[GraphBolt] remove single process DGL test (#6729)

tests/python/pytorch/graphbolt/__init__.py

+""" DGL graphbolt API tests"""

tests/python/pytorch/graphbolt/impl/test_fused_csc_sampling_graph.py

@@ -8,12 +8,13 @@ import backend as F
 
 import dgl
 import dgl.graphbolt as gb
-import gb_test_utils as gbt
 import pytest
 import torch
 import torch.multiprocessing as mp
 from scipy import sparse as spsp
 
+from .. import gb_test_utils as gbt
+
 torch.manual_seed(3407)
 mp.set_sharing_strategy("file_system")
 

tests/python/pytorch/graphbolt/impl/test_in_subgraph_sampler.py

 import dgl.graphbolt as gb
-import gb_test_utils
-import pytest
 import torch
 
+from .. import gb_test_utils
+
 
 def test_InSubgraphSampler_homo():
     """Original graph in COO:

tests/python/pytorch/graphbolt/impl/test_negative_sampler.py

 import dgl.graphbolt as gb
-import gb_test_utils
 import pytest
 import torch
 
+from .. import gb_test_utils
+
 
 def test_NegativeSampler_invoke():
     # Instantiate graph and required datapipes.

tests/python/pytorch/graphbolt/impl/test_ondisk_dataset.py

@@ -6,7 +6,6 @@ import tempfile
 import unittest
 import warnings
 
-import gb_test_utils as gbt
 import numpy as np
 import pandas as pd
 import pydantic
@@ -16,6 +15,8 @@ import yaml
 
 from dgl import graphbolt as gb
 
+from .. import gb_test_utils as gbt
+
 
 def write_yaml_file(yaml_content, dir):
     os.makedirs(os.path.join(dir, "preprocessed"), exist_ok=True)

tests/python/pytorch/graphbolt/test_base.py

@@ -4,10 +4,11 @@ import unittest
 import backend as F
 
 import dgl.graphbolt as gb
-import gb_test_utils
 import pytest
 import torch
 
+from . import gb_test_utils
+
 
 @unittest.skipIf(F._default_context_str == "cpu", "CopyTo needs GPU to test")
 def test_CopyTo():

tests/python/pytorch/graphbolt/test_multi_process_dataloader.py → tests/python/pytorch/graphbolt/test_dataloader.py

-import os
-import unittest
-
 import backend as F
 
 import dgl
 import dgl.graphbolt
-import gb_test_utils
 import torch
-from torchdata.datapipes.iter import Mapper
+
+from . import gb_test_utils
 
 
 def test_DataLoader():

tests/python/pytorch/graphbolt/test_feature_fetcher.py

@@ -2,11 +2,12 @@ import random
 from enum import Enum
 
 import dgl.graphbolt as gb
-import gb_test_utils
 import pytest
 import torch
 from torchdata.datapipes.iter import Mapper
 
+from . import gb_test_utils
+
 
 class MiniBatchType(Enum):
     MiniBatch = 1

tests/python/pytorch/graphbolt/test_minibatch_transformer.py

 import dgl.graphbolt as gb
-import gb_test_utils
 import torch
 
+from . import gb_test_utils
+
 
 def test_dgl_minibatch_converter():
     N = 32

tests/python/pytorch/graphbolt/test_single_process_dataloader.py

-import backend as F
-
-import dgl
-import dgl.graphbolt
-import gb_test_utils
-import torch
-from torchdata.datapipes.iter import Mapper
-
-
-def test_DataLoader():
-    N = 32
-    B = 4
-    itemset = dgl.graphbolt.ItemSet(torch.arange(N), names="seed_nodes")
-    graph = gb_test_utils.rand_csc_graph(200, 0.15, bidirection_edge=True)
-
-    features = {}
-    keys = [("node", None, "a"), ("node", None, "b")]
-    features[keys[0]] = dgl.graphbolt.TorchBasedFeature(torch.randn(200, 4))
-    features[keys[1]] = dgl.graphbolt.TorchBasedFeature(torch.randn(200, 4))
-    feature_store = dgl.graphbolt.BasicFeatureStore(features)
-
-    item_sampler = dgl.graphbolt.ItemSampler(itemset, batch_size=B)
-    subgraph_sampler = dgl.graphbolt.NeighborSampler(
-        item_sampler,
-        graph,
-        fanouts=[torch.LongTensor([2]) for _ in range(2)],
-    )
-    feature_fetcher = dgl.graphbolt.FeatureFetcher(
-        subgraph_sampler,
-        feature_store,
-        ["a"],
-    )
-    device_transferrer = dgl.graphbolt.CopyTo(feature_fetcher, F.ctx())
-
-    dataloader = dgl.graphbolt.DataLoader(device_transferrer)
-    assert len(list(dataloader)) == N // B

tests/python/pytorch/graphbolt/test_subgraph_sampler.py

 import dgl
 import dgl.graphbolt as gb
-import gb_test_utils
 import pytest
 import torch
 from torchdata.datapipes.iter import Mapper
 
+from . import gb_test_utils
+
 
 def test_SubgraphSampler_invoke():
     itemset = gb.ItemSet(torch.arange(10), names="seed_nodes")