dev-python/pandas: Bump to 2.1.4

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

3 files changed, 267 insertions, 0 deletions

diff --git a/dev-python/pandas/Manifest b/dev-python/pandas/Manifest
index 408349fbf7e5..eafa0a1db645 100644
--- a/dev-python/pandas/Manifest
+++ b/dev-python/pandas/Manifest
@@ -2,3 +2,4 @@ DIST pandas-1.5.3.tar.gz 5203060 BLAKE2B 846a6caa8ead6a2556303494f92e2091582d8fc
 DIST pandas-2.1.1.tar.gz 4266250 BLAKE2B 9d8256300db65b606e34b16d7f3884b602794fe3c14d66115056db4a4658603fbe7a7a3082dfc781760517d643d6e21a6d4d43441131ec6aed5965b6fa8a0e9c SHA512 346e13e06805948c569712028c9b8d756c67ca2fab040999a51a4e759206d50d8ee25be38ac3c7cb3d2e692a02d85c1ec198841311fe43cb04b64f2741627a4e
 DIST pandas-2.1.2.tar.gz 4271419 BLAKE2B 2f22035c9b7d6a0403b28ebe506550328e096f0ed87984dba28d2f49950393ed8a5a3c0c71c5391385d9f5361319431d5de3736f7381eb7670045415b05137a7 SHA512 dad04f1fcc5a59485c34ebfed0698fd47ad9eeffafaddf91642c084f4ffde8a55ce543f9dadb6e7a5066b1809489961d34fe3fe8860f64f9bd2d90b445d92063
 DIST pandas-2.1.3.tar.gz 4272855 BLAKE2B 73717926bb0af44acb9a2d12964df7bf70affd61ca19468b3af4376f9c8084d3612895a9ac5117be97658fd5b7ff0275cd7b2c6f1b9b0e00afb951fd50a1f103 SHA512 71012b47b76444e19507e444c8ef88497e3b286856690216886cbb67ad70786c2f5014a7cf6c89f8378769560109fb5593a8c82c21a034a1059769e15c843b92
+DIST pandas-2.1.4.tar.gz 4274800 BLAKE2B 71cd374964d5e751a9ea0b29cb341d7de521de8de3bbe0963fc821eb6fa2babf2a5cb87c0f49fab88fd451bd331d83fd216f199c75876d7c1ea367ba53c20178 SHA512 de02944829ea4df6a3913788c2ae30683504d70a1cef5789c927245dab8700981edd50de18af3619d30ac97f4658fce67f6c7637acf496bd464ad3a4109bfd1c
diff --git a/dev-python/pandas/files/pandas-2.1.4-khash-dep.patch b/dev-python/pandas/files/pandas-2.1.4-khash-dep.patch
new file mode 100644
index 000000000000..f8bbff5fed37
--- /dev/null
+++ b/dev-python/pandas/files/pandas-2.1.4-khash-dep.patch
@@ -0,0 +1,45 @@
+From e31a6865958442435ae9b31f312129c44f66eb5e Mon Sep 17 00:00:00 2001
+From: Ville Aikas <11279988+vaikas@users.noreply.github.com>
+Date: Mon, 6 Nov 2023 09:24:40 -0800
+Subject: [PATCH] Add missing dependencies for: _khash_primitive_helper
+ (#55795)
+
+* Add missing dependencies for: _khash_primitive_helper
+
+Signed-off-by: Ville Aikas <vaikas@chainguard.dev>
+
+* source->dep change.
+
+Signed-off-by: Ville Aikas <vaikas@chainguard.dev>
+
+* Remove dep from arrays.
+
+Signed-off-by: Ville Aikas <vaikas@chainguard.dev>
+
+---------
+
+Signed-off-by: Ville Aikas <vaikas@chainguard.dev>
+---
+ pandas/_libs/meson.build | 6 +++---
+ 1 file changed, 3 insertions(+), 3 deletions(-)
+
+diff --git a/pandas/_libs/meson.build b/pandas/_libs/meson.build
+index b4662d6bf8dd2..c27386743c6e9 100644
+--- a/pandas/_libs/meson.build
++++ b/pandas/_libs/meson.build
+@@ -61,12 +61,12 @@ subdir('tslibs')
+ libs_sources = {
+     # Dict of extension name -> dict of {sources, include_dirs, and deps}
+     # numpy include dir is implicitly included
+-    'algos': {'sources': ['algos.pyx', _algos_common_helper, _algos_take_helper, _khash_primitive_helper]},
++    'algos': {'sources': ['algos.pyx', _algos_common_helper, _algos_take_helper], 'deps': _khash_primitive_helper_dep},
+     'arrays': {'sources': ['arrays.pyx']},
+     'groupby': {'sources': ['groupby.pyx']},
+     'hashing': {'sources': ['hashing.pyx']},
+-    'hashtable': {'sources': ['hashtable.pyx', _khash_primitive_helper, _hashtable_class_helper, _hashtable_func_helper]},
+-    'index': {'sources': ['index.pyx', _index_class_helper]},
++    'hashtable': {'sources': ['hashtable.pyx', _hashtable_class_helper, _hashtable_func_helper], 'deps': _khash_primitive_helper_dep},
++    'index': {'sources': ['index.pyx', _index_class_helper], 'deps': _khash_primitive_helper_dep},
+     'indexing': {'sources': ['indexing.pyx']},
+     'internals': {'sources': ['internals.pyx']},
+     'interval': {'sources': ['interval.pyx', _intervaltree_helper],
diff --git a/dev-python/pandas/pandas-2.1.4.ebuild b/dev-python/pandas/pandas-2.1.4.ebuild
new file mode 100644
index 000000000000..f4a73c081111
--- /dev/null
+++ b/dev-python/pandas/pandas-2.1.4.ebuild
@@ -0,0 +1,221 @@
+# Copyright 1999-2023 Gentoo Authors
+# Distributed under the terms of the GNU General Public License v2
+
+EAPI=8
+
+DISTUTILS_EXT=1
+DISTUTILS_USE_PEP517=meson-python
+PYTHON_COMPAT=( python3_{10..12} )
+PYTHON_REQ_USE="threads(+)"
+
+VIRTUALX_REQUIRED="manual"
+
+inherit distutils-r1 optfeature pypi virtualx
+
+DESCRIPTION="Powerful data structures for data analysis and statistics"
+HOMEPAGE="
+	https://pandas.pydata.org/
+	https://github.com/pandas-dev/pandas/
+	https://pypi.org/project/pandas/
+"
+
+SLOT="0"
+LICENSE="BSD"
+KEYWORDS="~amd64 ~arm ~arm64 ~hppa ~ppc ~ppc64 ~riscv ~s390 ~sparc ~x86"
+IUSE="full-support minimal test X"
+RESTRICT="!test? ( test )"
+
+RECOMMENDED_DEPEND="
+	>=dev-python/bottleneck-1.3.4[${PYTHON_USEDEP}]
+	>=dev-python/numexpr-2.8.0[${PYTHON_USEDEP}]
+"
+
+# TODO: add pandas-gbq to the tree
+# TODO: Re-add dev-python/statsmodel[python3_11] dep once it supports python3_11
+# https://github.com/statsmodels/statsmodels/issues/8287
+OPTIONAL_DEPEND="
+	>=dev-python/beautifulsoup4-4.11.1[${PYTHON_USEDEP}]
+	dev-python/blosc[${PYTHON_USEDEP}]
+	>=dev-python/html5lib-1.1[${PYTHON_USEDEP}]
+	>=dev-python/jinja-3.1.2[${PYTHON_USEDEP}]
+	>=dev-python/lxml-4.8.0[${PYTHON_USEDEP}]
+	>=dev-python/matplotlib-3.6.1[${PYTHON_USEDEP}]
+	>=dev-python/openpyxl-3.0.7[${PYTHON_USEDEP}]
+	>=dev-python/sqlalchemy-1.4.36[${PYTHON_USEDEP}]
+	>=dev-python/tables-3.7.0[${PYTHON_USEDEP}]
+	>=dev-python/tabulate-0.8.10[${PYTHON_USEDEP}]
+	>=dev-python/xarray-2022.3.0[${PYTHON_USEDEP}]
+	>=dev-python/xlrd-2.0.1[${PYTHON_USEDEP}]
+	>=dev-python/xlsxwriter-3.0.3[${PYTHON_USEDEP}]
+	>=dev-python/xlwt-1.3.0[${PYTHON_USEDEP}]
+	!hppa? (
+		$(python_gen_cond_dep '
+			dev-python/statsmodels[${PYTHON_USEDEP}]
+		' python3_{8..10} )
+		>=dev-python/scipy-1.8.1[${PYTHON_USEDEP}]
+	)
+	X? (
+		|| (
+			>=dev-python/PyQt5-5.15.6[${PYTHON_USEDEP}]
+			>=dev-python/QtPy-2.2.0[${PYTHON_USEDEP}]
+			x11-misc/xclip
+			x11-misc/xsel
+		)
+	)
+"
+DEPEND="
+	>=dev-python/numpy-1.23.2[${PYTHON_USEDEP}]
+"
+COMMON_DEPEND="
+	${DEPEND}
+	>=dev-python/python-dateutil-2.8.2[${PYTHON_USEDEP}]
+	>=dev-python/pytz-2020.1[${PYTHON_USEDEP}]
+"
+BDEPEND="
+	${COMMON_DEPEND}
+	>=dev-util/meson-1.2.1
+	>=dev-python/cython-0.29.33[${PYTHON_USEDEP}]
+	>=dev-python/versioneer-0.28[${PYTHON_USEDEP}]
+	test? (
+		${VIRTUALX_DEPEND}
+		${RECOMMENDED_DEPEND}
+		${OPTIONAL_DEPEND}
+		>=dev-python/beautifulsoup4-4.11.1[${PYTHON_USEDEP}]
+		>=dev-python/hypothesis-6.46.1[${PYTHON_USEDEP}]
+		>=dev-python/openpyxl-3.0.10[${PYTHON_USEDEP}]
+		>=dev-python/pymysql-1.0.2[${PYTHON_USEDEP}]
+		>=dev-python/psycopg-2.9.3:2[${PYTHON_USEDEP}]
+		>=dev-python/xlsxwriter-3.0.3[${PYTHON_USEDEP}]
+		x11-misc/xclip
+		x11-misc/xsel
+	)
+"
+RDEPEND="
+	${COMMON_DEPEND}
+	dev-python/tzdata[${PYTHON_USEDEP}]
+	!minimal? ( ${RECOMMENDED_DEPEND} )
+	full-support? ( ${OPTIONAL_DEPEND} )
+"
+
+PATCHES=(
+	# a quick hack, it's already fixed in main
+	"${FILESDIR}/pandas-2.1.1-which.patch"
+	# https://github.com/pandas-dev/pandas/commit/e31a6865958442435ae9b31f312129c44f66eb5e
+	"${FILESDIR}/pandas-2.1.4-khash-dep.patch"
+)
+
+EPYTEST_XDIST=1
+distutils_enable_tests pytest
+
+src_test() {
+	virtx distutils-r1_src_test
+}
+
+python_test() {
+	local EPYTEST_DESELECT=(
+		# test for rounding errors, fails if we have better precision
+		# e.g. on amd64 with FMA or on arm64
+		# https://github.com/pandas-dev/pandas/issues/38921
+		tests/window/test_rolling.py::test_rolling_var_numerical_issues
+
+		# TODO; unhappy about DISPLAY?
+		tests/test_downstream.py::test_seaborn
+
+		# OOMs
+		tests/io/parser/test_c_parser_only.py::test_bytes_exceed_2gb
+
+		# TODO: numexpr says "forbidden control characters"
+		tests/computation/test_eval.py::TestOperations::test_multi_line_expression_local_variable
+		'tests/computation/test_eval.py::test_query_token[numexpr-Temp(\xb0C)]'
+		tests/frame/test_query_eval.py::TestDataFrameQueryBacktickQuoting::test_lots_of_operators_string
+		tests/frame/test_query_eval.py::TestDataFrameQueryBacktickQuoting::test_multiple_spaces
+		tests/frame/test_query_eval.py::TestDataFrameQueryBacktickQuoting::test_parenthesis
+		tests/frame/test_query_eval.py::TestDataFrameQueryBacktickQuoting::test_start_with_spaces
+		tests/frame/test_query_eval.py::TestDataFrameQueryNumExprPandas::test_local_syntax
+		tests/frame/test_query_eval.py::TestDataFrameQueryNumExprPandas::test_local_variable_with_in
+		tests/frame/test_query_eval.py::TestDataFrameQueryNumExprPandas::test_nested_scope
+		tests/frame/test_query_eval.py::TestDataFrameQueryNumExprPandas::test_query_scope
+
+		# TODO: missing data not covered by --no-strict-data-files?
+		tests/io/xml/test_xml.py::test_empty_stylesheet
+		tests/io/xml/test_xml.py::test_wrong_file_path
+
+		# TODO
+		tests/frame/test_arithmetic.py::TestFrameFlexArithmetic::test_floordiv_axis0_numexpr_path
+
+		# deprecation warning
+		tests/io/pytables/test_retain_attributes.py::test_retain_index_attributes2
+		'tests/computation/test_eval.py::TestEval::test_scalar_unary[numexpr-pandas]'
+
+		# Needs 64-bit time_t (TODO: split into 32-bit arch only section)
+		tests/tseries/offsets/test_year.py::test_add_out_of_pydatetime_range
+		'tests/tseries/offsets/test_common.py::test_apply_out_of_range[tzlocal()-BusinessDay]'
+		'tests/tseries/offsets/test_common.py::test_apply_out_of_range[tzlocal()-BusinessHour]'
+		'tests/tseries/offsets/test_common.py::test_apply_out_of_range[tzlocal()-BusinessMonthEnd]'
+		'tests/tseries/offsets/test_common.py::test_apply_out_of_range[tzlocal()-BusinessMonthBegin]'
+		'tests/tseries/offsets/test_common.py::test_apply_out_of_range[tzlocal()-BQuarterEnd]'
+		'tests/tseries/offsets/test_common.py::test_apply_out_of_range[tzlocal()-BQuarterBegin]'
+		'tests/tseries/offsets/test_common.py::test_apply_out_of_range[tzlocal()-CustomBusinessDay]'
+		'tests/tseries/offsets/test_common.py::test_apply_out_of_range[tzlocal()-CustomBusinessHour]'
+		'tests/tseries/offsets/test_common.py::test_apply_out_of_range[tzlocal()-CustomBusinessMonthEnd]'
+		'tests/tseries/offsets/test_common.py::test_apply_out_of_range[tzlocal()-CustomBusinessMonthBegin]'
+		'tests/tseries/offsets/test_common.py::test_apply_out_of_range[tzlocal()-MonthEnd]'
+		'tests/tseries/offsets/test_common.py::test_apply_out_of_range[tzlocal()-MonthBegin]'
+		'tests/tseries/offsets/test_common.py::test_apply_out_of_range[tzlocal()-SemiMonthBegin]'
+		'tests/tseries/offsets/test_common.py::test_apply_out_of_range[tzlocal()-SemiMonthEnd]'
+		'tests/tseries/offsets/test_common.py::test_apply_out_of_range[tzlocal()-QuarterEnd]'
+		'tests/tseries/offsets/test_common.py::test_apply_out_of_range[tzlocal()-LastWeekOfMonth]'
+		'tests/tseries/offsets/test_common.py::test_apply_out_of_range[tzlocal()-WeekOfMonth]'
+		'tests/tseries/offsets/test_common.py::test_apply_out_of_range[tzlocal()-Week]'
+
+		# hdf / pytables have alignment problems: bug #911660
+		# https://github.com/pandas-dev/pandas/issues/54396
+		tests/io/pytables/test_append.py::test_append_frame_column_oriented
+		tests/io/pytables/test_store.py::test_select_filter_corner
+
+		# blosc2 version
+		tests/io/pytables/test_file_handling.py::test_complibs\[blosc2-{1..9}\]
+	)
+
+	case ${EPYTHON} in
+		python3.12)
+			EPYTEST_DESELECT+=(
+				tests/io/pytables/test_select.py::test_select_dtypes
+				tests/io/pytables/test_select.py::test_frame_select
+				# deprecation warnings
+				tests/io/excel/test_writers.py::TestRoundTrip::test_read_excel_parse_dates
+				# some pytest-xdist problem?
+				tests/base/test_unique.py::test_unique_bad_unicode
+			)
+			;;
+	esac
+
+	local -x LC_ALL=C.UTF-8
+	cd "${BUILD_DIR}/install$(python_get_sitedir)" || die
+	"${EPYTHON}" -c "import pandas; pandas.show_versions()" || die
+	# --no-strict-data-files is necessary since upstream prevents data
+	# files from even being included in GitHub archives, sigh
+	# https://github.com/pandas-dev/pandas/issues/54907
+	local -x PYTEST_DISABLE_PLUGIN_AUTOLOAD=1
+	epytest pandas/tests \
+		--no-strict-data-files \
+		-m "not single and not slow and not network" ||
+		die "Tests failed with ${EPYTHON}"
+	rm test-data.xml || die
+}
+
+pkg_postinst() {
+	optfeature "accelerating certain types of NaN evaluations, using specialized cython routines to achieve large speedups." dev-python/bottleneck
+	optfeature "accelerating certain numerical operations, using multiple cores as well as smart chunking and caching to achieve large speedups" ">=dev-python/numexpr-2.1"
+	optfeature "needed for pandas.io.html.read_html" dev-python/beautifulsoup4 dev-python/html5lib dev-python/lxml
+	optfeature "for msgpack compression using blosc" dev-python/blosc
+	optfeature "Template engine for conditional HTML formatting" dev-python/jinja
+	optfeature "Plotting support" dev-python/matplotlib
+	optfeature "Needed for Excel I/O" ">=dev-python/openpyxl-3.0.10" dev-python/xlsxwriter dev-python/xlrd dev-python/xlwt
+	optfeature "necessary for HDF5-based storage" ">=dev-python/tables-3.7.0"
+	optfeature "R I/O support" dev-python/rpy
+	optfeature "Needed for parts of pandas.stats" dev-python/statsmodels
+	optfeature "SQL database support" ">=dev-python/sqlalchemy-1.4.36"
+	optfeature "miscellaneous statistical functions" dev-python/scipy
+	optfeature "necessary to use pandas.io.clipboard.read_clipboard support" dev-python/PyQt5 dev-python/QtPy dev-python/pygtk x11-misc/xclip x11-misc/xsel
+}