Skip to content

Refactor Python test to improve performance #5627

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
BradReesWork wants to merge 17 commits into
base: main
Choose a base branch
from
Open

Refactor Python test to improve performance #5627

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
17 commits
Select commit Hold shift + click to select a range
65d4be7
BC test
BradReesWork Aug 13, 2026
e7b7894
BC refactor
BradReesWork Aug 14, 2026
f21a87f
dropped karate - reduced by 64 tests
BradReesWork Aug 14, 2026
5e67a11
reduced number of datasets, dropped similar ones
BradReesWork Aug 17, 2026
644f3ff
address transpose warning message
BradReesWork Aug 17, 2026
bc87383
updated to use Dataset API and just two test files
BradReesWork Aug 18, 2026
7190365
style
BradReesWork Aug 18, 2026
f2c5d73
Merge branch 'main' into Refactor_Test_Perf
BradReesWork Aug 18, 2026
c4f43f5
Merge branch 'main' into Refactor_Test_Perf
BradReesWork Aug 19, 2026
135e569
Merge branch 'main' into Refactor_Test_Perf
BradReesWork Aug 24, 2026
9dda4f3
Merge branch 'main' into Refactor_Test_Perf
BradReesWork Aug 26, 2026
ec4c781
Merge branch 'main' into Refactor_Test_Perf
BradReesWork Aug 27, 2026
657d2b1
updated result set call
BradReesWork Aug 27, 2026
cfe3bdd
style fix
BradReesWork Aug 28, 2026
a52e33b
updted copyright text
BradReesWork Aug 28, 2026
d6c68d3
extra spacing
BradReesWork Aug 28, 2026
47ac62a
one more fix
BradReesWork Aug 28, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
78 changes: 51 additions & 27 deletions python/cugraph/cugraph/tests/centrality/test_betweenness_centrality.py
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,4 +1,4 @@
# SPDX-FileCopyrightText: Copyright (c) 2020-2025, NVIDIA CORPORATION.:
# SPDX-FileCopyrightText: Copyright (c) 2020-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved.
# SPDX-License-Identifier: Apache-2.0

import gc
Expand All @@ -10,12 +10,14 @@

import cudf
import cugraph
from cugraph.datasets import karate_disjoint
from cugraph.testing import SMALL_DATASETS
from cugraph.datasets import karate_disjoint, netscience, polbooks, karate

# =============================================================================
# Parameters
# =============================================================================
TEST_DATASETS = [netscience, polbooks]


DIRECTED_GRAPH_OPTIONS = [False, True]
WEIGHTED_GRAPH_OPTIONS = [False, True]
ENDPOINTS_OPTIONS = [False, True]
Expand All @@ -25,9 +27,7 @@
SUBSET_SIZE_OPTIONS = [4, None]
SUBSET_SEED_OPTIONS = [42]

# NOTE: The following is not really being exploited in the tests as the
# datasets that are used are too small to compare, but it ensures that both
# path are actually sane
# NOTE: The output conversion is done after the centrality is computed.
RESULT_DTYPE_OPTIONS = [np.float32, np.float64]


Expand Down Expand Up @@ -301,16 +301,49 @@ def compare_scores(sorted_df, first_key, second_key, epsilon=DEFAULT_EPSILON):
# =============================================================================
@pytest.mark.sg
@pytest.mark.requires_nx(min_ver="3.5")
@pytest.mark.parametrize("graph_file", SMALL_DATASETS)
@pytest.mark.parametrize("graph_file", TEST_DATASETS)
@pytest.mark.parametrize("directed", [False, True])
@pytest.mark.parametrize("subset_size", SUBSET_SIZE_OPTIONS)
@pytest.mark.parametrize("normalized", NORMALIZED_OPTIONS)
@pytest.mark.parametrize("weight", [None])
@pytest.mark.parametrize("endpoints", ENDPOINTS_OPTIONS)
@pytest.mark.parametrize("subset_seed", SUBSET_SEED_OPTIONS)
@pytest.mark.parametrize("result_dtype", RESULT_DTYPE_OPTIONS)
@pytest.mark.parametrize("edgevals", WEIGHTED_GRAPH_OPTIONS)
def test_betweenness_centrality(
graph_file,
directed,
subset_size,
normalized,
weight,
endpoints,
subset_seed,
edgevals,
):
sorted_df = calc_betweenness_centrality(
graph_file,
directed=directed,
normalized=normalized,
k=subset_size,
weight=weight,
endpoints=endpoints,
seed=subset_seed,
edgevals=edgevals,
)
compare_scores(sorted_df, first_key="cu_bc", second_key="ref_bc")


@pytest.mark.sg
@pytest.mark.requires_nx(min_ver="3.5")
@pytest.mark.parametrize("graph_file", [karate])
@pytest.mark.parametrize("directed", [False])
@pytest.mark.parametrize("subset_size", [4])
@pytest.mark.parametrize("normalized", [False])
@pytest.mark.parametrize("weight", [None])
@pytest.mark.parametrize("endpoints", [False])
@pytest.mark.parametrize("subset_seed", SUBSET_SEED_OPTIONS)
@pytest.mark.parametrize("result_dtype", RESULT_DTYPE_OPTIONS)
@pytest.mark.parametrize("edgevals", [False])
def test_betweenness_centrality_return_dtypes(
graph_file,
directed,
subset_size,
Expand All @@ -336,14 +369,13 @@ def test_betweenness_centrality(


@pytest.mark.sg
@pytest.mark.parametrize("graph_file", SMALL_DATASETS)
@pytest.mark.parametrize("graph_file", [netscience])
@pytest.mark.parametrize("directed", DIRECTED_GRAPH_OPTIONS)
@pytest.mark.parametrize("subset_size", [None])
@pytest.mark.parametrize("normalized", NORMALIZED_OPTIONS)
@pytest.mark.parametrize("normalized", [False])
@pytest.mark.parametrize("weight", [None])
@pytest.mark.parametrize("endpoints", ENDPOINTS_OPTIONS)
@pytest.mark.parametrize("endpoints", [False])
@pytest.mark.parametrize("subset_seed", SUBSET_SEED_OPTIONS)
@pytest.mark.parametrize("result_dtype", RESULT_DTYPE_OPTIONS)
@pytest.mark.parametrize("use_k_full", [True])
@pytest.mark.parametrize("edgevals", WEIGHTED_GRAPH_OPTIONS)
@pytest.mark.skip(reason="Skipping large tests")
Expand All @@ -355,7 +387,6 @@ def test_betweenness_centrality_k_full(
weight,
endpoints,
subset_seed,
result_dtype,
use_k_full,
edgevals,
):
Expand All @@ -369,7 +400,6 @@ def test_betweenness_centrality_k_full(
weight=weight,
endpoints=endpoints,
seed=subset_seed,
result_dtype=result_dtype,
use_k_full=use_k_full,
edgevals=edgevals,
)
Expand All @@ -388,7 +418,6 @@ def test_betweenness_centrality_k_full(
@pytest.mark.parametrize("weight", [None])
@pytest.mark.parametrize("endpoints", ENDPOINTS_OPTIONS)
@pytest.mark.parametrize("subset_seed", [None])
@pytest.mark.parametrize("result_dtype", RESULT_DTYPE_OPTIONS)
@pytest.mark.parametrize("edgevals", WEIGHTED_GRAPH_OPTIONS)
@pytest.mark.skip(reason="Skipping large tests")
def test_betweenness_centrality_fixed_sample(
Expand All @@ -399,7 +428,6 @@ def test_betweenness_centrality_fixed_sample(
weight,
endpoints,
subset_seed,
result_dtype,
edgevals,
):
"""Test Betweenness Centrality using a subset
Expand All @@ -413,21 +441,19 @@ def test_betweenness_centrality_fixed_sample(
weight=weight,
endpoints=endpoints,
seed=subset_seed,
result_dtype=result_dtype,
edgevals=edgevals,
)
compare_scores(sorted_df, first_key="cu_bc", second_key="ref_bc")


@pytest.mark.sg
@pytest.mark.parametrize("graph_file", SMALL_DATASETS)
@pytest.mark.parametrize("graph_file", TEST_DATASETS)
@pytest.mark.parametrize("directed", DIRECTED_GRAPH_OPTIONS)
@pytest.mark.parametrize("subset_size", SUBSET_SIZE_OPTIONS)
@pytest.mark.parametrize("normalized", NORMALIZED_OPTIONS)
@pytest.mark.parametrize("weight", [[]])
@pytest.mark.parametrize("endpoints", ENDPOINTS_OPTIONS)
@pytest.mark.parametrize("subset_seed", SUBSET_SEED_OPTIONS)
@pytest.mark.parametrize("result_dtype", RESULT_DTYPE_OPTIONS)
@pytest.mark.parametrize("edgevals", WEIGHTED_GRAPH_OPTIONS)
@pytest.mark.skip(reason="Skipping large tests")
def test_betweenness_centrality_weight_except(
Expand All @@ -438,7 +464,6 @@ def test_betweenness_centrality_weight_except(
weight,
endpoints,
subset_seed,
result_dtype,
edgevals,
):
"""Calls betwenness_centrality with weight
Expand All @@ -454,22 +479,21 @@ def test_betweenness_centrality_weight_except(
weight=weight,
endpoints=endpoints,
seed=subset_seed,
result_dtype=result_dtype,
edgevals=edgevals,
)
compare_scores(sorted_df, first_key="cu_bc", second_key="ref_bc")


@pytest.mark.sg
@pytest.mark.parametrize("graph_file", SMALL_DATASETS)
@pytest.mark.parametrize("directed", DIRECTED_GRAPH_OPTIONS)
@pytest.mark.parametrize("normalized", NORMALIZED_OPTIONS)
@pytest.mark.parametrize("subset_size", SUBSET_SIZE_OPTIONS)
@pytest.mark.parametrize("graph_file", [karate])
@pytest.mark.parametrize("directed", [False])
@pytest.mark.parametrize("normalized", [False])
@pytest.mark.parametrize("subset_size", [None])
@pytest.mark.parametrize("weight", [None])
@pytest.mark.parametrize("endpoints", ENDPOINTS_OPTIONS)
@pytest.mark.parametrize("endpoints", [False])
@pytest.mark.parametrize("subset_seed", SUBSET_SEED_OPTIONS)
@pytest.mark.parametrize("result_dtype", [str])
@pytest.mark.parametrize("edgevals", WEIGHTED_GRAPH_OPTIONS)
@pytest.mark.parametrize("edgevals", [False])
def test_betweenness_invalid_dtype(
graph_file,
directed,
Expand Down
62 changes: 40 additions & 22 deletions python/cugraph/cugraph/tests/centrality/test_edge_betweenness_centrality.py
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,4 +1,4 @@
# SPDX-FileCopyrightText: Copyright (c) 2019-2025, NVIDIA CORPORATION.:
# SPDX-FileCopyrightText: Copyright (c) 2019-2026, NVIDIA CORPORATION & AFFILIATES. All rights reserved.
# SPDX-License-Identifier: Apache-2.0

import gc
Expand All @@ -11,24 +11,23 @@
import cupy
import cudf
import cugraph
from cugraph.datasets import karate_disjoint
from cugraph.testing import utils, SMALL_DATASETS
from cugraph.datasets import karate_disjoint, polbooks, karate
from cugraph.testing import utils

# NOTE: Endpoint parameter is not currently being tested, there could be a test
# to verify that python raise an error if it is used
# =============================================================================
# Parameters
# =============================================================================
TEST_DATASETS = [karate, polbooks]

DIRECTED_GRAPH_OPTIONS = [False, True]
WEIGHTED_GRAPH_OPTIONS = [False, True]
NORMALIZED_OPTIONS = [False, True]
DEFAULT_EPSILON = 0.0001

SUBSET_SIZE_OPTIONS = [4, None]

# NOTE: The following is not really being exploited in the tests as the
# datasets that are used are too small to compare, but it ensures that both
# path are actually sane
# NOTE: The output conversion is done after the centrality is computed.
RESULT_DTYPE_OPTIONS = [np.float32, np.float64]


Expand Down Expand Up @@ -302,21 +301,49 @@ def generate_upper_triangle(dataframe):

@pytest.mark.sg
@pytest.mark.requires_nx(min_ver="3.5", max_ver="3.5")
@pytest.mark.parametrize("graph_file", SMALL_DATASETS)
@pytest.mark.parametrize("graph_file", TEST_DATASETS)
@pytest.mark.parametrize("directed", DIRECTED_GRAPH_OPTIONS)
@pytest.mark.parametrize("subset_size", SUBSET_SIZE_OPTIONS)
@pytest.mark.parametrize("normalized", NORMALIZED_OPTIONS)
@pytest.mark.parametrize("weight", [None])
@pytest.mark.parametrize("result_dtype", RESULT_DTYPE_OPTIONS)
@pytest.mark.parametrize("edgevals", WEIGHTED_GRAPH_OPTIONS)
def test_edge_betweenness_centrality(
graph_file,
directed,
subset_size,
normalized,
weight,
result_dtype,
edgevals,
):
sorted_df = calc_edge_betweenness_centrality(
graph_file,
directed=directed,
normalized=normalized,
k=subset_size,
weight=weight,
seed=42,
edgevals=edgevals,
)
compare_scores(sorted_df, first_key="cu_bc", second_key="ref_bc")


@pytest.mark.sg
@pytest.mark.requires_nx(min_ver="3.5", max_ver="3.5")
@pytest.mark.parametrize("graph_file", [karate])
@pytest.mark.parametrize("directed", [False])
@pytest.mark.parametrize("subset_size", [None])
@pytest.mark.parametrize("normalized", [False])
@pytest.mark.parametrize("weight", [None])
@pytest.mark.parametrize("edgevals", [False])
@pytest.mark.parametrize("result_dtype", RESULT_DTYPE_OPTIONS)
def test_edge_betweenness_centrality_return_dtypes(
graph_file,
directed,
subset_size,
normalized,
weight,
edgevals,
result_dtype,
):
sorted_df = calc_edge_betweenness_centrality(
graph_file,
Expand All @@ -333,12 +360,11 @@ def test_edge_betweenness_centrality(

@pytest.mark.sg
@pytest.mark.requires_nx(min_ver="3.5")
@pytest.mark.parametrize("graph_file", SMALL_DATASETS)
@pytest.mark.parametrize("graph_file", TEST_DATASETS)
@pytest.mark.parametrize("directed", DIRECTED_GRAPH_OPTIONS)
@pytest.mark.parametrize("subset_size", [None])
@pytest.mark.parametrize("normalized", NORMALIZED_OPTIONS)
@pytest.mark.parametrize("weight", [None])
@pytest.mark.parametrize("result_dtype", RESULT_DTYPE_OPTIONS)
@pytest.mark.parametrize("use_k_full", [True])
@pytest.mark.parametrize("edgevals", WEIGHTED_GRAPH_OPTIONS)
def test_edge_betweenness_centrality_k_full(
Expand All @@ -347,7 +373,6 @@ def test_edge_betweenness_centrality_k_full(
subset_size,
normalized,
weight,
result_dtype,
use_k_full,
edgevals,
):
Expand All @@ -360,7 +385,6 @@ def test_edge_betweenness_centrality_k_full(
k=subset_size,
weight=weight,
seed=42,
result_dtype=result_dtype,
use_k_full=use_k_full,
edgevals=edgevals,
)
Expand All @@ -378,15 +402,13 @@ def test_edge_betweenness_centrality_k_full(
@pytest.mark.parametrize("subset_size", SUBSET_SIZE_OPTIONS)
@pytest.mark.parametrize("normalized", NORMALIZED_OPTIONS)
@pytest.mark.parametrize("weight", [None])
@pytest.mark.parametrize("result_dtype", RESULT_DTYPE_OPTIONS)
@pytest.mark.parametrize("edgevals", WEIGHTED_GRAPH_OPTIONS)
def test_edge_betweenness_centrality_fixed_sample(
graph_file,
directed,
subset_size,
normalized,
weight,
result_dtype,
edgevals,
):
"""Test Edge Betweenness Centrality using a subset
Expand All @@ -400,27 +422,24 @@ def test_edge_betweenness_centrality_fixed_sample(
normalized=normalized,
weight=weight,
seed=None,
result_dtype=result_dtype,
edgevals=edgevals,
)
compare_scores(sorted_df, first_key="cu_bc", second_key="ref_bc")


@pytest.mark.sg
@pytest.mark.parametrize("graph_file", SMALL_DATASETS)
@pytest.mark.parametrize("graph_file", TEST_DATASETS)
@pytest.mark.parametrize("directed", DIRECTED_GRAPH_OPTIONS)
@pytest.mark.parametrize("subset_size", SUBSET_SIZE_OPTIONS)
@pytest.mark.parametrize("normalized", NORMALIZED_OPTIONS)
@pytest.mark.parametrize("weight", [[]])
@pytest.mark.parametrize("result_dtype", RESULT_DTYPE_OPTIONS)
@pytest.mark.parametrize("edgevals", WEIGHTED_GRAPH_OPTIONS)
def test_edge_betweenness_centrality_weight_except(
graph_file,
directed,
subset_size,
normalized,
weight,
result_dtype,
edgevals,
):
"""Test calls edge_betweeness_centrality with weight parameter
Expand All @@ -436,14 +455,13 @@ def test_edge_betweenness_centrality_weight_except(
normalized=normalized,
weight=weight,
seed=42,
result_dtype=result_dtype,
edgevals=edgevals,
)
compare_scores(sorted_df, first_key="cu_bc", second_key="ref_bc")


@pytest.mark.sg
@pytest.mark.parametrize("graph_file", SMALL_DATASETS)
@pytest.mark.parametrize("graph_file", TEST_DATASETS)
@pytest.mark.parametrize("directed", DIRECTED_GRAPH_OPTIONS)
@pytest.mark.parametrize("normalized", NORMALIZED_OPTIONS)
@pytest.mark.parametrize("subset_size", SUBSET_SIZE_OPTIONS)
Expand Down
Loading
Loading