dev-python/pyarrow: Bump to 23.0.0

Signed-off-by: Michał Górny <mgorny@gentoo.org>

2 files changed, 110 insertions, 0 deletions

diff --git a/dev-python/pyarrow/Manifest b/dev-python/pyarrow/Manifest
index 72a6e8918261..54b219d734e8 100644
--- a/dev-python/pyarrow/Manifest
+++ b/dev-python/pyarrow/Manifest
@@ -1,6 +1,7 @@
 DIST apache-arrow-20.0.0.tar.gz 17862661 BLAKE2B 154d4783746abf21ee7e94748cee1f5b640e24d756b08067e3c1135449b907e1888728eff76002e01f405c0d14fd01c76ce4b9de0e49d1116c8216ae8ae297f7 SHA512 067e62d7d311cebfca43473b0aacaacc534da47c3450b75328517df69281d8be1e79b0430cc7e975eb613e05c62a62d6ca92a4c4f7882ae7733f826d774d9081
 DIST apache-arrow-21.0.0.tar.gz 17320680 BLAKE2B 56c3bd23ceb59346e094ed83601f30187d88b9148118a83ce413ba525fec23da84ded4576eeefcc1786ecc14e46b5fd0b6270e96bad9659d25f154d77731124e SHA512 89da6de7eb2513c797d6671e1addf40b8b156215b481cf2511fa69faa16547c52d8220727626eeda499e4384d276e03880cd920aaab41c3d15106743d51a90a6
 DIST apache-arrow-22.0.0.tar.gz 17063271 BLAKE2B 99b53160b209a9427500fb5bda515c46256f7a6498c80627ff930af1690ffe3a01c86c161fd1946f3b249f1ceaefad0da6c05ea88efbb43b5733c9eb98b5cc1f SHA512 8ec9ddaf7917c0e35c8bb32831fe6ea0a7b81de5723828a1289ba1b9e104b42af688d0f427a0ceff6f617d5f7ac67769431184b137e54f6987779e467c59d3ec
+DIST apache-arrow-23.0.0.tar.gz 17287604 BLAKE2B 588881dbb434ed7cbdea2fe14f53c7f42ee75d9a4433a4deaad34dff33b989473c1ae7a0a5d6d806ce0e515548cfe515fb69c988536f5419062e344fc6634f6b SHA512 ddaaf31e7f3f8246ce3aa6e77c812c55a7993d037aaebc8ca9d79d54c701717a135eebfede731a3239becfaed2fa2395fcb3d68b14c32b9de72b3f5598d9df4a
 DIST arrow-testing-725fd4a4b12d01c53c98e80274c0b23aa8397082.tar.gz 3752144 BLAKE2B bbfde361f8da376f228720de00a3f6c2b42051e53580f650ad5621086cbf4828a289d86815ac1143720d80f230a851c9602bf198a32e457c1a0f7767790968ee SHA512 0789bf65fc2609f84a9c2865bc8e12f8c63fdd138eff02af1b24ba2c0e925d8c0ecfbe57515786e91531d918ef1a0d92fd5becb553682eafdcedd7678ddacb9e
 DIST arrow-testing-d2a13712303498963395318a4eb42872e66aead7.tar.gz 3572581 BLAKE2B 58e0e0a887efb9f68b9cccc37b946352eb382fe6f621c0778992ec3e1d10cdd5de89674f8282ceb90d0fcef5825b24ff3d0043806c6e353be9df6d1566699d2d SHA512 e6b76e4ab34744bc804ecfe5baba3e1f483c874f883cc44e8e2504a718b847cb05f0c7559e40a446deab99da20d1f9bbfe03d553ee2bb5543bef889c2b555259
 DIST parquet-testing-18d17540097fca7c40be3d42c167e6bfad90763c.tar.gz 1077258 BLAKE2B 457bf3404b9f6ad8da5cbecade4c2365e39e6409c5b551092a7f1f18f81f557e94e60adfb381058c5fd3a4000dae24c1ffebfbfa965d88e81f098f016bfdc816 SHA512 2cadc694e6cd2e52869b4ff0d2659c77c613fc870b54ca7f4cb64068d0e7956723549b6d5513a117a2e147d4161e2fcdbe15992a3d8e56306b25923676f182a0
diff --git a/dev-python/pyarrow/pyarrow-23.0.0.ebuild b/dev-python/pyarrow/pyarrow-23.0.0.ebuild
new file mode 100644
index 000000000000..b91f7a7c2aa3
--- /dev/null
+++ b/dev-python/pyarrow/pyarrow-23.0.0.ebuild
@@ -0,0 +1,109 @@
+# Copyright 2023-2026 Gentoo Authors
+# Distributed under the terms of the GNU General Public License v2
+
+EAPI=8
+
+DISTUTILS_EXT=1
+DISTUTILS_USE_PEP517=setuptools
+PYTHON_COMPAT=( python3_{11..14} )
+
+inherit distutils-r1 multiprocessing
+
+# arrow.git: testing
+ARROW_DATA_GIT_HASH=725fd4a4b12d01c53c98e80274c0b23aa8397082
+# arrow.git: cpp/submodules/parquet-testing
+PARQUET_DATA_GIT_HASH=a3d96a65e11e2bbca7d22a894e8313ede90a33a3
+
+DESCRIPTION="Python library for Apache Arrow"
+HOMEPAGE="
+	https://arrow.apache.org/
+	https://github.com/apache/arrow/
+	https://pypi.org/project/pyarrow/
+"
+SRC_URI="
+	mirror://apache/arrow/arrow-${PV}/apache-arrow-${PV}.tar.gz
+	test? (
+		https://github.com/apache/parquet-testing/archive/${PARQUET_DATA_GIT_HASH}.tar.gz
+			-> parquet-testing-${PARQUET_DATA_GIT_HASH}.tar.gz
+		https://github.com/apache/arrow-testing/archive/${ARROW_DATA_GIT_HASH}.tar.gz
+			-> arrow-testing-${ARROW_DATA_GIT_HASH}.tar.gz
+	)
+"
+S="${WORKDIR}/apache-arrow-${PV}/python"
+
+LICENSE="Apache-2.0"
+SLOT="0"
+KEYWORDS="~amd64 ~arm64 ~riscv ~x86"
+IUSE="+parquet +snappy ssl"
+
+RDEPEND="
+	~dev-libs/apache-arrow-${PV}[compute,dataset,json,parquet?,re2,snappy?,ssl?]
+	>=dev-python/numpy-1.16.6:=[${PYTHON_USEDEP}]
+"
+BDEPEND="
+	test? (
+		dev-python/cffi[${PYTHON_USEDEP}]
+		dev-python/packaging[${PYTHON_USEDEP}]
+		dev-python/pandas[${PYTHON_USEDEP}]
+		dev-python/pytz[${PYTHON_USEDEP}]
+		dev-libs/apache-arrow[lz4,zlib]
+	)
+"
+
+EPYTEST_PLUGINS=( hypothesis )
+EPYTEST_XDIST=1
+distutils_enable_tests pytest
+
+src_prepare() {
+	distutils-r1_src_prepare
+
+	# cython's -Werror
+	sed -i -e '/--warning-errors/d' CMakeLists.txt || die
+}
+
+src_compile() {
+	export PYARROW_PARALLEL="$(makeopts_jobs)"
+	export PYARROW_BUILD_VERBOSE=1
+	export PYARROW_CXXFLAGS="${CXXFLAGS}"
+	export PYARROW_BUNDLE_ARROW_CPP_HEADERS=0
+	export PYARROW_CMAKE_GENERATOR=Ninja
+	export PYARROW_WITH_HDFS=1
+	if use parquet; then
+		export PYARROW_WITH_DATASET=1
+		export PYARROW_WITH_PARQUET=1
+		use ssl && export PYARROW_WITH_PARQUET_ENCRYPTION=1
+	fi
+	if use snappy; then
+		export PYARROW_WITH_SNAPPY=1
+	fi
+
+	distutils-r1_src_compile
+}
+
+python_test() {
+	local EPYTEST_DESELECT=(
+		# wtf?
+		tests/test_fs.py::test_localfs_errors
+		# these require apache-arrow with jemalloc that doesn't seem
+		# to be supported by the Gentoo package
+		tests/test_memory.py::test_env_var
+		tests/test_memory.py::test_specific_memory_pools
+		tests/test_memory.py::test_supported_memory_backends
+		# require mimalloc
+		tests/test_memory.py::test_memory_pool_factories
+		# fragile memory tests
+		tests/test_csv.py::TestSerialStreamingCSVRead::test_batch_lifetime
+		tests/test_csv.py::TestThreadedStreamingCSVRead::test_batch_lifetime
+		tests/test_array.py::test_to_pandas_zero_copy
+		# takes forever, and manages to generate timedeltas over 64 bits
+		tests/test_strategies.py
+		"tests/test_array.py::test_pickling[builtin_pickle]"
+		# scipy.sparse does not support dtype float16
+		"tests/test_sparse_tensor.py::test_sparse_coo_tensor_scipy_roundtrip[f2-arrow_type8]"
+	)
+
+	cd "${T}" || die
+	local -x PARQUET_TEST_DATA="${WORKDIR}/parquet-testing-${PARQUET_DATA_GIT_HASH}/data"
+	local -x ARROW_TEST_DATA="${WORKDIR}/arrow-testing-${ARROW_DATA_GIT_HASH}/data"
+	epytest --pyargs pyarrow
+}