PORTNAME=	datasets
DISTVERSION=	2.21.0
CATEGORIES=	misc python # machine-learning
MASTER_SITES=	PYPI
PKGNAMEPREFIX=	${PYTHON_PKGNAMEPREFIX}

MAINTAINER=	yuri@FreeBSD.org
COMMENT=	HuggingFace community-driven open-source library of datasets
WWW=		https://huggingface.co/docs/datasets/index

LICENSE=	MIT
LICENSE_FILE=	${WRKSRC}/LICENSE

BUILD_DEPENDS=	${PYTHON_PKGNAMEPREFIX}pyproject_hooks>0:devel/py-pyproject_hooks@${PY_FLAVOR} \
		${PY_SETUPTOOLS} \
		${PYTHON_PKGNAMEPREFIX}wheel>0:devel/py-wheel@${PY_FLAVOR}
RUN_DEPENDS=	${PYTHON_PKGNAMEPREFIX}aiohttp>0:www/py-aiohttp@${PY_FLAVOR} \
		${PYTHON_PKGNAMEPREFIX}dill>0.3.0:devel/py-dill@${PY_FLAVOR} \
		${PYTHON_PKGNAMEPREFIX}filelock>0:sysutils/py-filelock@${PY_FLAVOR} \
		${PYTHON_PKGNAMEPREFIX}fsspec>=2023.1.0:devel/py-fsspec@${PY_FLAVOR} \
		${PYTHON_PKGNAMEPREFIX}huggingface-hub>0.21.2:misc/py-huggingface-hub@${PY_FLAVOR} \
		${PYTHON_PKGNAMEPREFIX}multiprocess>0:devel/py-multiprocess@${PY_FLAVOR} \
		${PYNUMPY} \
		${PYTHON_PKGNAMEPREFIX}packaging>0:devel/py-packaging@${PY_FLAVOR} \
		${PYTHON_PKGNAMEPREFIX}pandas>0:math/py-pandas@${PY_FLAVOR} \
		${PYTHON_PKGNAMEPREFIX}pyarrow>=15.0.0:databases/py-pyarrow@${PY_FLAVOR} \
		${PYTHON_PKGNAMEPREFIX}pyyaml>=5.1:devel/py-pyyaml@${PY_FLAVOR} \
		${PYTHON_PKGNAMEPREFIX}requests>=2.32.2:www/py-requests@${PY_FLAVOR} \
		${PYTHON_PKGNAMEPREFIX}tqdm>=4.66.3:misc/py-tqdm@${PY_FLAVOR} \
		${PYTHON_PKGNAMEPREFIX}xxhash>0:devel/py-xxhash@${PY_FLAVOR}
RUN_DEPENDS+=	${PYTHON_PKGNAMEPREFIX}librosa>0:audio/py-librosa@${PY_FLAVOR} \
		${PYTHON_PKGNAMEPREFIX}SoundFile>=0.12.1:audio/py-SoundFile@${PY_FLAVOR}
RUN_DEPENDS+=	${PY_PILLOW}
TEST_DEPENDS=	${PYTHON_PKGNAMEPREFIX}absl-py>=0:devel/py-absl-py@${PY_FLAVOR} \
		${PYTHON_PKGNAMEPREFIX}elasticsearch>0:textproc/py-elasticsearch@${PY_FLAVOR} \
		${PYTHON_PKGNAMEPREFIX}faiss>=1.6.4:math/py-faiss@${PY_FLAVOR} \
		${PYTHON_PKGNAMEPREFIX}jax>=0.3.14:math/py-jax@${PY_FLAVOR} \
		${PYTHON_PKGNAMEPREFIX}jiwer>0:misc/py-jiwer@${PY_FLAVOR} \
		${PYTHON_PKGNAMEPREFIX}joblib>=1.3.0:devel/py-joblib@${PY_FLAVOR} \
		${PYTHON_PKGNAMEPREFIX}lz4>=0:archivers/py-lz4@${PY_FLAVOR} \
		${PYTHON_PKGNAMEPREFIX}polars>=0.20.0:misc/py-polars@${PY_FLAVOR} \
		${PYTHON_PKGNAMEPREFIX}protobuf>=4.0.0:devel/py-protobuf@${PY_FLAVOR} \
		${PYTHON_PKGNAMEPREFIX}pytest-datadir>=0:devel/py-pytest-datadir@${PY_FLAVOR} \
		${PYTHON_PKGNAMEPREFIX}pytest-xdist>=0:devel/py-pytest-xdist@${PY_FLAVOR} \
		${PYTHON_PKGNAMEPREFIX}pytest>=0:devel/py-pytest@${PY_FLAVOR} \
		${PYTHON_PKGNAMEPREFIX}pytorch>=2.0.0:misc/py-pytorch@${PY_FLAVOR} \
		${PYTHON_PKGNAMEPREFIX}rarfile>=4.0:archivers/py-rarfile@${PY_FLAVOR} \
		${PYTHON_PKGNAMEPREFIX}s3fs>=2021.11.1:devel/py-s3fs@${PY_FLAVOR} \
		${PYTHON_PKGNAMEPREFIX}tiktoken>=0:textproc/py-tiktoken@${PY_FLAVOR} \
		${PYTHON_PKGNAMEPREFIX}typing-extensions>=4.6.1:devel/py-typing-extensions@${PY_FLAVOR} \
		${PYTHON_PKGNAMEPREFIX}zstandard>=0:archivers/py-zstandard@${PY_FLAVOR}
# missing TEST_DEPENDS: jaxlib, joblibspark, py7zr, pyspark, tensorflow

USES=		python
USE_PYTHON=	pep517 concurrent autoplist pytest

TEST_ENV=	${MAKE_ENV} PYTHONPATH=${STAGEDIR}${PYTHONPREFIX_SITELIBDIR}

NO_ARCH=	yes

pre-test: # prevent failure due to missing pyspark
	@${RM} ${WRKSRC}/tests/packaged_modules/test_spark.py

.include <bsd.port.mk>