diff options
Diffstat (limited to 'sci-ml/datasets')
| -rw-r--r-- | sci-ml/datasets/Manifest | 4 | ||||
| -rw-r--r-- | sci-ml/datasets/datasets-2.21.0-r2.ebuild | 152 | ||||
| -rw-r--r-- | sci-ml/datasets/files/datasets-2.21.0-tests.patch | 22 | ||||
| -rw-r--r-- | sci-ml/datasets/metadata.xml | 12 |
4 files changed, 190 insertions, 0 deletions
diff --git a/sci-ml/datasets/Manifest b/sci-ml/datasets/Manifest new file mode 100644 index 000000000000..38d95f43d5d5 --- /dev/null +++ b/sci-ml/datasets/Manifest @@ -0,0 +1,4 @@ +AUX datasets-2.21.0-tests.patch 721 BLAKE2B cf85b02a4eb5da27e52bf819f00e333d6e3fb3fb0b9f7279afff24e3fddf76fcd46cb5e06d29b4780d0ce6b48c9378df301c70f6339e2f2898560219ebab4210 SHA512 43cdc5af6ba3fff3e7641006be6d8e10a5ad00136c1cf137628456019cccbd3d05b888b7eeec1cb8cd42d5ce52091b2a66d32a5b41ff2193658c4257900ea45c +DIST datasets-2.21.0.gh.tar.gz 2198480 BLAKE2B 41a5e82013d26644db4f52b16728cee31c6ff27995b8a9b2ddef1512fcda5e9386e47f6726303fec64d5ed7e40432583a6ddcc2bdfbb76096dce5f23b0540149 SHA512 f4193e29505300f98ec026b1ed052ffd3d43ae05a9b191bb56105d703ba2e44752e97c0eeb36dd1b2df6499662915591ab850cdec5248af8c7c5b5a003437c5c +EBUILD datasets-2.21.0-r2.ebuild 7922 BLAKE2B 19e99c9b6f1aa65ce999b84287f289140307992c7345ab7769cfe6b1b2b4032175c0c3d09067ea52b416f9aaa3ef3a1e69d68353a81e8accd56c6ada9b6eb7df SHA512 ab58615e32224d054341219c970fc37fae9e7ce9439dc4c91223cd5c61f8063d77c96f0467a1196e07ea5c94b617347222aa646cd6fc8fae764119da726471bc +MISC metadata.xml 379 BLAKE2B 48ebb9e7bfa8b58b0d15b82c4146def465e08cf3212ab4af04129d09c153b67b00d0fa05b94d6af54f643ec3a202f2335d3254b966f49d1394d3c7b9e5da56a5 SHA512 99560decfaa0e438980f372d99257695e9ca9585167d9aba091e0b775c2f8384657ddc017841c8f06f8b568017a54fb9e31da736f3c875da717e154cdce876d1 diff --git a/sci-ml/datasets/datasets-2.21.0-r2.ebuild b/sci-ml/datasets/datasets-2.21.0-r2.ebuild new file mode 100644 index 000000000000..6c7f8faeefb7 --- /dev/null +++ b/sci-ml/datasets/datasets-2.21.0-r2.ebuild @@ -0,0 +1,152 @@ +# 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..12} ) +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" + +IUSE="test" + +RDEPEND=" + ${PYTHON_DEPS} + sci-ml/pytorch[${PYTHON_SINGLE_USEDEP}] + sci-ml/caffe2[${PYTHON_SINGLE_USEDEP},numpy] + $(python_gen_cond_dep ' + dev-python/aiohttp[${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}] + sci-ml/huggingface_hub[${PYTHON_USEDEP}] + ') +" +DEPEND="${RDEPEND}" +BDEPEND="test? ( + $(python_gen_cond_dep ' + dev-python/absl-py[${PYTHON_USEDEP}] + dev-python/decorator[${PYTHON_USEDEP}] + dev-python/pytest-datadir[${PYTHON_USEDEP}] + dev-python/scikit-learn[${PYTHON_USEDEP}] + dev-python/sqlalchemy[${PYTHON_USEDEP}] + dev-python/zstandard[${PYTHON_USEDEP}] + sci-libs/jiwer[${PYTHON_USEDEP}] + sci-libs/seqeval[${PYTHON_USEDEP}] + ') + sci-ml/torchvision[${PYTHON_SINGLE_USEDEP}] +)" + +PATCHES=( + "${FILESDIR}"/${P}-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/test_upstream_hub.py + tests/packaged_modules/test_spark.py + tests/test_load.py + ) + + local EPYTEST_DESELECT=( + "tests/features/test_audio.py::test_dataset_with_audio_feature_map_undecoded" + "tests/features/test_audio.py::test_dataset_with_audio_feature_undecoded" + "tests/features/test_audio.py::test_formatted_dataset_with_audio_feature_undecoded" + "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_arrow_dataset.py::TaskTemplatesTest::test_task_automatic_speech_recognition" + "tests/test_data_files.py::test_DataFilesDict_from_patterns_locally_or_remote_hashing" + "tests/test_data_files.py::test_DataFilesList_from_patterns_locally_with_extra_files" + "tests/test_distributed.py::test_torch_distributed_run" + "tests/test_exceptions.py::test_error_deprecated" + "tests/test_file_utils.py::TestxPath::test_xpath_rglob" + "tests/test_file_utils.py::TestxPath::test_xpath_glob" + "tests/test_fingerprint.py::HashingTest::test_hash_torch_compiled_module" + "tests/test_fingerprint.py::HashingTest::test_hash_torch_generator" + "tests/test_fingerprint.py::HashingTest::test_hash_torch_tensor" + "tests/test_fingerprint.py::HashingTest::test_set_doesnt_depend_on_order" + "tests/test_fingerprint.py::HashingTest::test_set_stable" + "tests/test_fingerprint.py::RecurseHashTest::test_hash_ignores_line_definition_of_function" + "tests/test_fingerprint.py::RecurseHashTest::test_hash_ipython_function" + "tests/test_fingerprint.py::test_move_script_doesnt_change_hash" + "tests/test_fingerprint.py::TokenizersHashTest::test_hash_regex" + "tests/test_fingerprint.py::TokenizersHashTest::test_hash_tokenizer" + "tests/test_fingerprint.py::TokenizersHashTest::test_hash_tokenizer_with_cache" + "tests/test_hf_gcp.py::TestDatasetOnHfGcp::test_dataset_info_available_natural_questions/default@19ba7767b174ad046a84f46af056517a3910ee57" + "tests/test_hf_gcp.py::TestDatasetOnHfGcp::test_dataset_info_available_wiki40b/en@7b21a2e64b90323b2d3d1b81aa349bb4bc76d9bf" + "tests/test_hf_gcp.py::TestDatasetOnHfGcp::test_dataset_info_available_wiki_dpr/psgs_w100.multiset.no_index@b24a417d802a583f8922946c1c75210290e93108" + "tests/test_hf_gcp.py::TestDatasetOnHfGcp::test_dataset_info_available_wiki_dpr/psgs_w100.nq.compressed@b24a417d802a583f8922946c1c75210290e93108" + "tests/test_hf_gcp.py::TestDatasetOnHfGcp::test_dataset_info_available_wiki_dpr/psgs_w100.nq.no_index@b24a417d802a583f8922946c1c75210290e93108" + "tests/test_hf_gcp.py::TestDatasetOnHfGcp::test_dataset_info_available_wikipedia/20220301.de@4d013bdd32c475c8536aae00a56efc774f061649" + "tests/test_hf_gcp.py::TestDatasetOnHfGcp::test_dataset_info_available_wikipedia/20220301.en@4d013bdd32c475c8536aae00a56efc774f061649" + "tests/test_hf_gcp.py::TestDatasetOnHfGcp::test_dataset_info_available_wikipedia/20220301.fr@4d013bdd32c475c8536aae00a56efc774f061649" + "tests/test_hf_gcp.py::TestDatasetOnHfGcp::test_dataset_info_available_wikipedia/20220301.frr@4d013bdd32c475c8536aae00a56efc774f061649" + "tests/test_hf_gcp.py::TestDatasetOnHfGcp::test_dataset_info_available_wikipedia/20220301.it@4d013bdd32c475c8536aae00a56efc774f061649" + "tests/test_hf_gcp.py::TestDatasetOnHfGcp::test_dataset_info_available_wikipedia/20220301.simple@4d013bdd32c475c8536aae00a56efc774f061649" + "tests/test_hf_gcp.py::test_as_dataset_from_hf_gcs" + "tests/test_hf_gcp.py::test_as_streaming_dataset_from_hf_gcs" + "tests/test_hub.py::test_convert_to_parquet" + "tests/test_hub.py::test_delete_from_hub" + "tests/test_iterable_dataset.py::test_iterable_dataset_from_file" + "tests/test_iterable_dataset.py::test_iterable_dataset_from_hub_torch_dataloader_parallel" + "tests/test_metric_common.py::LocalMetricTest::test_load_metric_bertscore" + "tests/test_metric_common.py::LocalMetricTest::test_load_metric_bleurt" + "tests/test_metric_common.py::LocalMetricTest::test_load_metric_chrf" + "tests/test_metric_common.py::LocalMetricTest::test_load_metric_code_eval" + "tests/test_metric_common.py::LocalMetricTest::test_load_metric_competition_math" + "tests/test_metric_common.py::LocalMetricTest::test_load_metric_coval" + "tests/test_metric_common.py::LocalMetricTest::test_load_metric_cuad" + "tests/test_metric_common.py::LocalMetricTest::test_load_metric_f1" + "tests/test_metric_common.py::LocalMetricTest::test_load_metric_frugalscore" + "tests/test_metric_common.py::LocalMetricTest::test_load_metric_glue" + "tests/test_metric_common.py::LocalMetricTest::test_load_metric_google_bleu" + "tests/test_metric_common.py::LocalMetricTest::test_load_metric_mauve" + "tests/test_metric_common.py::LocalMetricTest::test_load_metric_mean_iou" + "tests/test_metric_common.py::LocalMetricTest::test_load_metric_meteor" + "tests/test_metric_common.py::LocalMetricTest::test_load_metric_mse" + "tests/test_metric_common.py::LocalMetricTest::test_load_metric_precision" + "tests/test_metric_common.py::LocalMetricTest::test_load_metric_recall" + "tests/test_metric_common.py::LocalMetricTest::test_load_metric_roc_auc" + "tests/test_metric_common.py::LocalMetricTest::test_load_metric_rouge" + "tests/test_metric_common.py::LocalMetricTest::test_load_metric_sacrebleu" + "tests/test_metric_common.py::LocalMetricTest::test_load_metric_sari" + "tests/test_metric_common.py::LocalMetricTest::test_load_metric_spearmanr" + "tests/test_metric_common.py::LocalMetricTest::test_load_metric_super_glue" + "tests/test_metric_common.py::LocalMetricTest::test_load_metric_ter" + "tests/test_metric_common.py::LocalMetricTest::test_load_metric_wiki_split" + "tests/test_metric_common.py::LocalMetricTest::test_load_metric_xnli" + "tests/test_metric_common.py::LocalMetricTest::test_load_metric_xtreme_s" + "tests/test_offline_util.py::test_offline_with_timeout" + ) + distutils-r1_src_test +} diff --git a/sci-ml/datasets/files/datasets-2.21.0-tests.patch b/sci-ml/datasets/files/datasets-2.21.0-tests.patch new file mode 100644 index 000000000000..0a55459db109 --- /dev/null +++ b/sci-ml/datasets/files/datasets-2.21.0-tests.patch @@ -0,0 +1,22 @@ +--- a/tests/test_arrow_dataset.py 2024-02-20 21:53:24.248470991 +0100 ++++ b/tests/test_arrow_dataset.py 2024-02-20 21:53:29.441804737 +0100 +@@ -4131,7 +4131,6 @@ + [ + "relative/path", + "/absolute/path", +- "s3://bucket/relative/path", + "hdfs://relative/path", + "hdfs:///absolute/path", + ], +--- a/tests/packaged_modules/test_audiofolder.py 2023-05-06 14:00:39.560876163 +0200 ++++ b/tests/packaged_modules/test_audiofolder.py 2023-05-06 14:01:26.005212423 +0200 +@@ -1,9 +1,8 @@ + import shutil + import textwrap + + import numpy as np + import pytest +-import soundfile as sf + + from datasets import Audio, ClassLabel, Features, Value + from datasets.builder import InvalidConfigName diff --git a/sci-ml/datasets/metadata.xml b/sci-ml/datasets/metadata.xml new file mode 100644 index 000000000000..94c112402049 --- /dev/null +++ b/sci-ml/datasets/metadata.xml @@ -0,0 +1,12 @@ +<?xml version="1.0" encoding="UTF-8"?> +<!DOCTYPE pkgmetadata SYSTEM "https://www.gentoo.org/dtd/metadata.dtd"> +<pkgmetadata> + <maintainer type="person"> + <email>tupone@gentoo.org</email> + <name>Tupone Alfredo</name> + </maintainer> + <upstream> + <remote-id type="github">huggingface/datasets</remote-id> + <remote-id type="pypi">datasets</remote-id> + </upstream> +</pkgmetadata> |