{ stdenv , buildPythonPackage , fetchPypi , dask , numpy, toolz # dask[array] , numba , pandas , scikitlearn , scipy , dask-glm , six , multipledispatch , packaging , pytest , xgboost , tensorflow , joblib , distributed }: buildPythonPackage rec { version = "1.0.0"; pname = "dask-ml"; src = fetchPypi { inherit pname version; sha256 = "dde926478653bd03a3fbc501d3873a1534836608217b94d04320d1e1c07e59dc"; }; checkInputs = [ pytest xgboost tensorflow joblib distributed ]; propagatedBuildInputs = [ dask numpy toolz numba pandas scikitlearn scipy dask-glm six multipledispatch packaging ]; # dask-ml has some heavy test requirements # and requires some very new packages doCheck = false; meta = with stdenv.lib; { homepage = https://github.com/dask/dask-ml; description = "Scalable Machine Learn with Dask"; license = licenses.bsd3; maintainers = [ maintainers.costrouc ]; }; }