Adding metadataHEADdevelop

5 files changed, 340 insertions, 0 deletions

diff --git a/dev-python/pyarrow/Manifest b/dev-python/pyarrow/Manifest
new file mode 100644
index 000000000000..71f0ce798d6f
--- /dev/null
+++ b/dev-python/pyarrow/Manifest
@@ -0,0 +1,6 @@
+DIST apache-arrow-23.0.1.tar.gz 17296212 BLAKE2B 7023df6e6900dc239edc757fef76a89d2169e4fd3e1eaf3471f805a6e1c39014b9ec1e5a3d7c9e7a590b8d007c4adc1d15b8ed22ecf4724b24a33ffa96557db2 SHA512 c687e50dfcdbf7e0e39710224360d35d9aa734452b3a47adc8c101f3019b6b4116310c05b9f3cd0a5ed4ad9b7bd8fb88edb70e79b3cbd413a57e5e35e4554a6c
+DIST apache-arrow-24.0.0.tar.gz 17360520 BLAKE2B 1d50493b3b74e1c16ba406b5cea93712a332b8ea2db405c31aca844a8fa2d6ef9bcecf3a3bc84dc4bc6ddf946e55251e6d270c85c15b7f461765b6fc72b3a492 SHA512 ffbac0c198c74f26b2ac173aa2b054cfd3902927ca3db36fbbae25353e8696c32b744e805c25801ced22641af71a5a8ea441202d8d3494aa803b797995d9d408
+DIST arrow-testing-249079a810caedda6898464003c7ef8a47efeeae.tar.gz 4155027 BLAKE2B 4fc4cf5153382b476d66265d5ab2fd5e512fdd0766bcf94bf675e5fefd15523fc8e0f87d9708285fd94e59b8d2a895f4f70855f0e1a826587897bef303887584 SHA512 a7134f3a4dc9c1ce09c1b522ee38e1cb9e04aef08e32bdd18297b9b3a73971f632b97d18f955e59e1cefa38f7f6417a9a2bc2fef787f5174ddcd71733305d2d3
+DIST arrow-testing-725fd4a4b12d01c53c98e80274c0b23aa8397082.tar.gz 3752144 BLAKE2B bbfde361f8da376f228720de00a3f6c2b42051e53580f650ad5621086cbf4828a289d86815ac1143720d80f230a851c9602bf198a32e457c1a0f7767790968ee SHA512 0789bf65fc2609f84a9c2865bc8e12f8c63fdd138eff02af1b24ba2c0e925d8c0ecfbe57515786e91531d918ef1a0d92fd5becb553682eafdcedd7678ddacb9e
+DIST parquet-testing-a3d96a65e11e2bbca7d22a894e8313ede90a33a3.tar.gz 1150001 BLAKE2B 2c991bca5679ba25767fea1c9e17991ec7661e302adc55bf353a9d7b1c3a91269a9d464cc15743ab0a056591939ed43f0d500ef51275d7d10cb96055c773d44c SHA512 2559f3c3d5f4cc22299482f352ad67a5145a9d0727675932031f8b5977acaabfe29838b0e147d72066bd8cad6c046ee60cc60b364118cd9187be37fd22937205
+DIST parquet-testing-e74785d85a4ecee829e1e405444d6a1b24b8bc9c.tar.gz 1231649 BLAKE2B c9dea116ba68b896d4542ee182e42fb52a579022a89eba2a95222274a6d8bc1c8a83039c22dccdce434e01967e4010229cf864c68568a40a888bc410fc2c032b SHA512 d46330962c64ab86ec7ed8ec5416820428c0e87460bd12a7fcf04ac860b9d6985f37d5bfbfc4c53fba762f84cdc6f2acfe52900c4c19d2b73a5a2787a50853bd
diff --git a/dev-python/pyarrow/metadata.xml b/dev-python/pyarrow/metadata.xml
new file mode 100644
index 000000000000..1366c37c470f
--- /dev/null
+++ b/dev-python/pyarrow/metadata.xml
@@ -0,0 +1,19 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<!DOCTYPE pkgmetadata SYSTEM "https://docs.baldeagleos.com/dtd/metadata.dtd">
+<pkgmetadata>
+  <maintainer type="person">
+    <email>tupone@gentoo.org</email>
+    <name>Tupone Alfredo</name>
+  </maintainer>
+  <maintainer type="project">
+    <email>python@gentoo.org</email>
+    <name>Python</name>
+  </maintainer>
+  <use>
+    <flag name="parquet">Enables read/write parquet data format</flag>
+  </use>
+  <upstream>
+    <remote-id type="cpe">cpe:/a:apache:arrow</remote-id>
+  </upstream>
+  <origin>baldeagleos-repo</origin>
+</pkgmetadata>
diff --git a/dev-python/pyarrow/pyarrow-23.0.1.ebuild b/dev-python/pyarrow/pyarrow-23.0.1.ebuild
new file mode 100644
index 000000000000..41f0b7393b42
--- /dev/null
+++ b/dev-python/pyarrow/pyarrow-23.0.1.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_{13..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-${ARROW_DATA_GIT_HASH}/data"
+	epytest --pyargs pyarrow
+}
diff --git a/dev-python/pyarrow/pyarrow-24.0.0.ebuild b/dev-python/pyarrow/pyarrow-24.0.0.ebuild
new file mode 100644
index 000000000000..c5fa6a50fc71
--- /dev/null
+++ b/dev-python/pyarrow/pyarrow-24.0.0.ebuild
@@ -0,0 +1,111 @@
+# Copyright 2023-2026 Gentoo Authors
+# Distributed under the terms of the GNU General Public License v2
+
+EAPI=8
+
+DISTUTILS_EXT=1
+DISTUTILS_USE_PEP517=scikit-build-core
+# upstream's standalone wrapper modifies sdist only
+DISTUTILS_UPSTREAM_PEP517=standalone
+PYTHON_COMPAT=( python3_{13..14} )
+
+inherit distutils-r1 multiprocessing
+
+# arrow.git: testing
+ARROW_DATA_GIT_HASH=249079a810caedda6898464003c7ef8a47efeeae
+# arrow.git: cpp/submodules/parquet-testing
+PARQUET_DATA_GIT_HASH=e74785d85a4ecee829e1e405444d6a1b24b8bc9c
+
+DESCRIPTION="Python library for Apache Arrow"
+HOMEPAGE="
+	https://arrow.apache.org/
+	https://github.com/apache/arrow/
+	https://pypi.org/project/pyarrow/
+"
+SRC_URI="
+	mirror://apache/arrow/arrow-${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
+}
diff --git a/dev-python/pyarrow/pyarrow-9999.ebuild b/dev-python/pyarrow/pyarrow-9999.ebuild
new file mode 100644
index 000000000000..889a46a127ee
--- /dev/null
+++ b/dev-python/pyarrow/pyarrow-9999.ebuild
@@ -0,0 +1,95 @@
+# 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_{13..14} )
+
+inherit distutils-r1 git-r3 multiprocessing
+
+DESCRIPTION="Python library for Apache Arrow"
+HOMEPAGE="
+	https://arrow.apache.org/
+	https://github.com/apache/arrow/
+	https://pypi.org/project/pyarrow/
+"
+EGIT_REPO_URI="https://github.com/apache/arrow.git"
+EGIT_SUBMODULES=( '*' )
+S="${WORKDIR}/${P}/python"
+
+LICENSE="Apache-2.0"
+SLOT="0"
+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
+		# 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}/${P}/cpp/submodules/parquet-testing/data"
+	local -x ARROW_TEST_DATA="${WORKDIR}/${P}/testing/data"
+	epytest --pyargs pyarrow
+}