Changelog for python311-awkward-2.6.2-1.2.noarch.rpm :

* Mon Mar 18 2024 Steve Kowalik - Update to 2.6.2:
* New features + feat: add CUDA kernels that calculate length/sum + feat: add variable length loop kernels + feat: make ak.mean and ak.count differentiable
* Bug-fixes and performance + fix: remove redundant(?) Jax.is_tracer_type check in _layout + fix: throw more meaningful error message for non-NumPy buffers + fix: proper typetracer array slicing in BitMaskedArray + fix: ak.flatten on BitMaskedArray + fix: performance of ufunc resolution for non-nominal signatures- Add rpmlintrc to silence some rpmlint warnings
* Tue Feb 06 2024 Daniel Garcia - Update to version 2.6.1: The purpose of this release is to provide a lower-bound on the fsspec dependency version (#3006).- 2.6.0: From this point onward, awkward strictly depends on fsspec.
* feat: turn on CUDA unit tests for working kernels and add some CUDA kernels (gh#scikit-hep/awkward#2930)
* feat: add string builder (gh#scikit-hep/awkward#2899)
* feat: add to_parquet_row_groups (gh#scikit-hep/awkward#2979)
* feat!: array is no longer allowed on NumpyArray and EmptyArray (gh#scikit-hep/awkward#2997)
* fix: include header-only files in sdist (gh#scikit-hep/awkward#2909)
* fix: touching of unions (gh#scikit-hep/awkward#2906)
* fix: lazily project union contents (gh#scikit-hep/awkward#2913)
* fix: support typetracer in to_packed for IndexedOptionArray (gh#scikit-hep/awkward#2912)
* fix: cartesian nested record (gh#scikit-hep/awkward#2929)
* fix: cast bool_ to int8 for positional reducers (gh#scikit-hep/awkward#2935)
* fix: ak.std and ak.var for axis != -1. (gh#scikit-hep/awkward#2918)
* Mon Jan 15 2024 Atri Bhattacharya - Update to version 2.5.2:
* feat: add some simpler cuda_kernels (gh#scikit-hep/awkward#2880).
* feat: add recursive flag to Content.to_packed (gh#scikit-hep/awkward#2933).
* feat: configure Julia version and required packages (gh#scikit-hep/awkward#2896).
* fix: include header-only files in sdist (gh#scikit-hep/awkward#2909).
* fix: touching of unions (gh#scikit-hep/awkward#2906).
* fix: lazily project union contents (gh#scikit-hep/awkward#2913).
* fix: support typetracer in to_packed for IndexedOptionArray (gh#scikit-hep/awkward#2912).
* fix: cartesian nested record (gh#scikit-hep/awkward#2929).
* fix: cast bool_ to int8 for positional reducers (gh#scikit-hep/awkward#2935).
* fix: ak.std and ak.var for axis != -1. (gh#scikit-hep/awkward#2918).- Bump required version of awkward-cpp to 28.
* Fri Dec 15 2023 Atri Bhattacharya - Update to version 2.5.1:
* feat: add enforce_concatenated_form (gh#scikit-hep/awkward#2860).
* feat: add Form.is_equal_to (gh#scikit-hep/awkward#2862).
* feat: Implement Indexed builder (gh#scikit-hep/awkward#2883).
* fix: ufunc handling of attrs (gh#scikit-hep/awkward#2837).
* fix: only compare scalars in full_like (gh#scikit-hep/awkward#2857).
* fix: Array.__getitem__ should set attrs on result (gh#scikit-hep/awkward#2866).
* fix: errors in existing cuda_kernels (gh#scikit-hep/awkward#2877).
* fix: register Numba in test (gh#scikit-hep/awkward#2879).
* fix: failing tests after #2877 (gh#scikit-hep/awkward#2886).
* fix: support revertable for concatenate in pyarrow logic (gh#scikit-hep/awkward#2889).- Bump required version of python-awkward-cpp to 27.
* Wed Nov 22 2023 Daniel Garcia - Update required version for python-awkward-cpp to 26.
* Sun Oct 08 2023 Atri Bhattacharya - Update to version 2.4.5: - feat: add recursive argument to expected_from_buffers (gh#scikit-hep/awkward#2724). - feat: infer unknown lengths from context in from_buffers (gh#scikit-hep/awkward#2732). - fix: don\'t use np.asarray on Index or Content objects (gh#scikit-hep/awkward#2740). - fix: swap names for __dlpack__ (gh#scikit-hep/awkward#2741). - fix: __array__ accepts positional arguments (gh#scikit-hep/awkward#2744).
* Sun Sep 17 2023 Atri Bhattacharya - Update to version 2.4.2:
* fix: support placeholder arrays in ArrayModuleNumpyLike.frombuffer (gh#scikit-hep/awkward#2693).
* fix: drop zero-cost views of ak.Array (gh#scikit-hep/awkward#2697).- Changes from version 2.4.1:
* Fix docs version selector- Changes from version 2.4.0:
* New features: - Form.expected_from_buffers for names/dtypes ak.from_buffers needs (gh#scikit-hep/awkward#2660). - support out-of-band buffers in pickling (gh#scikit-hep/awkward#2665). - add touch_data to ak.typetracer (gh#scikit-hep/awkward#2672). - add support for custom picklers (gh#scikit-hep/awkward#2682). - deprecate forget_length, add parameters to typetracer_with_report (gh#scikit-hep/awkward#2671). - remove deprecations for 1.4.0 release (gh#scikit-hep/awkward#2688).
* Bug fixes: - support placeholders in shape-only routines (gh#scikit-hep/awkward#2652). - multiple-output ufuncs e.g. divmod (gh#scikit-hep/awkward#2654). - support unflattening a typetracer-backed array at axis != 0 (gh#scikit-hep/awkward#2657). - fix IPython inspection (gh#scikit-hep/awkward#2658). - using numba.core.errors.Numba instead of Error in a Numba typing context. (gh#scikit-hep/awkward#2659). - raise error for invalid object in from_dlpack (gh#scikit-hep/awkward#2662). - fix/harden broadcasting through tuples (gh#scikit-hep/awkward#2663). - update string to search for in determing parquet column list separator (gh#scikit-hep/awkward#2670). - guard broadcast_and_apply from mixed backends (gh#scikit-hep/awkward#2678). - support typetracer in ak.str. operations (gh#scikit-hep/awkward#2679).- Requires python-awkward-cpp = 23.
* Mon Sep 11 2023 Dirk Müller - update to 2.3.3:
* feat: add support for expected termination of transform functions
* feat: export error classes
* feat: unpickle arrays made in Awkward v1 (as v2).
* feat: use pyarrow for string functions
* feat: add support for dictionary encoding from Arrow
* feat: add CPU kernel for to_numpy support for strings/bytestrings
* feat: add support for __arrow_array__
* feat: add __dlpack__, from_dlpack support
* Mon Jul 10 2023 Atri Bhattacharya - Update to version 2.3.1:
* Bug fix: Added kernel specializations for awkward_ListOffsetArray_reduce_local_nextparents_64 [gh#scikit-hep/awkward#2572].- Changes from version 2.3.0:
* New features: - LayoutBuilder in Numba [gh#scikit-hep/awkward#2408]. - Move string features into core [gh#scikit-hep/awkward#2547]. - Add new __list__ parameter for list-based array types [gh#scikit-hep/awkward#2549]. - Deprecate string behavior classes [gh#scikit-hep/awkward#2528]. - Apply retroactive deprecations [gh#scikit-hep/awkward#2566].
* Bug fixes and performance - support dict-likes in form.from_dict, convert fields to list [gh#scikit-hep/awkward#2548]. - check validity error properly [gh#scikit-hep/awkward#2550]. - string broadcasting [gh#scikit-hep/awkward#2552]. - make C++ LayoutBuilder API consistent with the Numba\'s [gh#scikit-hep/awkward#2553]. - convert unknown-length to index in ListOffsetArray.to_RegularArray [gh#scikit-hep/awkward#2560]. - prefer known to unknown lengths in broadcasting [gh#scikit-hep/awkward#2561]. - broadcasting of regular strings [gh#scikit-hep/awkward#2564]. - restore NEP-18 overloads for concatenate and where [gh#scikit-hep/awkward#2567].- Require python-awkward-cpp >= 21, python >= 3.8
* Sat Jul 01 2023 Dirk Müller - update to 2.2.4:
* feat!: canonicalise union of indexed
* feat: add array dispatcher (inline)
* fix: patch NDArrayOperatorsMixin
* fix: improve performance of Form.select_columns
* refactor: add internal with_operation_context
* Tue Jun 20 2023 Dirk Müller - update to 2.2.3:
* feat!: deprecate to_NumpyForm\'s ``dtype` argument
* feat!: deprecate the time __unit__ parameter
* fix: don\'t project records during broadcasting; push index down
* fix: protect ak.to_parquet against memory explosion when args are swapped.
* fix: str of KeyError for <3.11
* fix: support axis != -1 for record reduction
* Sun Jun 11 2023 Dirk Müller - update to 2.2.2:
* drop string broadcasting overloading
* add Index._touch_XXX methods
* add ak.typetracer.length_one_if_typetracer
* add placeholder array
* bugfixes and other changes see https://github.com/scikit-hep/awkward/releases/tag/v2.2.2
* Mon May 29 2023 Dirk Müller - update to 2.2.1:
* feat: add validity error for duplicate record field
* feat: add empty_if_typetracer & expose more of typetracer to L2
* feat!: remove use of typestr in ak.types.Type
* feat: improve performance of behavior lookups
* feat: support record reducer overloads
* fix: ak.to_numpy does not properly convert lists of empty string
* fix: handle indexed nodes in ak.almost_equal
* feat: canonicalise union[?X, Y] into union[?X, ?Y]
* feat: add ak.forms.from_type
* feat: add ak.enforce_type
* feat!: deprecate length_zero_array highlevel argument
* Bug-fixes and performance
* fix: avoid reference cycles with the error-handling machinery
* fix: proper unknown size broadcasting for all-regular
* Sat May 27 2023 Dirk Müller - add sle15_python_module_pythons
* Thu May 04 2023 Atri Bhattacharya - Update required version for python-awkward-cpp to 15.
* Fri Apr 28 2023 Atri Bhattacharya - Update to version 2.1.4:
* feat: raise error for out-of-bounds axis in ak.cartesian (gh#scikit-hep/awkward#2411).
* feat: make ak.numba.GrowableBuffer visible (gh#scikit-hep/awkward#2403).
* fix: unify C++ and Python GrowableBuffer::extend algorithm (gh#scikit-hep/awkward#2396).
* fix: return empty list if broadcasting no arrays (gh#scikit-hep/awkward#2407).
* fix: only convert NEP18 arguments to layouts if required (gh#scikit-hep/awkward#2409).
* fix: partially fix string broadcasting (gh#scikit-hep/awkward#2410).
* fix: correctly convert unknown scalar from kernel to a length (gh#scikit-hep/awkward#2417).
* fix: support unknown lengths when broadcasting unions (gh#scikit-hep/awkward#2418).
* fix: consistently set a resize default value to 8 - that is passed to GrowableBuffer (gh#scikit-hep/awkward#2421).
* fix: is_equal_to bug in UnionArray (gh#scikit-hep/awkward#2426).
* fix: handle reordered contents in ak.almost_equal (gh#scikit-hep/awkward#2424).
* fix: typetracer nplike.repeat & ak.with_field (gh#scikit-hep/awkward#2429).
* refactor: add custom axis error (gh#scikit-hep/awkward#2412).
* Sun Apr 23 2023 Dirk Müller - update to 2.1.3:
* fix: support empty records in ak.with_field by AATTagoose77 in #2387
* fix: only format exception for primary context
* fix: only touch data in nplike.asarray if copy is required
* fix: update GrowableBuffer::move_to algorithm
* fix: don\'t touch for ascontiguousarray
* feat: use cppyy for JIT
* feat!: rename cpptype
* feat: implemented GrowableBuffer in Numba as a start toward LayoutBuilder
* feat: add to_backend to ak.record.Record
* feat: add type equality is_equal_to member
* fix: expose array interface for CUDA
* fix: test untested path in ak.cartesian & broadcasting
* fix: ak.cartesian for typetracer
* fix: Numba string reference count memory leak.
* fix: generate an array view when an Array C++ type is requested
* fix: do not expose an rdfentry_ column
* fix: generate RDataSource API based on ROOT attribute
* fix: simplify depth limit test to support bigger-than-depth values
* fix: use cppyy include
* fix: convert ufunc arguments to same backend
* fix: support N-d sequences in TypeTracer.asarray
* fix: make from_iter require iterables!
* fix: avoid stack overflow with std::unique_ptr linked list
* fix: ak.unzip visits all contents
* fix: remove spurious typetracer conversion
* feat: add support for arrays in from_buffers
* feat: validate axis in L1
* feat: add examples to header-only module
* fix: support typetracer in ak.unflatten
* fix: view for unknown scalar
* fix: common backend in functions accepting multiple arrays
* fix: support unknown scalars in __getitem__
* fix: fail early if index is too dimensional
* fix: uneccessary use of nplike_of
* fix: jupyterlite dependency changes
* Tue Mar 14 2023 Dirk Müller - update to 2.1.0:
* bump numpy to 1.17.0
* Tue Mar 07 2023 Dirk Müller - update to 2.0.9:
* feat: keep column order after filtering flag
* feat: Awkward Arrays in Numba\'s CUDA target
* feat: add \'highlevel\' and \'behavior\' arguments to \'ak.from_rdataframe\' function
* fix: set dtype in full_like
* fix: use simplified in ak._do.merge_as_union
* fix: support unknown content length in to_packed
* fix: run-lengths with typetracer
* fix: add trivial nan_to_num impl
* fix: add ak.broadcast_fields
* fix: keep column order in all tests
* Wed Mar 01 2023 Dirk Müller - update to 2.0.8:
* allow awkward type arrays filtering based on rdfentry
* feat!: re-introduce unknown-length
* fix: keep EmptyArray in remove_structure
* fix: add is_c_contiguous method to Cupy
* fix: boolean indexing with non-zero starting offsets
* fix: add license to _backends
* fix: use of size in to_RegularArray
* fix: indexing into RegularArray with typetracer
* fix: support length-zero outer arrays in almost_equal
* fix: support options in ak.merge_union_of_records
* fix: don\'t merge non-union parameters
* fix: boolean slicing with non-packed arrays
* refactor: add type hints for Content._getitem_XXX
* refactor: drop NumpyLike.known_shape
* refactor: change Content._getitem_range to use explicit indices
* docs: fix example for merge_union_of_records
* docs: add 1.10 legacy docs to switcher.json
* docs: reintroduce Content documentation from v1 reST files.
* docs: add how-to-examine-type.md
* docs: improve ragged indexing docs
* chore: include cuda kernel tests in sdist
* chore(deps): bump amannn/action-semantic-pull-request from 5.0.2 to 5.1.0
* chore: update pre-commit hooks
* Mon Feb 13 2023 Steve Kowalik - awkward 2.0.7 actually requires awkward-cpp 8.
* Fri Feb 10 2023 Dirk Müller - update to 2.0.7
* feat: add ability to forget length of typetracer created with typetracer_from_report
* feat: start hardening nplike signatures
* feat: implement all ufuncs on TypeTracer.
* feat: use None for unknown lengths (1 of 2)
* feat: coerce backends to same zero-copy type (2 of 2)
* feat: growable buffer move_to method
* feat: add ak.merge_union_of_records
* feat: add support for histogram module
* feat: add ak.approx_equal
* fix: re-order cases to handle NumPy scalar types properly
* fix: specify dtype for buffers in from_rdataframe.
* fix: unify typestr with _repr
* fix: update type_to_name for Layout builder cxx_14 target
* fix: Layout builders clean and length bug fixes
* fix: support mask_identity=True for axis=None in ptp, std, etc.
* fix: preserve dimensions for keepdims=True, axis=None reductions
* fix: some usages of len(layout) under typetracer
* fix: rdataframe memory check
* fix: rework parameter merging rules
* fix: don\'t raise NotImplementedError when reading empty array from Parquet
* fix: ignore object arrays
* fix: ak.values_astype now turns \'unknown\' type into the requested type.
* Fri Jan 27 2023 Ben Greiner - Update to 2.0.6
* feat: expose typetracer in public backend API by AATTagoose77 in #2066
* feat: add byteorder argument to to_buffers by AATTagoose77 in #2095
* feat: add exception for missing field by AATTagoose77 in #2120- Release 2.0.3 and 2.0.4
* The flatten_records argument of all reducers (ak.all, ak.any, ..., ak.var) has effectively been removed: setting it now raises an error (PR #2020). This argument applies a reducer to all contents of a record, merging fields, and it had to be removed to properly implement axis=None. The old default, flatten_records=False, is now the only behavior, and to get the equivalent of flatten_records=True, you can use ak.ravel: ak.sum(array, flatten_records=True) becomes ak.sum(ak.ravel(array))
* Removing the feature is still the right thing to do, but the function argument needs to go through a deprecation cycle, since libraries like dask-awkward pass arguments through to Awkward. Removing flatten_records as an argument introduces an error, even if the surviving case of flatten_records=False is desired.- Release 2.0.2
* feat: add ak.drop_none() by AATTioanaif in #1904- Release 2.0.1
* feat: add ak.without_field by AATTagoose77 in #1963- Release 2.0.0
* The Awkward Array version 2 project started in June of 2021 and has been developed alongside version 1 updates. For most of that time, it was available as a submodule, awkward._v2, so that it could be tested with the same tests as version 1 and could be experimented upon by early adopters.
* Despite the long list of PRs, the high-level interface changes from version 1 to version 2 were kept at a minimum. For the most part, the Awkward 1.x API is fine, but the internal implementation needed an overhaul to prevent technical debt. [#]# Summary of changes
* Nearly all of the code is written in Python now. Exceptions are the \"kernel\" functions, GrowableBuffer, LayoutBuilder, ArrayBuilder, AwkwardForth, and dynamically generated C++ code for RDataFrame.
* Maintains performance because any algorithms that scale with the size of an array are implemented in compiled \"kernel\" functions.
* Split into two packages: awkward-cpp for the C++ part (infrequently updated, binary distribution for most platforms and Python versions) and awkward, the Python part (frequently updated).
* Virtual arrays and Partitions (collectively, \"lazy arrays\") have been removed in favor of dask-awkward.
* Awkward Arrays can be converted to and from RDataFrame, generating C++ for ROOT to JIT-compile so that iteration over Awkward Array input is fast (adapted from the Numba implementation).
* Auto-differentiation of functions on Awkward Arrays using JAX. (But not JAX JIT-compilation.)
* Suite of header-only C++ that does not depend on Awkward Arrays, but can be used to produce them and quickly get them from C++ to Python. The header-only suite includes GrowableBuffer and LayoutBuilder.
* New documentation website (https://awkward-array.org/), based on JupyterBooks, the NumPy/SciPy/Pandas style and organization, as well as a notebook that can be executed in your web browser.
* More expressive error-messages, highlighting the ak.
* function that was in progress when the error occurred, with its arguments. (That is, highlighting ak.
* functions as the granularity of feedback to users of Awkward Array, rather than making you search through the stack trace to the hand-off from your code to ours.)
* Brackets are always balanced in the console representation of an array: >>> ak.Array([ ... [{\"x\": 1.1, \"y\": [1]}, {\"x\": 2.2, \"y\": [1, 2]}], ... [], ... [{\"x\": 3.3, \"y\": [1, 2, 3]}], ... ]) * var
* {x: float64, y: var...\'> as opposed to * var
* {\"x\": float64, \"y...\'> in version 1. Also, show methods for values [[{x: 1.1, y: [1]}, {x: 2.2, y: [1, 2]}], [], [{x: 3.3, y: [1, 2, 3]}]] and types 3
* var
* { x: float64, y: var
* int64 } This extended show output is the default representation in Jupyter.
* Round-trip fidelity in ak.to_arrow/ak.from_arrow: no Awkward Array metadata is lost. Same for ak.to_parquet/ak.from_parquet, to the extent that pyarrow can read and write Parquet.
* Parquet column selection using wildcards.
* Data exported with version 1 ak.to_buffers can be imported by version 2 ak.from_buffers, with custom buffer_keys.
* The majority of version 1 tests have been ported to version 2, to ensure that the interface and functionality doesn\'t change, except where intended (e.g. organizing naming conventions).
* Consistent handling of date-time and time-delta types (matches NumPy\'s system).
* Improved ak.to_json/ak.from_json arguments (for converting non-JSON types NaN, infinity, complex numbers) and using a known JSONSchema to accelerate ak.from_json. Removed ambiguities about newline-delimited JSON (requires explicit argument).
* The world\'s fastest Avro file reader in Python, ak.from_avro_file (uses AwkwardForth).
* \"nan\" versions of NumPy functions, such as np.nansum, np.nanmean, np.nanstd.
* Renamed ak.to_pandas → ak.to_dataframe, to clarify distinction from awkward-pandas.
* Organized Type and Form objects better, more consistent.
* Clear specification of NumPy dtypes that can be used in Awkward Arrays (bool, numbers, including complex, and date-time/time-delta).
* Organized naming conventions throughout the codebase, such as keys versus fields versus recordlookup.
* Carefully examined the public API (all modules, functions, classes, and methods that don\'t start with an underscore) to be sure that we can support it going forward. Any removal or change of an interface will require a minor version number increase and a deprecation cycle, on the order of months. (New features and bug-fixes can be immediate, on patch releases.)
* Flags and \"configuration\" function arguments are now keyword-only (order independent).
* Started adding Python type hints (nowhere near complete, but started).
* Removed the Identities from array nodes. They were never fully implemented—a placeholder for a feature that won\'t be developed within Awkward Array (SQL-style JOINs).
* TypeTracerArray does a \"dry run\" of a calculation to predict its type at the end. Used to build a computation graph for dask-awkward.
* Equivalent but ungainly type combinations, such as \"option-type of option-type of X\" or \"union-type containing union-types,\" have been outlawed with tools to squash them into a canonical layout. Operations on the data now have fewer possibilities to worry about.
* Simplified the semantics of nbytes.
* Clarified ak.ravel and ak.flatten\'s treatments of missing data.
* Added missing ArrayBuilder methods in Numba.
* Set up framework for performing ak.
* operations in CUDA, using CuPy JIT-compilation to get the right interface, with error message passing through asynchronous GPU computations. The CUDA implementation is not complete, but the groundwork has been laid.
* Parquet reading/writing uses fsspec for remote connections and handles the dataset protocol correctly.
* Public interface to our power-tools: ak.transform unifies our internal recursively_apply and broadcast_and_apply.
* All ak.
* functions were rewritten; some were redesigned in a simpler way using the above-mentioned power tools. There are likely fewer undiscovered bugs in these functions now.
* Use of records in numerical functions is more restricted, to prevent misapplication of ufuncs like + and reducers like np.sum on meaningful records such as Vectors.
* All docstrings updated for version 2, including the code examples.- Drop awkward-cmake-build-with-RelWithDebInfo.patch: No longer compiling here- Remove the devel package: The header-only files are now provided by awkward-devel built from python-awkward-cpp
* Fri Nov 11 2022 Daniel Garcia - Rvert the devel package requirement, python-awkward is not provided by any of subpackages, the real package provided is python3-awkward.
* Thu Nov 10 2022 Daniel Garcia - Remove not needed tests filter- Improve devel package requirement to use the package name variable instead of python3-awkward- Update to 1.10.2: [#] New features
* feat: add RegularArray._reduce_next implementation (backport) by AATTagoose77 in #1813 [#] Bug-fixes and performance
* fix: don\'t assume trailing . for module name in is_XXX_buffer (backport) by AATTagoose77 in #1746
* fix: use proper lengths in ByteMaskedArray.mergemany (backport) by AATTagoose77 in #1750
* fix: simplify ListOffsetArray_reduce_nonlocal_outstartsstops (backport) by AATTagoose77 in #1797 [#] Other
* chore: remove v2 Python highlevel LayoutBuilder from main-v1. by AATTjpivarski in #1863
* Fri Nov 04 2022 Daniel Garcia - Update tests to skip broken tests in i586. gh#scikit-hep/awkward#1862- Remove python_module macro definition
* Mon Oct 17 2022 Steve Kowalik - Update to 1.10.1:
* fix: support non-hashable parameter values in broadcasting
* feat: raise Error for Record.setattr
* fix: allow string arrays to merge without unions
* fix: carry parameters through broadcasting
* fix: don\'t project categorical in ak._v2.packed
* fix: reduce runtime dependency from setuptools to just packaging
* fix: ensure that __copy__ and __deepcopy__ are enabled.
* fix: pass memo to deepcopy
* perf: improve Array initialisation performance
* fix: Removed bytemask() in favour of mask_as_bool()- Add importlib-resources to BuildRequires.- Drop patches:
* awkward-correct-includedir.patch
* awkward-import-packging-not-setuptools_external.patch- No longer attempt to skip 3.6.
* Wed Aug 24 2022 Steve Kowalik - Copy the heaers ourselves now that setuptools does not.
* Sat Mar 05 2022 Atri Bhattacharya - Update to version 1.8.0:
* Bumped pybind11 to 2.9.1 [gh#scikit-hep/awkward-1.0#1316].
* Added optiontype_outside_record option to ak.zip [gh#scikit-hep/awkward-1.0#1308].
* Fixed typo in ArrayBuilder documentation [gh#scikit-hep/awkward-1.0#1326].
* Many more changes through various 1.8.0rc versions, see https://awkward-array.readthedocs.io/en/latest/_auto/changelog.html.- Add awkward-import-packging-not-setuptools_external.patch: Directly use packaging module instead of setuptools.extern; the latter is used by upstream to keep dependencies to a minimum, but importing setuptools and trying to then use setuptools.external does not work on openSUSE.- Add python-packaging BuildRequires, needed for tests in light of above patch, and Requires.
* Wed Dec 08 2021 Atri Bhattacharya - Update to version 1.7.0:
* Fixed a division-by-zero bug: [gh#scikit-hep/awkward-1.0#1168, gh#scikit-hep/awkward-1.0#1169].
* Applied the 1.7.0 deprecation (see Roadmap): [gh#scikit-hep/awkward-1.0#1118].
* Enabled codecov: [gh#scikit-hep/awkward-1.0#1120].
* Updated Azure\'s Windows VM: [gh#scikit-hep/awkward-1.0#1129].
* Fixed order in RecordForm.contents property: [gh#scikit-hep/awkward-1.0#1166].
* Fixed Parquet reading for files with zero RecordBatches: [gh#scikit-hep/awkward-1.0#1170].
* Sat Sep 18 2021 Atri Bhattacharya - Update to version 1.5.0:
* Reverting the spelling of ak.nplike.NumPy → ak.nplike.Numpy, etc. (including CuPy → Cupy) to avoid needlessly breaking interfaces.
* Unrestricted the jaxlib version bounds, thanks to a fix that he prompted in JAX: PR gh#scikit-hep/awkward#988.
* Added support for n-dimensional NumpyArrays in ak.where: PR gh#scikit-hep/awkward#980.
* Added axis parameter to ak.fill_none: PR gh#scikit-hep/awkward#917. Added np.ravel: PR gh#scikit-hep/awkward#985.
* Fixed ak.from_buffers to support empty buffers: PR gh#scikit-hep/awkward#1009.
* Fixed ak.packed for RegularArrays with size=0: PR gh#scikit-hep/awkward#1008.
* Fixed ak.zip to not convert regular to jagged: gh#scikit-hep/awkward#1013.
* Clarified the documentation of left and right broadcasting: PR gh#scikit-hep/awkward#1016.
* Fixed ak.broadcast_arrays by setting numpy_to_regular=True: PR gh#scikit-hep/awkward#1019.
* Added support for ak.Record in ak.packed: PR gh#scikit-hep/awkward#1021.
* Fixed gh#scikit-hep/awkward#1026, jagged slicing of a multidimensional NumpyArray: PR gh#scikit-hep/awkward#1029.
* Added a name parameter to ak.mixin_class: PR gh#scikit-hep/awkward#1030.
* Added a spell-checker in pre-commit that fixed a lot of spelling mistakes in comments and documentation: PR gh#scikit-hep/awkward#1003.
* Fixed the stack level of the deprecation warning: PR gh#scikit-hep/awkward#1058.
* Made the source code reproducible by allowing the build date to be overridden by SOURCE_DATE_EPOCH: PR gh#scikit-hep/awkward#1050.
* Optimized common array slice (\"take\") operations: PR gh#scikit-hep/awkward#1062.
* Added a better set of kernel unit tests to continuous integration, which will be important for testing GPU kernels: PRs gh#scikit-hep/awkward#1065 and gh#scikit-hep/awkward#1081.
* Added a default for NEP-18, so now any NumPy function can be used with Awkward Arrays, as long as the Awkward Array can be cast as NumPy (i.e. is rectilinear): gh#scikit-hep/awkward#1005.
* Reverted the ill-conceived gh#scikit-hep/awkward#694, SliceVarNewAxis: PR gh#scikit-hep/awkward#1028.
* Made ak.Record iterable (iterates over field names, like dict): PR gh#scikit-hep/awkward#1061.
* Fixed the UnionForm bug described in gh#scikit-hep/awkward#770 and gh#scikit-hep/awkward#930: PR gh#scikit-hep/awkward#991.
* Fixed gh#scikit-hep/awkward#546, ak.fill_none losing the replacement\'s dtype: PR gh#scikit-hep/awkward#995.
* Fixed double-masking of reducers and unmasking of ak.ptp: PR gh#scikit-hep/awkward#994.
* Fixed ak.Array.__dir__ to include methods/properties of overridden classees: PR gh#scikit-hep/awkward#993.
* Fixed gh#scikit-hep/awkward#982 by accounting for another kind of \"gap\" in nonlocal reducers: PR gh#scikit-hep/awkward#997.
* Fixed gh#scikit-hep/awkward#998 and gh#scikit-hep/awkward#1000, ak.argmax for ListOffsetArray with nonzero start: PR gh#scikit-hep/awkward#1001.
* Fixed ak.fill_none fill value\'s handling of NumPy dimension: PR gh#scikit-hep/awkward#1056.
* Fixed ak.to_numpy for structured arrays that are also masked arrays: PR gh#scikit-hep/awkward#1067.
* Finished __getitem__ for all array types: PRs gh#scikit-hep/awkward#959, gh#scikit-hep/awkward#1031, gh#scikit-hep/awkward#1036, and gh#scikit-hep/awkward#1073.
* Refactored ArrayBuilder to not depend on C++ that will be removed in v2: PRs gh#scikit-hep/awkward#977, gh#scikit-hep/awkward#1045, and gh#scikit-hep/awkward#1063.
* Implemented _localindex in v2: PR gh#scikit-hep/awkward#1059. And combinations: PR gh#scikit-hep/awkward#1074.
* Renamed _getitem_array as _carry and made it take an Index: PR gh#scikit-hep/awkward#1025. Added VirtualArray: PR gh#scikit-hep/awkward#1079.- Disable python2 builds explicitly by setting skip_python2 in specfile; otherwise builds for openSUSE <= 1550 fail.
* Sat Jul 03 2021 Atri Bhattacharya - Update to version 1.4.0:
* Long list of changes since last packaged version, see https://awkward-array.readthedocs.io/en/latest/_auto/changelog.html.- Adapt symlinks and file-list for changes to shared object locations (directly in python_sitearch instead of python_sitearch/awkward/).- Skip a test on 32-bit that uses float128, since float32 is not available.
* Sun Feb 21 2021 Ben Greiner - Update to 1.1.2
* The fix that motivates this patch release is AATThenryiii’s repair of the SDist distribution: PR #720. The addition of dlpack wasn’t included in the source distribution, so pip install awkward without an available wheel was failing due to not having all the headers it needed to compile.
* Also fixed jagged slices: issue #723, PRs #725 and #729, as well as a segfault in ak.flatten: issue #724, PR #726.- Release 1.1.1
* New patch release motivated by a bug discovered in Uproot’s tests: PR #719. (From now on, Uproot will test Awkward release candidates.)
* Also, #717 adds an optional behavior argument to every function that takes a highlevel=True/False.
* Also from now on, only critical bug-fixes (like this one) will be patch releases and other updates will all go into release candidates for the next non-patch release (e.g. 1.2.0, [scheduled for 2021-04-01](https://github.com/scikit-hep/ awkward-1.0#roadmap)).- Release 1.1.0
* Updates to Arrow 3.0. PR #709.
* Can now read and write Parquet datasets (conventions for sets of multiple files): PR #706, fixes #368 (old!).
* Also, Python 2 can use ellipsis because we have pybind11 2.6 now: PR #707, fixes #705.
* remove deprecated features
* Fixed #689, the dimension of arrays returned by empty slices: PR #691.
* AATTianna fixed tuple-building in ArrayBuilder to guard against specifying a nonexistent slot and segfaulting: #697.
* Added [ak.materialized](https://awkward-array.readthedocs.io/ en/latest/_auto/ak.materialized.html) to force-materialize arrays (opposite of [ak.virtual](https://awkward-array. readthedocs.io/en/latest/_auto/ak.virtual.html)) because debugging very lazy deferments is hard.
* Can now use np.newaxis to give one array the same number of dimensions as another for slicing (which closes #492).
* Fixed ak.min/ak.max with axis != None for partitioned data.
* Fixed a bug that affected boolean-masking of partitioned data.
* Reducing over-materialization of virtual arrays: PRs #643, [#644], #657, added Form-aware lazy Parquet: #688
* Fixed ak.from_buffers to allow unreachable elements: PR #673
* Developed the AwkwardForth VM: PRs #683, #648, #653, #661, [#664], #672
* Removed right-broadcasting from all operations except those that require it for NumPy compliance: #656, #680
* AATTianna improved sorting: PRs #639, #687, added complex numbers support: PRs #652
* AATTtrickarcher generalized CuPy interop to JAX: PRs #645, #662
* AATTreikdas cleaned up developer scripts: PRs #665, #666, #667
* AATThenryiii added more code quality tests: PRs #675, #681, #682, [#677], #676, #685, #683- Somehow the missing python-devel BuildRequires was okay until recently.
* Sun Feb 14 2021 Dirk Müller - skip python 3.6 (numbas requires scipy)
* Mon Jan 25 2021 Ben Greiner - Update to 1.0.2
* AATTnsmith- fixed virtual arrays in ak.to_arrow (PR #642).
* Fixed #628, allowing np.int64(N) to be used in a slice like integer N (PR #636).
* Added the ability to select fields inside of nested records with syntax like array[[“pt”, “eta”, “phi”], “nominal”] (#619).
* Fixed Apache Arrow → Awkward Array (ak.from_arrow) for sliced Arrow arrays: it has an offset attribute we had been ignoring (#625). Thanks to AATTtrickarcher for the extensive tests!
* Ensured that an ak.Array’s behavior always gets passed to output of any operation (#626), even if the input is a dict (#627, affected ak.zip, ak.cartesian, and ak.argcartesian).
* Fixed #629, ak.with_field dropping parameters (#631), and #634, ak.fill_none not accepting a Record as a fill value (#635).
* Added a roadmap and upvote/thumbs up instructions to the README.
* AATTjpivarski revised the ak.concatenate implementation (#610), fixed a bug related to axis handling (#612), and made ak.concatenate friendlier to lazy arrays (#613; “friendlier” = “materializes arrays in fewer cases”).
* AATTchrisburr added setuptools as a requirement because it is used for pkg_resources.
* AATTianna added validity-checking for strings, bytestrings, and categorical (#571), including some revisions of sorting and a partial implementation of checking for uniqueness and identifying unique elements from arrays that can later be used to implement np.unique.
* Updated Awkward <–> Arrow conversion to preserve nullability in type (PRs #602, #606). Preemptively avoiding warnings from NumPy 1.20 (#599) and fixed 32-bit Linux cases in the tests (#604).
* AATThenryiii fixed hardcoded threads in MacOS compilation target (for deploying to conda-forge).- Drop awkward-tests-on-32bit.patch merged upstream- Fix python_expand for primary flavor check