python3Packages.datasets: init at 1.0.0

This library provides open-access datasets an evaluation metrics for
natural language processing using PyTorch, TensorFlow, NumPy, and
Pandas.

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

@@ -0,0 +1,52 @@
+{ lib
+, buildPythonPackage
+, fetchFromGitHub
+, dill
+, filelock
+, numpy
+, pandas
+, pyarrow
+, requests
+, tqdm
+, xxhash
+}:
+
+buildPythonPackage rec {
+  pname = "datasets";
+  version = "1.0.0";
+
+  src = fetchFromGitHub {
+    owner = "huggingface";
+    repo = pname;
+    rev = version;
+    sha256 = "13l52r7nhj2c1a10isy5309d2g6pmaivyqs5w6yjbjj4195jxya5";
+  };
+
+  propagatedBuildInputs = [
+    dill
+    filelock
+    numpy
+    pandas
+    pyarrow
+    requests
+    tqdm
+    xxhash
+  ];
+
+  # Tests require pervasive internet access.
+  doCheck = false;
+
+  # Module import will attempt to create a cache directory.
+  postFixup = "export HF_MODULES_CACHE=$TMPDIR";
+
+  pythonImportsCheck = [ "datasets" ];
+
+  meta = with lib; {
+    homepage = "https://github.com/huggingface/datasets";
+    description = "Fast, efficient, open-access datasets and evaluation metrics for natural language processing";
+    changelog = "https://github.com/huggingface/datasets/releases/tag/${version}";
+    license = licenses.asl20;
+    platforms = platforms.unix;
+    maintainers = with maintainers; [ danieldk ];
+  };
+}

pkgs/top-level/python-packages.nix

@@ -1436,6 +1436,8 @@ in {
 
   datamodeldict = callPackage ../development/python-modules/datamodeldict { };
 
+  datasets = callPackage ../development/python-modules/datasets { };
+
   datasette = callPackage ../development/python-modules/datasette { };
 
   datashader = callPackage ../development/python-modules/datashader { };