_axis.py b/tests/test_benchmark_category_axis.py index 6c473fab..d6c3eb46 100644 --- a/tests/test_benchmark_category_axis.py +++ b/tests/test_benchmark_category_axis.py @@ -34,3 +34,39 @@ def run(h, data): h.fill(data) benchmark(run, h, tuple(values) if dtype is tuple else values.astype(dtype)) + + +@pytest.mark.benchmark(group="Pick") +def test_pick_only(benchmark): + + h = bh.Histogram( + bh.axis.StrCategory([str(i) for i in range(32)]), + bh.axis.StrCategory([str(i) for i in range(32)]), + bh.axis.StrCategory([str(i) for i in range(32)]), + bh.axis.Regular(32, 0, 320), + ) + + h[...] = 1.0 + + def run(h): + return h[bh.loc("13"), bh.loc("13"), bh.loc("13"), :].view() + + benchmark(run, h) + + +@pytest.mark.benchmark(group="Pick") +def test_pick_and_slice(benchmark): + + h = bh.Histogram( + bh.axis.StrCategory([str(i) for i in range(32)]), + bh.axis.StrCategory([str(i) for i in range(32)]), + bh.axis.StrCategory([str(i) for i in range(32)]), + bh.axis.Regular(32, 0, 320), + ) + + h[...] = 1.0 + + def run(h): + return h[3:29, bh.loc("13"), bh.loc("13"), :].view() + + benchmark(run, h) diff --git a/tests/test_histogram.py b/tests/test_histogram.py index bcc7ca30..46ac0042 100644 --- a/tests/test_histogram.py +++ b/tests/test_histogram.py @@ -1,8 +1,11 @@ import functools import operator +import pickle import sys +from collections import OrderedDict from io import BytesIO +import env import numpy as np import pytest from numpy.testing import assert_array_equal @@ -10,15 +13,6 @@ import boost_histogram as bh -try: - import cPickle as pickle -except ImportError: - import pickle - -from collections import OrderedDict - -import env - def test_init(): bh.Histogram() @@ -152,12 +146,8 @@ def test_fill_1d(flow): assert get(h, bh.overflow) == 1 -@pytest.mark.parametrize( - "storage", - [bh.storage.Int64, bh.storage.Double, bh.storage.Unlimited, bh.storage.AtomicInt64], -) -def test_setting(storage): - h = bh.Histogram(bh.axis.Regular(10, 0, 1), storage=storage()) +def test_setting(count_single_storage): + h = bh.Histogram(bh.axis.Regular(10, 0, 1), storage=count_single_storage()) h[bh.underflow] = 1 h[0] = 2 h[1] = 3 @@ -403,6 +393,50 @@ def test_add_2d_w(flow): assert h[bh.tag.at(i), bh.tag.at(j)] == 2 * m[i][j] +def test_sub_2d(flow, count_storage): + + h0 = bh.Histogram( + bh.axis.Integer(-1, 2, underflow=flow, overflow=flow), + bh.axis.Regular(4, -2, 2, underflow=flow, overflow=flow), + storage=count_storage(), + ) + + h0.fill(-1, -2) + h0.fill(-1, -1) + h0.fill(0, 0) + h0.fill(0, 1) + h0.fill(1, 0) + h0.fill(3, -1) + h0.fill(0, -3) + + m = h0.values(flow=True).copy() + + if count_storage not in {bh.storage.AtomicInt64, bh.storage.Weight}: + h = h0.copy() + h -= h0 + assert h.values(flow=True) == approx(m * 0) + + h -= h0 + assert h.values(flow=True) == approx(-m) + + h2 = h0 - (h0 + h0 + h0) + assert h2.values(flow=True) == approx(-2 * m) + + h3 = h0 - h0.view(flow=True) * 4 + assert h3.values(flow=True) == approx(-3 * m) + + h4 = h0.copy() + h4 -= h0.view(flow=True) * 5 + assert h4.values(flow=True) == approx(-4 * m) + + h5 = h0.copy() + h5 -= 2 + assert h5.values(flow=True) == approx(m - 2) + + h6 = h0 - 3 + assert h6.values(flow=True) == approx(m - 3) + + def test_repr(): hrepr = """Histogram( Regular(3, 0, 1), @@ -733,7 +767,7 @@ def test_pickle_bool(): assert_array_equal(a.view(), b.view()) -# Numpy tests +# NumPy tests def test_numpy_conversion_0(): @@ -1240,3 +1274,12 @@ def test_sum_empty_axis(): ) assert hist.sum().value == 0 assert "Str" in repr(hist) + + +# Issue 618 +def test_negative_fill(count_storage): + h = bh.Histogram(bh.axis.Integer(0, 3), storage=count_storage()) + h.fill(1, weight=-1) + + answer = np.array([0, -1, 0]) + assert h.values() == approx(answer) diff --git a/tests/test_minihist_title.py b/tests/test_minihist_title.py index d2042745..794b1f3a 100644 --- a/tests/test_minihist_title.py +++ b/tests/test_minihist_title.py @@ -41,6 +41,11 @@ def name(self): """ return tuple(ax.name for ax in self) + @name.setter + def name(self, values): + for ax, val in zip(self, values): + ax._ax.metadata["name"] = f"test: {val}" + # When you subclass Histogram or an Axes, you should register your family so # boost-histogram will know what to convert C++ objects into. @@ -152,3 +157,22 @@ def test_access(): assert hist_conv.axes["a"] == hist_conv.axes[0] assert hist_conv.axes["b"] == hist_conv.axes[1] + + +def test_hist_name_set(): + hist_1 = CustomHist(Regular(10, 0, 1, name="a"), Regular(20, 0, 4, name="b")) + + hist_1.axes.name = ("c", "d") + assert hist_1.axes.name == ("test: c", "test: d") + + with pytest.raises(AttributeError): + hist_1.axes[0].name = "a" + + hist_1.axes.label = ("one", "two") + assert hist_1.axes.label == ("one", "two") + + with pytest.raises(ValueError): + hist_1.axes.label = ("one",) + + with pytest.raises(ValueError): + hist_1.axes.label = ("one", "two", "three") diff --git a/tests/test_pickles.py b/tests/test_pickles.py index ca332740..b902169f 100644 --- a/tests/test_pickles.py +++ b/tests/test_pickles.py @@ -8,7 +8,7 @@ DIR = os.path.abspath(os.path.dirname(__file__)) -@pytest.mark.parametrize("version", ["0.10.2", "0.6.2", "0.11.1"]) +@pytest.mark.parametrize("version", ["0.10.2", "0.6.2", "0.11.1", "1.1.0"]) def test_read_pickle(version): filename = os.path.join(DIR, "pickles", f"bh_{version}.pkl") @@ -21,7 +21,9 @@ def test_read_pickle(version): h2 = d["h2"] h3 = d["h3"] - assert h1._storage_type == bh.storage.Double + assert h1._storage_type == ( + bh.storage.Double if version[0] == "0" else bh.storage.Int64 + ) assert h2._storage_type == bh.storage.Weight assert h3._storage_type == bh.storage.Double diff --git a/tests/test_views.py b/tests/test_views.py index c2f5c746..95c6cb95 100644 --- a/tests/test_views.py +++ b/tests/test_views.py @@ -65,9 +65,14 @@ def test_view_add(v): assert_allclose(v2.value, [2, 5, 4, 3]) assert_allclose(v2.variance, [4, 7, 6, 5]) - v += 2 - assert_allclose(v.value, [2, 5, 4, 3]) - assert_allclose(v.variance, [4, 7, 6, 5]) + v2 = v.copy() + v2 += 2 + assert_allclose(v2.value, [2, 5, 4, 3]) + assert_allclose(v2.variance, [4, 7, 6, 5]) + + v2 = v + v + assert_allclose(v2.value, v.value * 2) + assert_allclose(v2.variance, v.variance * 2) def test_view_sub(v): @@ -83,9 +88,14 @@ def test_view_sub(v): assert_allclose(v2.value, [1, -2, -1, 0]) assert_allclose(v2.variance, [1, 4, 3, 2]) - v -= 2 - assert_allclose(v.value, [-2, 1, 0, -1]) - assert_allclose(v.variance, [4, 7, 6, 5]) + v2 = v.copy() + v2 -= 2 + assert_allclose(v2.value, [-2, 1, 0, -1]) + assert_allclose(v2.variance, [4, 7, 6, 5]) + + v2 = v - v + assert_allclose(v2.value, [0, 0, 0, 0]) + assert_allclose(v2.variance, v.variance * 2) def test_view_unary(v):