Apply flake8 to Cython code.

.pre-commit-config.yaml

@@ -24,7 +24,7 @@ repos:
  args: ["--config=setup.cfg"]
  files: python/.*$
  types: [file]
- types_or: [python] # TODO: Enable [python, cython]
+ types_or: [python, cython]
  additional_dependencies: ["flake8-force"]
  - repo: https://github.com/asottile/yesqa
  rev: v1.3.0

python/cugraph/cugraph/centrality/betweenness_centrality_wrapper.pyx

@@ -16,7 +16,8 @@
 # cython: embedsignature = True
 # cython: language_level = 3
 
-from cugraph.centrality.betweenness_centrality cimport betweenness_centrality as c_betweenness_centrality
+from cugraph.centrality.betweenness_centrality cimport \
+ betweenness_centrality as c_betweenness_centrality
 from cugraph.structure.graph_primtypes cimport *
 from libc.stdint cimport uintptr_t
 from libcpp cimport bool
@@ -37,7 +38,7 @@ def get_output_df(number_of_vertices, result_dtype):
 
 def get_batch(sources, number_of_workers, current_worker):
  batch_size = len(sources) // number_of_workers
- begin =  current_worker * batch_size
+ begin = current_worker * batch_size
  end = (current_worker + 1) * batch_size
  if current_worker == (number_of_workers - 1):
  end = len(sources)
@@ -55,23 +56,25 @@ cdef void run_c_betweenness_centrality(uintptr_t c_handle,
  uintptr_t c_batch,
  result_dtype):
  if result_dtype == np.float64:
- c_betweenness_centrality[int, int, double, double]((<handle_t *> c_handle)[0],
- (<GraphCSRView[int, int, double] *> c_graph)[0],
- <double *> c_betweenness,
- normalized,
- endpoints,
- <double *> c_weights,
- number_of_sources_in_batch,
- <int *> c_batch)
+ c_betweenness_centrality[int, int, double, double](
+ (<handle_t *> c_handle)[0],
+ (<GraphCSRView[int, int, double] *> c_graph)[0],
+ <double *> c_betweenness,
+ normalized,
+ endpoints,
+ <double *> c_weights,
+ number_of_sources_in_batch,
+ <int *> c_batch)
  elif result_dtype == np.float32:
- c_betweenness_centrality[int, int, float, float]((<handle_t *> c_handle)[0],
- (<GraphCSRView[int, int, float] *> c_graph)[0],
- <float *> c_betweenness,
- normalized,
- endpoints,
- <float *> c_weights,
- number_of_sources_in_batch,
- <int *> c_batch)
+ c_betweenness_centrality[int, int, float, float](
+ (<handle_t *> c_handle)[0],
+ (<GraphCSRView[int, int, float] *> c_graph)[0],
+ <float *> c_betweenness,
+ normalized,
+ endpoints,
+ <float *> c_weights,
+ number_of_sources_in_batch,
+ <int *> c_batch)
  else:
  raise ValueError("result_dtype can only be np.float64 or np.float32")
 
@@ -127,7 +130,8 @@ def run_internal_work(handle, input_data, normalized, endpoints,
  raise ValueError("result_dtype can only be np.float64 or np.float32")
 
  c_identifier = result_df['vertex'].__cuda_array_interface__['data'][0]
- c_betweenness = result_df['betweenness_centrality'].__cuda_array_interface__['data'][0]
+ c_betweenness = \
+ result_df['betweenness_centrality'].__cuda_array_interface__['data'][0]
  if weights is not None:
  c_weights = weights.__cuda_array_interface__['data'][0]
 
@@ -152,6 +156,7 @@ def run_internal_work(handle, input_data, normalized, endpoints,
 
  return result_df
 
+
 def run_mg_work(input_data, normalized, endpoints,
  weights, sources,
  result_dtype, session_id):
@@ -170,20 +175,20 @@ def run_mg_work(input_data, normalized, endpoints,
 
 
 def batch_betweenness_centrality(input_graph, normalized, endpoints,
-  weights, vertices, result_dtype):
+ weights, vertices, result_dtype):
  df = None
  client = get_client()
  comms = Comms.get_comms()
  replicated_adjlists = input_graph.batch_adjlists
- work_futures =  [client.submit(run_mg_work,
-  (data, input_graph.is_directed()),
-  normalized,
-  endpoints,
-  weights,
-  vertices,
-  result_dtype,
-  comms.sessionId,
-  workers=[worker]) for
+ work_futures = [client.submit(run_mg_work,
+ (data, input_graph.is_directed()),
+ normalized,
+ endpoints,
+ weights,
+ vertices,
+ result_dtype,
+ comms.sessionId,
+ workers=[worker]) for
  idx, (worker, data) in enumerate(replicated_adjlists.items())]
  dask.distributed.wait(work_futures)
  df = work_futures[0].result()
@@ -215,13 +220,13 @@ def betweenness_centrality(input_graph, normalized, endpoints, weights,
  if not input_graph.adjlist:
  input_graph.view_adj_list()
 
- if Comms.is_initialized() and input_graph.batch_enabled == True:
+ if Comms.is_initialized() and input_graph.batch_enabled is True:
  df = batch_betweenness_centrality(input_graph,
-  normalized,
-  endpoints,
-  weights,
-  vertices,
-  result_dtype)
+ normalized,
+ endpoints,
+ weights,
+ vertices,
+ result_dtype)
  else:
  df = sg_betweenness_centrality(input_graph,
  normalized,

python/cugraph/cugraph/centrality/edge_betweenness_centrality_wrapper.pyx

@@ -16,7 +16,8 @@
 # cython: embedsignature = True
 # cython: language_level = 3
 
-from cugraph.centrality.betweenness_centrality cimport edge_betweenness_centrality as c_edge_betweenness_centrality
+from cugraph.centrality.betweenness_centrality cimport \
+ edge_betweenness_centrality as c_edge_betweenness_centrality
 from cugraph.structure import graph_primtypes_wrapper
 from cugraph.structure.graph_primtypes cimport *
 from libc.stdint cimport uintptr_t
@@ -40,7 +41,7 @@ def get_output_df(indices, result_dtype):
 
 def get_batch(sources, number_of_workers, current_worker):
  batch_size = len(sources) // number_of_workers
- begin =  current_worker * batch_size
+ begin = current_worker * batch_size
  end = (current_worker + 1) * batch_size
  if current_worker == (number_of_workers - 1):
  end = len(sources)
@@ -49,7 +50,7 @@ def get_batch(sources, number_of_workers, current_worker):
 
 
 def run_mg_work(input_data, normalized, weights, sources,
- result_dtype, session_id):
+  result_dtype, session_id):
  result = None
 
  number_of_workers = Comms.get_n_workers(session_id)
@@ -80,7 +81,7 @@ def run_internal_work(handle, input_data, normalized, weights, batch,
  cdef GraphCSRViewDouble graph_double
  cdef GraphCSRViewFloat graph_float
 
- (offsets, indices, graph_weights), is_directed =  input_data
+ (offsets, indices, graph_weights), is_directed = input_data
 
  if graph_weights is not None:
  c_graph_weights = graph_weights.__cuda_array_interface__['data'][0]
@@ -93,7 +94,8 @@ def run_internal_work(handle, input_data, normalized, weights, batch,
  result_df = get_output_df(indices, result_dtype)
  c_src_identifier = result_df['src'].__cuda_array_interface__['data'][0]
  c_dst_identifier = result_df['dst'].__cuda_array_interface__['data'][0]
- c_betweenness = result_df['betweenness_centrality'].__cuda_array_interface__['data'][0]
+ c_betweenness = \
+ result_df['betweenness_centrality'].__cuda_array_interface__['data'][0]
 
  number_of_sources_in_batch = len(batch)
  if result_dtype == np.float64:
@@ -140,38 +142,41 @@ cdef void run_c_edge_betweenness_centrality(uintptr_t c_handle,
  uintptr_t c_batch,
  result_dtype):
  if result_dtype == np.float64:
- c_edge_betweenness_centrality[int, int, double, double]((<handle_t *> c_handle)[0],
- (<GraphCSRView[int, int, double] *> c_graph)[0],
- <double *> c_betweenness,
- normalized,
- <double *> c_weights,
- number_of_sources_in_batch,
- <int *> c_batch)
+ c_edge_betweenness_centrality[int, int, double, double](
+ (<handle_t *> c_handle)[0],
+ (<GraphCSRView[int, int, double] *> c_graph)[0],
+ <double *> c_betweenness,
+ normalized,
+ <double *> c_weights,
+ number_of_sources_in_batch,
+ <int *> c_batch)
  elif result_dtype == np.float32:
- c_edge_betweenness_centrality[int, int, float, float]((<handle_t *> c_handle)[0],
- (<GraphCSRView[int, int, float] *> c_graph)[0],
- <float *> c_betweenness,
- normalized,
- <float *> c_weights,
- number_of_sources_in_batch,
- <int *> c_batch)
+ c_edge_betweenness_centrality[int, int, float, float](
+ (<handle_t *> c_handle)[0],
+ (<GraphCSRView[int, int, float] *> c_graph)[0],
+ <float *> c_betweenness,
+ normalized,
+ <float *> c_weights,
+ number_of_sources_in_batch,
+ <int *> c_batch)
  else:
  raise ValueError("result_dtype can only be np.float64 or np.float32")
 
+
 def batch_edge_betweenness_centrality(input_graph,
-  normalized,
-  weights, vertices, result_dtype):
+ normalized,
+ weights, vertices, result_dtype):
  client = get_client()
  comms = Comms.get_comms()
  replicated_adjlists = input_graph.batch_adjlists
- work_futures =  [client.submit(run_mg_work,
-  (data, input_graph.is_directed()),
-  normalized,
-  weights,
-  vertices,
-  result_dtype,
-  comms.sessionId,
-  workers=[worker]) for
+ work_futures = [client.submit(run_mg_work,
+ (data, input_graph.is_directed()),
+ normalized,
+ weights,
+ vertices,
+ result_dtype,
+ comms.sessionId,
+ workers=[worker]) for
  (worker, data) in replicated_adjlists.items()]
  dask.distributed.wait(work_futures)
  df = work_futures[0].result()
@@ -200,25 +205,26 @@ def edge_betweenness_centrality(input_graph, normalized, weights,
  cdef GraphCSRViewDouble graph_double
  cdef GraphCSRViewFloat graph_float
 
-
  df = None
 
  if not input_graph.adjlist:
  input_graph.view_adj_list()
 
- if Comms.is_initialized() and input_graph.batch_enabled == True:
+ if Comms.is_initialized() and input_graph.batch_enabled is True:
  df = batch_edge_betweenness_centrality(input_graph, normalized,
-  weights, vertices,
-  result_dtype)
+ weights, vertices,
+ result_dtype)
  else:
  df = sg_edge_betweenness_centrality(input_graph, normalized,
  weights, vertices, result_dtype)
 
  if result_dtype == np.float64:
  graph_double = get_graph_view[GraphCSRViewDouble](input_graph)
- graph_double.get_source_indices(<int*>(<uintptr_t>df['src'].__cuda_array_interface__['data'][0]))
+ graph_double.get_source_indices(
+ <int*>(<uintptr_t>df['src'].__cuda_array_interface__['data'][0]))
  elif result_dtype == np.float32:
  graph_float = get_graph_view[GraphCSRViewFloat](input_graph)
- graph_float.get_source_indices(<int*>(<uintptr_t>df['src'].__cuda_array_interface__['data'][0]))
+ graph_float.get_source_indices(
+ <int*>(<uintptr_t>df['src'].__cuda_array_interface__['data'][0]))
 
  return df

python/cugraph/cugraph/community/ecg.pxd

@@ -1,4 +1,4 @@
-# Copyright (c) 2019-2021, NVIDIA CORPORATION.
+# Copyright (c) 2019-2022, NVIDIA CORPORATION.
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.
 # You may obtain a copy of the License at
@@ -21,9 +21,9 @@ from cugraph.structure.graph_primtypes cimport *
 
 cdef extern from "cugraph/algorithms.hpp" namespace "cugraph":
 
- cdef void ecg[VT,ET,WT](
+ cdef void ecg[VT, ET, WT](
  const handle_t &handle,
- const GraphCSRView[VT,ET,WT] &graph,
+ const GraphCSRView[VT, ET, WT] &graph,
  WT min_weight,
  VT ensemble_size,
  VT* ecg_parts) except +

python/cugraph/cugraph/community/ecg_wrapper.pyx

@@ -38,9 +38,13 @@ def ecg(input_graph, min_weight=.05, ensemble_size=16):
  cdef unique_ptr[handle_t] handle_ptr
  handle_ptr.reset(new handle_t())
 
- [offsets, indices] = graph_primtypes_wrapper.datatype_cast([input_graph.adjlist.offsets,
- input_graph.adjlist.indices], [np.int32, np.int64])
- [weights] = graph_primtypes_wrapper.datatype_cast([input_graph.adjlist.weights], [np.float32, np.float64])
+ [offsets, indices] = graph_primtypes_wrapper.datatype_cast(
+ [
+ input_graph.adjlist.offsets,
+ input_graph.adjlist.indices,
+ ], [np.int32, np.int64])
+ [weights] = graph_primtypes_wrapper.datatype_cast(
+ [input_graph.adjlist.weights], [np.float32, np.float64])
 
  num_verts = input_graph.number_of_vertices()
  num_edges = input_graph.number_of_edges(directed_edges=True)
@@ -55,30 +59,38 @@ def ecg(input_graph, min_weight=.05, ensemble_size=16):
  cdef uintptr_t c_partition = df['partition'].__cuda_array_interface__['data'][0]
  cdef uintptr_t c_weights = weights.__cuda_array_interface__['data'][0]
 
- cdef GraphCSRView[int,int,float] graph_float
- cdef GraphCSRView[int,int,double] graph_double
+ cdef GraphCSRView[int, int, float] graph_float
+ cdef GraphCSRView[int, int, double] graph_double
 
  if weights.dtype == np.float32:
- graph_float = GraphCSRView[int,int,float](<int*>c_offsets, <int*>c_indices,
- <float*>c_weights, num_verts, num_edges)
+ graph_float = GraphCSRView[int, int, float](
+ <int*> c_offsets,
+ <int*> c_indices,
+ <float*> c_weights,
+ num_verts,
+ num_edges)
 
  graph_float.get_vertex_identifiers(<int*>c_identifier)
 
- c_ecg[int,int,float](handle_ptr.get()[0],
- graph_float,
- min_weight,
- ensemble_size,
- <int*> c_partition)
+ c_ecg[int, int, float](handle_ptr.get()[0],
+  graph_float,
+  min_weight,
+  ensemble_size,
+  <int*> c_partition)
  else:
- graph_double = GraphCSRView[int,int,double](<int*>c_offsets, <int*>c_indices,
- <double*>c_weights, num_verts, num_edges)
+ graph_double = GraphCSRView[int, int, double](
+ <int*> c_offsets,
+ <int*> c_indices,
+ <double*> c_weights,
+ num_verts,
+ num_edges)
 
  graph_double.get_vertex_identifiers(<int*>c_identifier)
 
- c_ecg[int,int,double](handle_ptr.get()[0],
- graph_double,
- min_weight,
- ensemble_size,
- <int*> c_partition)
+ c_ecg[int, int, double](handle_ptr.get()[0],
+  graph_double,
+  min_weight,
+  ensemble_size,
+  <int*> c_partition)
 
  return df

python/cugraph/cugraph/community/ktruss_subgraph.pxd

@@ -1,4 +1,4 @@
-# Copyright (c) 2019-2021, NVIDIA CORPORATION.
+# Copyright (c) 2019-2022, NVIDIA CORPORATION.
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.
 # You may obtain a copy of the License at
@@ -21,6 +21,6 @@ from cugraph.structure.graph_primtypes cimport *
 
 cdef extern from "cugraph/algorithms.hpp" namespace "cugraph":
 
- cdef unique_ptr[GraphCOO[VT,ET,WT]] k_truss_subgraph[VT,ET,WT](
- const GraphCOOView[VT,ET,WT] &graph,
+ cdef unique_ptr[GraphCOO[VT, ET, WT]] k_truss_subgraph[VT, ET, WT](
+ const GraphCOOView[VT, ET, WT] &graph,
  int k) except +