Fix isort config, fix build scripts using dask-cudf instead of dask_cudf

ci/gpu/build.sh

@@ -61,7 +61,7 @@ conda list
 ################################################################################
 
 logger "Build libcudf..."
-$WORKSPACE/build.sh clean libcudf cudf dask-cudf
+$WORKSPACE/build.sh clean libcudf cudf dask_cudf
 
 ################################################################################
 # TEST - Run GoogleTest and py.tests for libcudf and cuDF

conda/recipes/dask-cudf/build.sh

@@ -1,4 +1,4 @@
 # Copyright (c) 2018-2019, NVIDIA CORPORATION.
 
 # This assumes the script is executed from the root of the repo directory
-./build.sh dask-cudf
+./build.sh dask_cudf

python/cudf/cudf/__init__.py

@@ -1,5 +1,7 @@
 # Copyright (c) 2018-2019, NVIDIA CORPORATION.
 
+from librmm_cffi import librmm as rmm
+
 from cudf import dataframe, datasets
 from cudf._version import get_versions
 from cudf.dataframe import (
@@ -36,7 +38,6 @@ from cudf.ops import (
 )
 from cudf.reshape import get_dummies, melt
 from cudf.settings import set_options
-from librmm_cffi import librmm as rmm
 
 __version__ = get_versions()["version"]
 del get_versions

python/cudf/cudf/_sort.py

@@ -4,13 +4,15 @@
 import collections
 import logging
 
-import cudf.bindings.sort as cpp_sort
 import numpy as np
+
+from librmm_cffi import librmm as rmm
+
+import cudf.bindings.sort as cpp_sort
 from cudf.dataframe import columnops
 from cudf.dataframe.buffer import Buffer
 from cudf.dataframe.column import Column
 from cudf.utils import cudautils
-from librmm_cffi import librmm as rmm
 
 logging.basicConfig(format="%(levelname)s:%(message)s")
 

python/cudf/cudf/comm/gpu_ipc_broker.py

@@ -7,9 +7,10 @@ import pickle
 import socket
 import threading
 
-from librmm_cffi import librmm as rmm
 from numba import cuda
 
+from librmm_cffi import librmm as rmm
+
 try:
     import zmq
 

python/cudf/cudf/dataframe/buffer.py

@@ -1,7 +1,9 @@
 import numpy as np
+
+from librmm_cffi import librmm as rmm
+
 from cudf.comm.serialize import register_distributed_serializer
 from cudf.utils import cudautils, utils
-from librmm_cffi import librmm as rmm
 
 
 class Buffer(object):

python/cudf/cudf/dataframe/categorical.py

@@ -1,10 +1,11 @@
 # Copyright (c) 2018, NVIDIA CORPORATION.
 
-import cudf.bindings.copying as cpp_copying
-import cudf.bindings.replace as cpp_replace
 import numpy as np
 import pandas as pd
 import pyarrow as pa
+
+import cudf.bindings.copying as cpp_copying
+import cudf.bindings.replace as cpp_replace
 from cudf.comm.serialize import register_distributed_serializer
 from cudf.dataframe import columnops, numerical
 from cudf.dataframe.buffer import Buffer

python/cudf/cudf/dataframe/column.py

@@ -6,18 +6,20 @@ LibGDF operates on column.
 """
 from numbers import Number
 
+import numpy as np
+import pandas as pd
+from numba.cuda.cudadrv.devicearray import DeviceNDArray
+
+import nvstrings
+from librmm_cffi import librmm as rmm
+
 import cudf.bindings.copying as cpp_copying
 import cudf.bindings.quantile as cpp_quantile
-import numpy as np
-import nvstrings
-import pandas as pd
 from cudf.bindings.concat import _column_concat
 from cudf.bindings.cudf_cpp import column_view_pointer, count_nonzero_mask
 from cudf.comm.serialize import register_distributed_serializer
 from cudf.dataframe.buffer import Buffer
 from cudf.utils import cudautils, ioutils, utils
-from librmm_cffi import librmm as rmm
-from numba.cuda.cudadrv.devicearray import DeviceNDArray
 
 
 class Column(object):

python/cudf/cudf/dataframe/columnops.py

@@ -5,11 +5,15 @@ view of Columns.
 
 import warnings
 
-import cudf.bindings.copying as cpp_copying
 import numpy as np
-import nvstrings
 import pandas as pd
 import pyarrow as pa
+from numba import cuda, njit
+
+import nvstrings
+from librmm_cffi import librmm as rmm
+
+import cudf.bindings.copying as cpp_copying
 from cudf.bindings.cudf_cpp import np_to_pa_dtype
 from cudf.bindings.stream_compaction import (
     cpp_apply_boolean_mask,
@@ -19,8 +23,6 @@ from cudf.dataframe.buffer import Buffer
 from cudf.dataframe.column import Column
 from cudf.utils import cudautils, utils
 from cudf.utils.utils import buffers_from_pyarrow, min_scalar_type
-from librmm_cffi import librmm as rmm
-from numba import cuda, njit
 
 
 class TypedColumnBase(Column):

python/cudf/cudf/dataframe/dataframe.py

@@ -12,12 +12,16 @@ from collections import OrderedDict
 from collections.abc import Mapping, Sequence
 from types import GeneratorType
 
-import cudf
-import cudf.bindings.hash as cpp_hash
-import cudf.bindings.join as cpp_join
 import numpy as np
 import pandas as pd
 import pyarrow as pa
+from pandas.api.types import is_dict_like
+
+from librmm_cffi import librmm as rmm
+
+import cudf
+import cudf.bindings.hash as cpp_hash
+import cudf.bindings.join as cpp_join
 from cudf import formatting
 from cudf._sort import get_sorted_inds
 from cudf.bindings import copying as cpp_copying
@@ -36,8 +40,6 @@ from cudf.settings import NOTSET, settings
 from cudf.utils import applyutils, cudautils, ioutils, queryutils, utils
 from cudf.utils.docutils import copy_docstring
 from cudf.window import Rolling
-from librmm_cffi import librmm as rmm
-from pandas.api.types import is_dict_like
 
 
 def _unique_name(existing_names, suffix="_unique_name"):

python/cudf/cudf/dataframe/datetime.py

@@ -1,13 +1,14 @@
 import datetime as dt
 
+import numpy as np
+import pandas as pd
+import pyarrow as pa
+
 import cudf.bindings.binops as cpp_binops
 import cudf.bindings.copying as cpp_copying
 import cudf.bindings.reduce as cpp_reduce
 import cudf.bindings.replace as cpp_replace
 import cudf.bindings.unaryops as cpp_unaryops
-import numpy as np
-import pandas as pd
-import pyarrow as pa
 from cudf._sort import get_sorted_inds
 from cudf.bindings.cudf_cpp import get_ctype_ptr, np_to_pa_dtype
 from cudf.bindings.nvtx import nvtx_range_pop, nvtx_range_push

python/cudf/cudf/dataframe/index.py

@@ -5,11 +5,15 @@ from __future__ import division, print_function
 import pickle
 from copy import copy, deepcopy
 
+import numpy as np
+import pandas as pd
+from numba.cuda.cudadrv.devicearray import DeviceNDArray
+
+import nvstrings
+from librmm_cffi import librmm as rmm
+
 import cudf
 import cudf.bindings.copying as cpp_copying
-import numpy as np
-import nvstrings
-import pandas as pd
 from cudf.comm.serialize import register_distributed_serializer
 from cudf.dataframe import columnops
 from cudf.dataframe.buffer import Buffer
@@ -20,8 +24,6 @@ from cudf.dataframe.numerical import NumericalColumn
 from cudf.dataframe.string import StringColumn
 from cudf.indexing import _IndexLocIndexer
 from cudf.utils import cudautils, ioutils, utils
-from librmm_cffi import librmm as rmm
-from numba.cuda.cudadrv.devicearray import DeviceNDArray
 
 
 class Index(object):

python/cudf/cudf/dataframe/multiindex.py

@@ -6,6 +6,7 @@ from collections.abc import Sequence
 
 import numpy as np
 import pandas as pd
+
 from cudf.comm.serialize import register_distributed_serializer
 from cudf.dataframe import columnops
 from cudf.dataframe.index import Index, StringIndex, as_index

python/cudf/cudf/dataframe/numerical.py

@@ -2,6 +2,13 @@
 
 from __future__ import division, print_function
 
+import numpy as np
+import pandas as pd
+import pyarrow as pa
+from pandas.api.types import is_integer_dtype
+
+from librmm_cffi import librmm as rmm
+
 import cudf.bindings.binops as cpp_binops
 import cudf.bindings.copying as cpp_copying
 import cudf.bindings.hash as cpp_hash
@@ -9,9 +16,6 @@ import cudf.bindings.reduce as cpp_reduce
 import cudf.bindings.replace as cpp_replace
 import cudf.bindings.sort as cpp_sort
 import cudf.bindings.unaryops as cpp_unaryops
-import numpy as np
-import pandas as pd
-import pyarrow as pa
 from cudf._sort import get_sorted_inds
 from cudf.bindings.cudf_cpp import get_ctype_ptr, np_to_pa_dtype
 from cudf.bindings.nvtx import nvtx_range_pop, nvtx_range_push
@@ -19,8 +23,6 @@ from cudf.comm.serialize import register_distributed_serializer
 from cudf.dataframe import columnops, datetime, string
 from cudf.dataframe.buffer import Buffer
 from cudf.utils import cudautils, utils
-from librmm_cffi import librmm as rmm
-from pandas.api.types import is_integer_dtype
 
 
 class NumericalColumn(columnops.TypedColumnBase):

python/cudf/cudf/dataframe/series.py

@@ -5,9 +5,14 @@ import warnings
 from collections import OrderedDict
 from numbers import Number
 
-import cudf.bindings.copying as cpp_copying
 import numpy as np
 import pandas as pd
+from numba.cuda.cudadrv.devicearray import DeviceNDArray
+from pandas.api.types import is_dict_like, is_scalar
+
+from librmm_cffi import librmm as rmm
+
+import cudf.bindings.copying as cpp_copying
 from cudf import formatting
 from cudf.bindings.nvtx import nvtx_range_pop, nvtx_range_push
 from cudf.comm.serialize import register_distributed_serializer
@@ -21,9 +26,6 @@ from cudf.settings import NOTSET, settings
 from cudf.utils import cudautils, ioutils, utils
 from cudf.utils.docutils import copy_docstring
 from cudf.window import Rolling
-from librmm_cffi import librmm as rmm
-from numba.cuda.cudadrv.devicearray import DeviceNDArray
-from pandas.api.types import is_dict_like, is_scalar
 
 
 class Series(object):

python/cudf/cudf/dataframe/string.py

@@ -3,19 +3,21 @@
 import warnings
 from numbers import Number
 
-import cudf.bindings.binops as cpp_binops
 import numpy as np
-import nvstrings
 import pandas as pd
 import pyarrow as pa
+from numba.cuda.cudadrv.devicearray import DeviceNDArray
+
+import nvstrings
+from librmm_cffi import librmm as rmm
+
+import cudf.bindings.binops as cpp_binops
 from cudf.bindings.cudf_cpp import get_ctype_ptr
 from cudf.bindings.nvtx import nvtx_range_pop, nvtx_range_push
 from cudf.comm.serialize import register_distributed_serializer
 from cudf.dataframe import column, columnops, numerical, series
 from cudf.dataframe.buffer import Buffer
 from cudf.utils import cudautils, utils
-from librmm_cffi import librmm as rmm
-from numba.cuda.cudadrv.devicearray import DeviceNDArray
 
 _str_to_numeric_typecast_functions = {
     np.dtype("int32"): nvstrings.nvstrings.stoi,

python/cudf/cudf/datasets.py

@@ -1,7 +1,8 @@
-import cudf
 import numpy as np
 import pandas as pd
 
+import cudf
+
 __all__ = ["timeseries", "randomdata"]
 
 

python/cudf/cudf/groupby/legacy_groupby.py

@@ -4,6 +4,10 @@ from collections import OrderedDict, defaultdict, namedtuple
 from itertools import chain
 
 import numpy as np
+from numba import cuda
+
+from librmm_cffi import librmm as rmm
+
 from cudf.bindings.sort import apply_segsort
 from cudf.comm.serialize import register_distributed_serializer
 from cudf.dataframe.buffer import Buffer
@@ -12,8 +16,6 @@ from cudf.dataframe.dataframe import DataFrame
 from cudf.dataframe.series import Series
 from cudf.multi import concat
 from cudf.utils import cudautils
-from librmm_cffi import librmm as rmm
-from numba import cuda
 
 
 def _auto_generate_grouper_agg(members):

python/cudf/cudf/indexing.py

@@ -1,9 +1,10 @@
-import cudf
 import numpy as np
 import pandas as pd
-from cudf.utils.utils import is_single_value
 from numba.cuda.cudadrv.devicearray import DeviceNDArray
 
+import cudf
+from cudf.utils.utils import is_single_value
+
 
 class _SeriesLocIndexer(object):
     """

python/cudf/cudf/io/feather.py

@@ -2,9 +2,10 @@
 
 import warnings
 
+from pyarrow import feather
+
 from cudf.dataframe.dataframe import DataFrame
 from cudf.utils import ioutils
-from pyarrow import feather
 
 
 @ioutils.doc_read_feather()

python/cudf/cudf/io/hdf.py

@@ -2,8 +2,9 @@
 
 import warnings
 
-import cudf
 import pandas as pd
+
+import cudf
 from cudf.utils import ioutils
 
 

python/cudf/cudf/io/json.py

@@ -3,8 +3,9 @@
 import warnings
 from io import BytesIO, StringIO
 
-import cudf
 import pandas as pd
+
+import cudf
 from cudf.bindings.json import cpp_read_json
 from cudf.utils import ioutils
 

python/cudf/cudf/io/orc.py

@@ -3,6 +3,7 @@
 import warnings
 
 import pyarrow.orc as orc
+
 from cudf.bindings.orc import cpp_read_orc
 from cudf.dataframe.dataframe import DataFrame
 from cudf.utils import ioutils

python/cudf/cudf/io/parquet.py

@@ -3,6 +3,7 @@
 import warnings
 
 import pyarrow.parquet as pq
+
 from cudf.bindings.parquet import cpp_read_parquet
 from cudf.dataframe.dataframe import DataFrame
 from cudf.utils import ioutils

python/cudf/cudf/reshape/general.py

@@ -2,6 +2,7 @@
 
 import numpy as np
 import pandas as pd
+
 from cudf.dataframe import Buffer, DataFrame, Series
 from cudf.dataframe.categorical import CategoricalColumn
 from cudf.utils import cudautils

python/cudf/cudf/tests/conftest.py

@@ -1,6 +1,7 @@
 import pathlib
 
 import pytest
+
 from librmm_cffi import librmm
 
 

python/cudf/cudf/tests/test_applymap.py

@@ -5,6 +5,7 @@ from math import floor
 
 import numpy as np
 import pytest
+
 from cudf import Series
 from cudf.tests import utils
 

python/cudf/cudf/tests/test_binops.py

@@ -6,10 +6,11 @@ import operator
 import random
 from itertools import product
 
-import cudf
 import numpy as np
 import pandas as pd
 import pytest
+
+import cudf
 from cudf.dataframe import Series
 from cudf.dataframe.index import as_index
 from cudf.tests import utils

python/cudf/cudf/tests/test_categorical.py

@@ -3,6 +3,7 @@
 import numpy as np
 import pandas as pd
 import pytest
+
 from cudf.dataframe import DataFrame, Series
 from cudf.tests.utils import assert_eq
 

python/cudf/cudf/tests/test_copying.py

@@ -1,10 +1,12 @@
 from __future__ import division, print_function
 
-import cudf.bindings.copying as cpp_copying
 import numpy as np
-from cudf.dataframe import columnops
+
 from librmm_cffi import librmm as rmm
 
+import cudf.bindings.copying as cpp_copying
+from cudf.dataframe import columnops
+
 
 def test_gather_single_col():
     col = columnops.as_column(np.arange(100), dtype=np.int32)

python/cudf/cudf/tests/test_csv.py

@@ -7,10 +7,11 @@ from collections import OrderedDict
 from io import BytesIO, StringIO
 from pathlib import Path
 
-import cudf
 import numpy as np
 import pandas as pd
 import pytest
+
+import cudf
 from cudf import read_csv
 from cudf.tests.utils import assert_eq
 

python/cudf/cudf/tests/test_cuda_apply.py

@@ -6,9 +6,10 @@ Test method that apply GPU kernel to a frame.
 
 import numpy as np
 import pytest
-from cudf import DataFrame
 from numba import cuda
 
+from cudf import DataFrame
+
 
 @pytest.mark.parametrize("nelem", [1, 2, 64, 128, 1000, 5000])
 def test_df_apply_rows(nelem):

python/cudf/cudf/tests/test_dask.py

@@ -1,10 +1,12 @@
 # Copyright (c) 2019, NVIDIA CORPORATION.
 
-import cudf
 import pytest
+
 from dask.dataframe.utils import is_dataframe_like  # noqa: E402
 from dask.dataframe.utils import is_index_like, is_series_like
 
+import cudf
+
 dask = pytest.importorskip(
     "dask",
     minversion="1.1.0",

python/cudf/cudf/tests/test_dataframe.py

@@ -4,17 +4,19 @@ import array as arr
 import operator
 from itertools import combinations
 
-import cudf as gd
 import numpy as np
 import pandas as pd
 import pyarrow as pa
 import pytest
+
+from librmm_cffi import librmm as rmm
+
+import cudf as gd
 from cudf.dataframe.buffer import Buffer
 from cudf.dataframe.dataframe import DataFrame, Series
 from cudf.settings import set_options
 from cudf.tests import utils
 from cudf.tests.utils import assert_eq, gen_rand
-from librmm_cffi import librmm as rmm
 
 
 def test_buffer_basic():

python/cudf/cudf/tests/test_dataframe_copy.py

@@ -6,9 +6,10 @@ from copy import deepcopy  # noqa:F401
 import numpy as np
 import pandas as pd
 import pytest
+from numba import cuda
+
 from cudf.dataframe.dataframe import DataFrame
 from cudf.tests.utils import assert_eq
-from numba import cuda
 
 
 """

python/cudf/cudf/tests/test_datasets.py

@@ -1,5 +1,6 @@
-import cudf as gd
 import pandas as pd
+
+import cudf as gd
 from cudf.tests.utils import assert_eq
 
 

python/cudf/cudf/tests/test_datetime.py

@@ -1,18 +1,19 @@
 import datetime as dt
 
-import cudf
 import numpy as np
 import pandas as pd
 import pytest
-from cudf.dataframe import DataFrame, Series
-from cudf.dataframe.index import DatetimeIndex
-from cudf.tests.utils import assert_eq
 from pandas.util.testing import (
     assert_frame_equal,
     assert_index_equal,
     assert_series_equal,
 )
 
+import cudf
+from cudf.dataframe import DataFrame, Series
+from cudf.dataframe.index import DatetimeIndex
+from cudf.tests.utils import assert_eq
+
 
 def data1():
     return pd.date_range("20010101", "20020215", freq="400h")

python/cudf/cudf/tests/test_dlpack.py

@@ -3,9 +3,10 @@
 import itertools
 from contextlib import ExitStack as does_not_raise
 
-import cudf
 import numpy as np
 import pytest
+
+import cudf
 from cudf.tests.utils import assert_eq
 
 try:

python/cudf/cudf/tests/test_dropna.py

@@ -1,7 +1,8 @@
-import cudf
 import numpy as np
 import pandas as pd
 import pytest
+
+import cudf
 from cudf.tests.utils import assert_eq
 
 

python/cudf/cudf/tests/test_duplicates.py

@@ -1,9 +1,10 @@
-import cudf
 import numpy as np
 import pytest
+from pandas import DataFrame, date_range
+
+import cudf
 from cudf.multi import concat
 from cudf.tests.utils import assert_eq
-from pandas import DataFrame, date_range
 
 
 def assert_df(g, p):

python/cudf/cudf/tests/test_factorize.py

@@ -2,6 +2,7 @@
 
 import numpy as np
 import pytest
+
 from cudf.dataframe import DataFrame, Series
 
 

python/cudf/cudf/tests/test_feather.py

@@ -4,11 +4,12 @@ import os
 from distutils.version import LooseVersion
 from string import ascii_letters
 
-import cudf
 import numpy as np
 import pandas as pd
 import pyarrow as pa
 import pytest
+
+import cudf
 from cudf.tests.utils import assert_eq
 
 if LooseVersion(pd.__version__) < LooseVersion("0.24"):

python/cudf/cudf/tests/test_generic_ptx.py

@@ -5,11 +5,12 @@ from __future__ import division
 import numba
 import numpy as np
 import pytest
-from cudf.bindings import binops
-from cudf.dataframe import Series
 from numba import cuda, types
 from packaging.version import Version
 
+from cudf.bindings import binops
+from cudf.dataframe import Series
+
 
 @pytest.mark.skipif(
     Version(numba.__version__) < Version("0.44.0a"),

python/cudf/cudf/tests/test_gpu_arrow_parser.py

@@ -3,9 +3,11 @@ import logging
 
 import numpy as np
 import pytest
-from cudf.comm.gpuarrow import GpuArrowReader
+
 from librmm_cffi import librmm as rmm
 
+from cudf.comm.gpuarrow import GpuArrowReader
+
 try:
     import pyarrow as pa
 

python/cudf/cudf/tests/test_groupby.py

@@ -1,12 +1,13 @@
 # Copyright (c) 2018, NVIDIA CORPORATION.
 
-import cudf
 import numpy as np
 import pandas as pd
 import pytest
+from numpy.testing import assert_array_equal
+
+import cudf
 from cudf.dataframe import DataFrame, Series
 from cudf.tests.utils import assert_eq
-from numpy.testing import assert_array_equal
 
 
 def make_frame(dataframe_class, nelem, seed=0, extra_levels=(), extra_vals=()):

python/cudf/cudf/tests/test_hdf.py

@@ -3,10 +3,11 @@
 import os
 from string import ascii_letters
 
-import cudf
 import numpy as np
 import pandas as pd
 import pytest
+
+import cudf
 from cudf.tests.utils import assert_eq
 
 try:

python/cudf/cudf/tests/test_index.py

@@ -6,6 +6,7 @@ Test related to Index
 import numpy as np
 import pandas as pd
 import pytest
+
 from cudf.dataframe import DataFrame
 from cudf.dataframe.index import (
     CategoricalIndex,

python/cudf/cudf/tests/test_joining.py

@@ -2,10 +2,11 @@
 
 from timeit import default_timer as timer
 
-import cudf
 import numpy as np
 import pandas as pd
 import pytest
+
+import cudf
 from cudf.dataframe import DataFrame
 from cudf.tests.utils import assert_eq
 

python/cudf/cudf/tests/test_json.py

@@ -6,10 +6,11 @@ import os
 from io import BytesIO, StringIO
 from pathlib import Path
 
-import cudf
 import numpy as np
 import pandas as pd
 import pytest
+
+import cudf
 from cudf.tests.utils import assert_eq
 
 

python/cudf/cudf/tests/test_label_encode.py

@@ -5,6 +5,7 @@ from itertools import product
 
 import numpy as np
 import pytest
+
 from cudf.dataframe import DataFrame
 
 

python/cudf/cudf/tests/test_libgdf_groupby.py

@@ -3,6 +3,7 @@
 import numpy as np
 import pandas as pd
 import pytest
+
 from cudf.dataframe import DataFrame
 from cudf.tests.utils import assert_eq
 

python/cudf/cudf/tests/test_melt.py

@@ -1,6 +1,7 @@
 import numpy as np
 import pandas as pd
 import pytest
+
 from cudf import melt as cudf_melt
 from cudf.dataframe import DataFrame
 

python/cudf/cudf/tests/test_multi.py

@@ -1,9 +1,10 @@
 # Copyright (c) 2018, NVIDIA CORPORATION.
 
-import cudf as gd
 import numpy as np
 import pandas as pd
 import pytest
+
+import cudf as gd
 from cudf.tests.utils import assert_eq
 
 

python/cudf/cudf/tests/test_multiindex.py

@@ -3,10 +3,11 @@
 """
 Test related to MultiIndex
 """
-import cudf
 import numpy as np
 import pandas as pd
 import pytest
+
+import cudf
 from cudf.tests.utils import assert_eq, assert_neq
 
 

python/cudf/cudf/tests/test_numpy_interop.py

@@ -1,5 +1,6 @@
 import numpy as np
 import pytest
+
 from cudf.dataframe import DataFrame
 
 

python/cudf/cudf/tests/test_onehot.py

@@ -2,6 +2,7 @@
 
 import numpy as np
 import pytest
+
 from cudf.dataframe import DataFrame, GenericIndex, Series
 from cudf.reshape import get_dummies
 from cudf.tests import utils

python/cudf/cudf/tests/test_ops.py

@@ -1,7 +1,8 @@
-import cudf
 import numpy as np
 import pandas as pd
 import pytest
+
+import cudf
 from cudf.tests.utils import assert_eq, gen_rand
 
 

python/cudf/cudf/tests/test_orc.py

@@ -3,11 +3,12 @@
 import os
 from io import BytesIO
 
-import cudf
 import numpy as np
 import pandas as pd
 import pyarrow as pa
 import pytest
+
+import cudf
 from cudf.tests.utils import assert_eq
 
 

python/cudf/cudf/tests/test_pandas_interop.py

@@ -2,6 +2,7 @@
 
 import numpy as np
 import pandas as pd
+
 from cudf.dataframe import DataFrame
 
 

python/cudf/cudf/tests/test_parquet.py

@@ -6,11 +6,12 @@ from distutils.version import LooseVersion
 from io import BytesIO
 from string import ascii_letters
 
-import cudf
 import numpy as np
 import pandas as pd
 import pyarrow as pa
 import pytest
+
+import cudf
 from cudf.tests.utils import assert_eq
 
 

python/cudf/cudf/tests/test_pickling.py

@@ -5,9 +5,11 @@ import sys
 
 import numpy as np
 import pandas as pd
+
+from librmm_cffi import librmm as rmm
+
 from cudf.dataframe import DataFrame, GenericIndex
 from cudf.dataframe.buffer import Buffer
-from librmm_cffi import librmm as rmm
 
 
 def check_serialization(df):

python/cudf/cudf/tests/test_query.py

@@ -6,14 +6,15 @@ import datetime
 import inspect
 from itertools import product
 
-import cudf
 import numpy as np
 import pandas as pd
 import pytest
+from pandas.util.testing import assert_frame_equal
+
+import cudf
 from cudf.dataframe import DataFrame
 from cudf.tests.utils import assert_eq
 from cudf.utils import queryutils
-from pandas.util.testing import assert_frame_equal
 
 _params_query_parser = []
 _params_query_parser.append(("a > @b", ("a", "__CUDF_ENVREF__b")))

python/cudf/cudf/tests/test_reductions.py

@@ -5,6 +5,7 @@ from itertools import product
 
 import numpy as np
 import pytest
+
 from cudf.dataframe import Series
 from cudf.tests import utils
 from cudf.tests.utils import gen_rand

python/cudf/cudf/tests/test_replace.py

@@ -1,6 +1,7 @@
 import numpy as np
 import pandas as pd
 import pytest
+
 from cudf.dataframe import DataFrame, Series
 from cudf.tests.utils import assert_eq
 

python/cudf/cudf/tests/test_rolling.py

@@ -1,7 +1,8 @@
-import cudf
 import numpy as np
 import pandas as pd
 import pytest
+
+import cudf
 from cudf.tests.utils import assert_eq
 
 

python/cudf/cudf/tests/test_scan.py

@@ -3,6 +3,7 @@ from itertools import product
 import numpy as np
 import pandas as pd
 import pytest
+
 from cudf.dataframe.dataframe import DataFrame, Series
 from cudf.tests.utils import assert_eq, gen_rand
 

python/cudf/cudf/tests/test_serialize.py

@@ -2,14 +2,16 @@
 
 import functools
 
-import cudf
 import msgpack  # noqa: E402
 import numpy as np
 import pandas as pd
 import pytest
+
+from distributed.protocol import deserialize, serialize  # noqa: E402
+
+import cudf
 from cudf.tests import utils
 from cudf.tests.utils import assert_eq
-from distributed.protocol import deserialize, serialize  # noqa: E402
 
 pytest.importorskip("dask.distributed")
 pytest.importorskip("distributed.protocol")

python/cudf/cudf/tests/test_sorting.py

@@ -5,6 +5,7 @@ from itertools import product
 import numpy as np
 import pandas as pd
 import pytest
+
 from cudf.dataframe import DataFrame, Series
 from cudf.tests.utils import assert_eq
 

python/cudf/cudf/tests/test_sparse_df.py

@@ -3,10 +3,12 @@ import os.path
 
 import numpy as np
 import pytest
+
+from librmm_cffi import librmm as rmm
+
 from cudf.comm.gpuarrow import GpuArrowReader
 from cudf.dataframe import DataFrame, Series
 from cudf.tests.utils import assert_eq
-from librmm_cffi import librmm as rmm
 
 try:
     import pyarrow as pa

python/cudf/cudf/tests/test_stats.py

@@ -3,6 +3,7 @@
 import numpy as np
 import pandas as pd
 import pytest
+
 from cudf.dataframe import Series
 
 params_dtypes = [np.int32, np.float32, np.float64]

python/cudf/cudf/tests/test_string.py

@@ -6,12 +6,14 @@ import numpy as np
 import pandas as pd
 import pyarrow as pa
 import pytest
+from numba import cuda
+
+from librmm_cffi import librmm as rmm
+
 from cudf import concat
 from cudf.dataframe import DataFrame, Series
 from cudf.dataframe.index import StringColumn, StringIndex
 from cudf.tests.utils import assert_eq
-from librmm_cffi import librmm as rmm
-from numba import cuda
 
 data_list = [
     ["AbC", "de", "FGHI", "j", "kLm"],

python/cudf/cudf/tests/test_unaops.py

@@ -1,9 +1,10 @@
 from __future__ import division
 
-import cudf
 import numpy as np
 import pandas as pd
 import pytest
+
+import cudf
 from cudf.dataframe import Series
 from cudf.tests import utils
 

python/cudf/cudf/tests/utils.py

@@ -1,6 +1,7 @@
 import numpy as np
 import pandas as pd
 import pandas.util.testing as tm
+
 from cudf.utils import utils
 
 

python/cudf/cudf/utils/applyutils.py

@@ -2,12 +2,14 @@
 
 import functools
 
+from numba import cuda, six
+from numba.utils import exec_, pysignature
+
+from librmm_cffi import librmm as rmm
+
 from cudf.dataframe.series import Series
 from cudf.utils import cudautils
 from cudf.utils.docutils import docfmt_partial
-from librmm_cffi import librmm as rmm
-from numba import cuda, six
-from numba.utils import exec_, pysignature
 
 _doc_applyparams = """
 func : function

python/cudf/cudf/utils/cudautils.py

@@ -3,7 +3,11 @@
 from math import fmod, isnan
 
 import numpy as np
+from numba import cuda, int32, numpy_support
+
 import nvstrings
+from librmm_cffi import librmm as rmm
+
 from cudf.utils.utils import (
     check_equals_float,
     check_equals_int,
@@ -12,8 +16,6 @@ from cudf.utils.utils import (
     mask_get,
     mask_set,
 )
-from librmm_cffi import librmm as rmm
-from numba import cuda, int32, numpy_support
 
 
 def optimal_block_count(minblkct):

python/cudf/cudf/utils/queryutils.py

@@ -5,9 +5,10 @@ import datetime as dt
 
 import numpy as np
 import six
-from librmm_cffi import librmm as rmm
 from numba import cuda
 
+from librmm_cffi import librmm as rmm
+
 ENVREF_PREFIX = "__CUDF_ENVREF__"
 
 

python/cudf/cudf/utils/utils.py

@@ -5,9 +5,10 @@ from math import ceil, isinf, isnan
 import numpy as np
 import pandas as pd
 import pyarrow as pa
-from librmm_cffi import librmm as rmm
 from numba import njit
 
+from librmm_cffi import librmm as rmm
+
 mask_dtype = np.dtype(np.int8)
 mask_bitsize = mask_dtype.itemsize * 8
 mask_byte_padding = 64

python/cudf/cudf/window/rolling.py

@@ -1,5 +1,6 @@
-import cudf
 import pandas as pd
+
+import cudf
 from cudf.bindings.rolling import apply_rolling
 from cudf.utils import cudautils
 

python/cudf/setup.cfg

@@ -14,6 +14,11 @@ parentdir_prefix = cudf-
 
 [flake8]
 exclude = __init__.py
+ignore =
+    # line break before binary operator
+    W503
+    # whitespace before :
+    E203
 
 [isort]
 line_length=79
@@ -21,6 +26,33 @@ multi_line_output=3
 include_trailing_comma=True
 force_grid_wrap=0
 combine_as_imports=True
+order_by_type=True
+known_third_party= 
+    cython,
+    msgpack,
+    numba,
+    numpy,
+    packaging,
+    pandas,
+    pyarrow,
+    pytest,
+    setuptools,
+    six,
+    toolz,
+    versioneer
+known_dask=
+    dask
+    distributed
+    dask_cuda
+known_rapids=
+    librmm_cffi
+    nvstrings
+    nvcategory
+    nvtext
+    cuml
+    cugraph
+    dask_cudf
+sections=FUTURE,STDLIB,THIRDPARTY,DASK,RAPIDS,FIRSTPARTY,LOCALFOLDER
 skip=
     thirdparty
     .eggs
@@ -32,4 +64,4 @@ skip=
     _build
     buck-out
     build
-    dist
+    dist

python/cudf/setup.py

@@ -2,10 +2,10 @@
 
 from distutils.sysconfig import get_python_lib
 
+import versioneer
 from setuptools import find_packages, setup
 from setuptools.extension import Extension
 
-import versioneer
 from Cython.Build import cythonize
 
 install_requires = ["numba", "cython"]

python/dask_cudf/dask_cudf/accessor.py

@@ -12,12 +12,13 @@ accessor properties.
 
 """
 
+from toolz import partial
+
 import cudf
 from cudf.dataframe.categorical import (
     CategoricalAccessor as GdfCategoricalAccessor,
 )
 from cudf.dataframe.series import DatetimeProperties
-from toolz import partial
 
 # Adapted from
 # https://github.com/dask/dask/blob/master/dask/dataframe/accessor.py

python/dask_cudf/dask_cudf/backends.py

@@ -1,7 +1,8 @@
-import cudf
 from dask.dataframe.core import get_parallel_type, make_meta, meta_nonempty
 from dask.dataframe.methods import concat_dispatch
 
+import cudf
+
 from .core import DataFrame, Index, Series
 
 get_parallel_type.register(cudf.DataFrame, lambda _: DataFrame)

python/dask_cudf/dask_cudf/batcher_sortnet.py

@@ -4,9 +4,10 @@ Adapted from https://en.wikipedia.org/wiki/Batcher_odd%E2%80%93even_mergesort
 """
 import math
 
-import cudf as gd
 from dask import compute, delayed
 
+import cudf as gd
+
 
 def get_oversized(length):
     """

python/dask_cudf/dask_cudf/core.py

@@ -2,12 +2,12 @@
 import warnings
 from collections import OrderedDict
 
-import cudf
-import cudf.bindings.reduce as cpp_reduce
-import dask
-import dask.dataframe as dd
 import numpy as np
 import pandas as pd
+from toolz import partition_all
+
+import dask
+import dask.dataframe as dd
 from dask import compute
 from dask.base import normalize_token, tokenize
 from dask.compatibility import apply
@@ -19,13 +19,16 @@ from dask.dataframe.utils import raise_on_meta_error
 from dask.delayed import delayed
 from dask.optimization import cull, fuse
 from dask.utils import M, OperatorMethodMixin, derived_from, funcname
+
+import cudf
+import cudf.bindings.reduce as cpp_reduce
+
 from dask_cudf import batcher_sortnet, join_impl
 from dask_cudf.accessor import (
     CachedAccessor,
     CategoricalAccessor,
     DatetimeAccessor,
 )
-from toolz import partition_all
 
 
 def optimize(dsk, keys, **kwargs):

python/dask_cudf/dask_cudf/io/csv.py

@@ -2,14 +2,15 @@ import os
 from glob import glob
 from warnings import warn
 
-import cudf
 import dask.dataframe as dd
-from cudf.bindings.GDFError import GDFError
 from dask.base import tokenize
 from dask.compatibility import apply
 from dask.dataframe.io.csv import make_reader
 from dask.utils import parse_bytes
 
+import cudf
+from cudf.bindings.GDFError import GDFError
+
 
 def read_csv(path, chunksize="256 MiB", **kwargs):
     if "://" in str(path):

python/dask_cudf/dask_cudf/io/json.py

@@ -1,6 +1,7 @@
 from functools import partial
 
-import cudf
 import dask
 
+import cudf
+
 read_json = partial(dask.dataframe.read_json, engine=cudf.read_json)

python/dask_cudf/dask_cudf/io/orc.py

@@ -1,10 +1,11 @@
 from glob import glob
 
-import cudf
 import dask.dataframe as dd
 from dask.base import tokenize
 from dask.compatibility import apply
 
+import cudf
+
 
 def read_orc(path, **kwargs):
     """ Read ORC files into a Dask DataFrame

python/dask_cudf/dask_cudf/io/parquet.py

@@ -1,11 +1,12 @@
 from glob import glob
 
-import cudf
 import dask.dataframe as dd
 from dask.base import tokenize
 from dask.compatibility import apply
 from dask.utils import natural_sort_key
 
+import cudf
+
 
 def read_parquet(path, **kwargs):
     """ Read parquet files into a Dask DataFrame

python/dask_cudf/dask_cudf/io/tests/test_csv.py

@@ -1,12 +1,14 @@
 import warnings
 
-import dask
-import dask.dataframe as dd
-import dask_cudf
 import numpy as np
 import pandas as pd
 import pytest
 
+import dask
+import dask.dataframe as dd
+
+import dask_cudf
+
 
 def test_read_csv(tmp_path):
     df = dask.datasets.timeseries(

python/dask_cudf/dask_cudf/io/tests/test_json.py

@@ -1,10 +1,12 @@
-import dask
-import dask.dataframe as dd
-import dask_cudf
 import pandas as pd
 import pytest
+
+import dask
+import dask.dataframe as dd
 from dask.utils import tmpfile
 
+import dask_cudf
+
 
 def test_read_json(tmp_path):
     df1 = dask.datasets.timeseries(

python/dask_cudf/dask_cudf/io/tests/test_orc.py

@@ -1,10 +1,13 @@
 import os
 
-import cudf
-import dask.dataframe as dd
-import dask_cudf
 import pytest
 
+import dask.dataframe as dd
+
+import cudf
+
+import dask_cudf
+
 # import pyarrow.orc as orc
 
 cur_dir = os.path.dirname(__file__)

python/dask_cudf/dask_cudf/io/tests/test_parquet.py

@@ -1,11 +1,13 @@
 import os
 
-import dask.dataframe as dd
-import dask_cudf
 import pandas as pd
+
+import dask.dataframe as dd
 from dask.dataframe.utils import assert_eq
 from dask.utils import natural_sort_key
 
+import dask_cudf
+
 nrows = 40
 npartitions = 15
 df = pd.DataFrame(

python/dask_cudf/dask_cudf/io/tests/test_s3.py

@@ -1,10 +1,12 @@
 import os
 from contextlib import contextmanager
 
-import dask_cudf
 import pytest
+
 from dask.bytes.s3 import DaskS3FileSystem
 
+import dask_cudf
+
 s3fs = pytest.importorskip("s3fs")
 boto3 = pytest.importorskip("boto3")
 moto = pytest.importorskip("moto")

python/dask_cudf/dask_cudf/join_impl.py

@@ -1,7 +1,8 @@
-import cudf
 import dask.dataframe as dd
 from dask import delayed
 
+import cudf
+
 
 @delayed
 def local_shuffle(frame, num_new_parts, key_columns):

python/dask_cudf/dask_cudf/tests/test_accessor.py

@@ -1,10 +1,12 @@
-import dask_cudf as dgd
 import numpy as np
 import pandas as pd
 import pytest
-from cudf.dataframe import Series
 from pandas.util.testing import assert_series_equal
 
+from cudf.dataframe import Series
+
+import dask_cudf as dgd
+
 #############################################################################
 #                        Datetime Accessor                                  #
 #############################################################################

python/dask_cudf/dask_cudf/tests/test_batcher_sortnet.py

@@ -1,6 +1,8 @@
-import cudf
 import numpy as np
 import pytest
+
+import cudf
+
 from dask_cudf import batcher_sortnet
 
 

python/dask_cudf/dask_cudf/tests/test_binops.py

@@ -1,11 +1,13 @@
 import operator
 
-import cudf
-import dask.dataframe as dd
 import numpy as np
 import pandas as pd
 import pytest
 
+import dask.dataframe as dd
+
+import cudf
+
 
 def _make_empty_frame(npartitions=2):
     df = pd.DataFrame({"x": [], "y": []})

python/dask_cudf/dask_cudf/tests/test_core.py

@@ -1,12 +1,15 @@
-import cudf
-import dask
-import dask.dataframe as dd
-import dask_cudf as dgd
 import numpy as np
 import pandas as pd
 import pandas.util.testing as tm
 import pytest
 
+import dask
+import dask.dataframe as dd
+
+import cudf
+
+import dask_cudf as dgd
+
 
 def test_from_cudf():
     np.random.seed(0)

python/dask_cudf/dask_cudf/tests/test_delayed_io.py

@@ -1,13 +1,16 @@
 """
 Test IO with dask.delayed API
 """
-import cudf as gd
-import dask_cudf as dgd
 import numpy as np
 import pytest
-from dask.delayed import delayed
 from pandas.util.testing import assert_frame_equal
 
+from dask.delayed import delayed
+
+import cudf as gd
+
+import dask_cudf as dgd
+
 
 @delayed
 def load_data(nelem, ident):

python/dask_cudf/dask_cudf/tests/test_distributed.py

@@ -1,10 +1,12 @@
-import cudf
+import pytest
+
 import dask
 import dask.dataframe as dd
-import pytest
 from dask.distributed import Client
 from distributed.utils_test import loop  # noqa: F401
 
+import cudf
+
 dask_cuda = pytest.importorskip("dask_cuda")