nixpkgs/pkgs/development/python-modules/datasets/default.nix

78 lines
1.5 KiB
Nix

{ lib
, aiohttp
, buildPythonPackage
, dill
, fetchFromGitHub
, fetchpatch
, fsspec
, huggingface-hub
, importlib-metadata
, multiprocess
, numpy
, packaging
, pandas
, pyarrow
, pythonOlder
, requests
, responses
, tqdm
, xxhash
}:
buildPythonPackage rec {
pname = "datasets";
version = "2.12.0";
format = "setuptools";
disabled = pythonOlder "3.7";
src = fetchFromGitHub {
owner = "huggingface";
repo = pname;
rev = "refs/tags/${version}";
hash = "sha256-o/LUzRmpM4tjiCh31KoQXzU1Z/p/91uamh7G4SGnxQM=";
};
postPatch = ''
substituteInPlace setup.py \
--replace "responses<0.19" "responses"
'';
propagatedBuildInputs = [
aiohttp
dill
fsspec
huggingface-hub
multiprocess
numpy
packaging
pandas
pyarrow
requests
responses
tqdm
xxhash
] ++ lib.optionals (pythonOlder "3.8") [
importlib-metadata
];
# Tests require pervasive internet access.
doCheck = false;
# Module import will attempt to create a cache directory.
postFixup = "export HF_MODULES_CACHE=$TMPDIR";
pythonImportsCheck = [
"datasets"
];
meta = with lib; {
description = "Open-access datasets and evaluation metrics for natural language processing";
homepage = "https://github.com/huggingface/datasets";
changelog = "https://github.com/huggingface/datasets/releases/tag/${version}";
license = licenses.asl20;
platforms = platforms.unix;
maintainers = with maintainers; [ ];
};
}