[GraphBolt][Doc] Minor fixes and clarifications. (#7135)

python/dgl/graphbolt/base.py

@@ -90,9 +90,9 @@ def expand_indptr(indptr, dtype=None, node_ids=None, output_size=None):
         argument avoids a stream synchronization to calculate the output shape.
 
     Returns
-        -------
-        torch.Tensor
-            The converted COO tensor with values from node_ids.
+    -------
+    torch.Tensor
+        The converted COO tensor with values from node_ids.
     """
     assert indptr.dim() == 1, "Indptr should be 1D tensor."
     assert not (
@@ -127,9 +127,9 @@ def index_select(tensor, index):
         The 1-D tensor containing the indices to index.
 
     Returns
-        -------
-        torch.Tensor
-            The indexed input tensor, equivalent to tensor[index].
+    -------
+    torch.Tensor
+        The indexed input tensor, equivalent to tensor[index].
     """
     assert index.dim() == 1, "Index should be 1D tensor."
     return torch.ops.graphbolt.index_select(tensor, index)

python/dgl/graphbolt/dataloader.py

@@ -84,7 +84,9 @@ class DataLoader(torch.utils.data.DataLoader):
     everything after feature fetching in the main process. The datapipe
     is modified in-place as a result.
 
-    Only works on single GPU.
+    When the copy_to operation is placed earlier in the data pipeline, the
+    num_workers argument is required to be 0 as utilizing CUDA in multiple
+    worker processes is not supported.
 
     Parameters
     ----------

python/dgl/graphbolt/impl/torch_based_feature_store.py

@@ -63,6 +63,7 @@ class TorchBasedFeature(Feature):
     tensor([[1, 2]])
 
     3. Pinned CPU feature.
+
     >>> torch_feat = torch.arange(10).reshape(2, -1).pin_memory()
     >>> feature = gb.TorchBasedFeature(torch_feat)
     >>> feature.read().device

python/dgl/graphbolt/internal/sample_utils.py

@@ -32,11 +32,11 @@ def unique_and_compact(
     Returns
     -------
     Tuple[unique_nodes, compacted_node_list]
-    The Unique nodes (per type) of all nodes in the input. And the compacted
-    nodes list, where IDs inside are replaced with compacted node IDs.
-    "Compacted node list" indicates that the node IDs in the input node
-    list are replaced with mapped node IDs, where each type of node is
-    mapped to a contiguous space of IDs ranging from 0 to N.
+        The Unique nodes (per type) of all nodes in the input. And the compacted
+        nodes list, where IDs inside are replaced with compacted node IDs.
+        "Compacted node list" indicates that the node IDs in the input node
+        list are replaced with mapped node IDs, where each type of node is
+        mapped to a contiguous space of IDs ranging from 0 to N.
     """
     is_heterogeneous = isinstance(nodes, dict)
 

python/dgl/graphbolt/itemset.py

@@ -36,6 +36,7 @@ class ItemSet:
     >>> from dgl import graphbolt as gb
 
     1. Integer: number of nodes.
+
     >>> num = 10
     >>> item_set = gb.ItemSet(num, names="seed_nodes")
     >>> list(item_set)
@@ -47,6 +48,7 @@ class ItemSet:
     ('seed_nodes',)
 
     2. Single iterable: seed nodes.
+
     >>> node_ids = torch.arange(0, 5)
     >>> item_set = gb.ItemSet(node_ids, names="seed_nodes")
     >>> list(item_set)
@@ -57,6 +59,7 @@ class ItemSet:
     ('seed_nodes',)
 
     3. Tuple of iterables with same shape: seed nodes and labels.
+
     >>> node_ids = torch.arange(0, 5)
     >>> labels = torch.arange(5, 10)
     >>> item_set = gb.ItemSet(
@@ -70,6 +73,7 @@ class ItemSet:
     ('seed_nodes', 'labels')
 
     4. Tuple of iterables with different shape: node pairs and negative dsts.
+
     >>> node_pairs = torch.arange(0, 10).reshape(-1, 2)
     >>> neg_dsts = torch.arange(10, 25).reshape(-1, 3)
     >>> item_set = gb.ItemSet(