commit: da7700303f28481c4766716f0e6a1de721f5d2f8
Author: Alfredo Tupone <tupone <AT> gentoo <DOT> org>
AuthorDate: Mon Apr 14 18:12:07 2025 +0000
Commit: Alfredo Tupone <tupone <AT> gentoo <DOT> org>
CommitDate: Mon Apr 14 18:12:22 2025 +0000
URL: https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=da770030
sci-ml/datasets: add 3.2.0
Signed-off-by: Alfredo Tupone <tupone <AT> gentoo.org>
sci-ml/datasets/Manifest | 1 +
sci-ml/datasets/datasets-3.2.0.ebuild | 112 ++++++++++++++++++++++++++++++++++
2 files changed, 113 insertions(+)
diff --git a/sci-ml/datasets/Manifest b/sci-ml/datasets/Manifest
index 5b7eec02ccc8..86d90f578b59 100644
--- a/sci-ml/datasets/Manifest
+++ b/sci-ml/datasets/Manifest
@@ -1,2 +1,3 @@
DIST datasets-3.0.2.gh.tar.gz 1867037 BLAKE2B
92475ac566bd825c542dbd3e138c83e1733e6d1137f30a2cfd8f82f7dd54ed78c0566aa766182625e8787413be18f5a0c0c0a289b5823fa9112dcb2876074b7e
SHA512
aa55dd2f53df7217c30999f759c76a13d78851233bce4df9b1ad85081e101a7280e905822160874d24d15ea1e1a5ecee1bae9cc66785e89af45ad16dd0339595
DIST datasets-3.1.0.gh.tar.gz 1877057 BLAKE2B
818cbf10282f775cda5956ae37607bb5357a753229fa91788b19437eb28df609a99c9000f283a8b9b27a920efb3b32f991d729db94a0acf08aaca8cdcd6c4f4b
SHA512
7e2a22b6ca423f4358153137f50d963075d37ad01e834c78176dc432f1b3adc3b80df44ac35b22d727db4e1aa8331c938efca1ab2d678e58401e8e7ce571abfc
+DIST datasets-3.2.0.gh.tar.gz 1877261 BLAKE2B
0169ee27e1128f8b64fda0e3a4dd63d5d603874cdbf703cd73f100ad1ea69294c300b93cddaa4f2dd8c795570b76adb34e40475d38730f4733679008fb593d84
SHA512
2e381d6919ea7f2e934e832d6bff30a4b23cfe929c0b4345703ce39c7882daf8a234fe71ca89d1d9bcd75000e35845156f7036f9604504ac584feb1b8d3239b1
diff --git a/sci-ml/datasets/datasets-3.2.0.ebuild
b/sci-ml/datasets/datasets-3.2.0.ebuild
new file mode 100644
index 000000000000..206da6ebe884
--- /dev/null
+++ b/sci-ml/datasets/datasets-3.2.0.ebuild
@@ -0,0 +1,112 @@
+# Copyright 2023-2025 Gentoo Authors
+# Distributed under the terms of the GNU General Public License v2
+
+EAPI=8
+
+DISTUTILS_USE_PEP517=setuptools
+PYTHON_COMPAT=( python3_{10..13} )
+DISTUTILS_SINGLE_IMPL=1
+inherit distutils-r1
+
+DESCRIPTION="Access and share datasets for Audio, Computer Vision, and NLP
tasks"
+HOMEPAGE="https://pypi.org/project/datasets/";
+SRC_URI="https://github.com/huggingface/${PN}/archive/refs/tags/${PV}.tar.gz
+ -> ${P}.gh.tar.gz"
+
+LICENSE="Apache-2.0"
+SLOT="0"
+KEYWORDS="~amd64"
+
+RDEPEND="
+ ${PYTHON_DEPS}
+ sci-ml/caffe2[${PYTHON_SINGLE_USEDEP},numpy]
+ sci-ml/huggingface_hub[${PYTHON_SINGLE_USEDEP}]
+ sci-ml/pytorch[${PYTHON_SINGLE_USEDEP}]
+ $(python_gen_cond_dep '
+ dev-python/aiohttp[${PYTHON_USEDEP}]
+ dev-python/dill[${PYTHON_USEDEP}]
+ dev-python/filelock[${PYTHON_USEDEP}]
+ dev-python/fsspec[${PYTHON_USEDEP}]
+ dev-python/multiprocess[${PYTHON_USEDEP}]
+ dev-python/numpy[${PYTHON_USEDEP}]
+ dev-python/packaging[${PYTHON_USEDEP}]
+ dev-python/pandas[${PYTHON_USEDEP}]
+ dev-python/pyarrow[${PYTHON_USEDEP},parquet,snappy]
+ dev-python/pyyaml[${PYTHON_USEDEP}]
+ dev-python/requests[${PYTHON_USEDEP}]
+ dev-python/tqdm[${PYTHON_USEDEP}]
+ dev-python/xxhash[${PYTHON_USEDEP}]
+ ')
+"
+DEPEND="${RDEPEND}"
+BDEPEND="test? (
+ sci-ml/torchvision[${PYTHON_SINGLE_USEDEP}]
+ $(python_gen_cond_dep '
+ dev-python/absl-py[${PYTHON_USEDEP}]
+ dev-python/decorator[${PYTHON_USEDEP}]
+ dev-python/lz4[${PYTHON_USEDEP}]
+ dev-python/pytest-datadir[${PYTHON_USEDEP}]
+ dev-python/scikit-learn[${PYTHON_USEDEP}]
+ dev-python/soundfile[${PYTHON_USEDEP}]
+ dev-python/sqlalchemy[${PYTHON_USEDEP}]
+ dev-python/zstandard[${PYTHON_USEDEP}]
+ sci-ml/jiwer[${PYTHON_USEDEP}]
+ sci-ml/seqeval[${PYTHON_USEDEP}]
+ ')
+)"
+
+PATCHES=(
+ "${FILESDIR}"/${PN}-3.0.2-tests.patch
+)
+
+distutils_enable_tests pytest
+
+src_prepare() {
+ distutils-r1_src_prepare
+ sed -i -e \
+ "/pyarrow_hotfix/d" \
+ src/datasets/features/features.py || die
+}
+
+src_test() {
+ local EPYTEST_IGNORE=(
+ tests/features/test_audio.py
+ tests/packaged_modules/test_audiofolder.py
+ tests/packaged_modules/test_spark.py
+ tests/test_fingerprint.py
+ tests/test_iterable_dataset.py
+ tests/test_inspect.py
+ tests/test_load.py
+ tests/test_upstream_hub.py
+ )
+
+ local EPYTEST_DESELECT=(
+ tests/commands/test_test.py::test_test_command
+ tests/io/test_parquet.py::test_parquet_read_geoparquet
+ tests/packaged_modules/test_cache.py::test_cache_multi_configs
+ tests/packaged_modules/test_cache.py::test_cache_single_config
+
tests/test_arrow_dataset.py::BaseDatasetTest::test_filter_caching_on_disk
+
tests/test_arrow_dataset.py::BaseDatasetTest::test_map_caching_on_disk
+ tests/test_distributed.py::test_torch_distributed_run
+ tests/test_file_utils.py::TestxPath::test_xpath_rglob
+ tests/test_file_utils.py::TestxPath::test_xpath_glob
+ tests/test_file_utils.py::test_xexists_private
+ tests/test_file_utils.py::test_xlistdir_private
+ tests/test_file_utils.py::test_xisdir_private
+ tests/test_file_utils.py::test_xisfile_private
+ tests/test_file_utils.py::test_xgetsize_private
+ tests/test_file_utils.py::test_xglob_private
+ tests/test_file_utils.py::test_xwalk_private
+ tests/test_hub.py::test_convert_to_parquet
+ tests/packaged_modules/test_cache.py::test_cache_capital_letters
+
tests/packaged_modules/test_folder_based_builder.py::test_data_files_with_different_levels_no_metadata
+
tests/packaged_modules/test_folder_based_builder.py::test_data_files_with_one_label_no_metadata
+
tests/test_data_files.py::test_DataFilesList_from_patterns_locally_with_extra_files
+
tests/test_data_files.py::test_DataFilesDict_from_patterns_locally_or_remote_hashing
+ tests/test_file_utils.py::test_xopen_remote
+ tests/test_hub.py::test_delete_from_hub
+ tests/test_offline_util.py::test_offline_with_timeout
+ tests/test_search.py::ElasticSearchIndexTest::test_elasticsearch
+ )
+ distutils-r1_src_test
+}
