fix for LargeGraphIndexer from_triples ordering issue

CHANGELOG.md

@@ -40,6 +40,7 @@ The format is based on [Keep a Changelog](http://keepachangelog.com/en/1.0.0/).
 
 ### Fixed
 
+- fixed unique ordering in `LargeGraphIndexer.from_triplets` ([#9952](https://github.com/pyg-team/pytorch_geometric/pull/9952))
 - Fixed the `k_hop_subgraph()` method for directed graphs ([#9756](https://github.com/pyg-team/pytorch_geometric/pull/9756))
 - Fixed `utils.group_cat` concatenating dimension ([#9766](https://github.com/pyg-team/pytorch_geometric/pull/9766))
 - Fixed `WebQSDataset.process` raising exceptions ([#9665](https://github.com/pyg-team/pytorch_geometric/pull/9665))

test/data/test_large_graph_indexer.py

@@ -20,7 +20,10 @@
 
 # create possible nodes and edges for graph
 strkeys = string.ascii_letters + string.digits
+
 NODE_POOL = list({"".join(random.sample(strkeys, 10)) for i in range(1000)})
+# insert a duplicate node to test the unique ordering of from_triples
+NODE_POOL[1] = NODE_POOL[0]
 EDGE_POOL = list({"".join(random.sample(strkeys, 10)) for i in range(50)})
 
 
@@ -64,9 +67,15 @@ def test_basic_collate():
     assert len(set(big_indexer._nodes.values())) == len(big_indexer._nodes)
     assert len(set(big_indexer._edges.values())) == len(big_indexer._edges)
 
-    for node in (indexer_0._nodes.keys() | indexer_1._nodes.keys()):
+    for i, node in enumerate(indexer_0._nodes.keys()
+                             | indexer_1._nodes.keys()):
         assert big_indexer.node_attr[NODE_PID][
             big_indexer._nodes[node]] == node
+        if i == 1:
+            node_0 = list(indexer_0._nodes.keys())[0]
+            assert big_indexer.node_attr[NODE_PID][
+                big_indexer._nodes[node_0]] != big_indexer.node_attr[NODE_PID][
+                    big_indexer._nodes[node]]
 
 
 def test_large_graph_index():

torch_geometric/data/large_graph_indexer.py

@@ -31,6 +31,7 @@
 
 
 def ordered_set(values: Iterable[str]) -> List[str]:
+    # dicts mantain order while keeping unique keys
     return list(dict.fromkeys(values))
 
 
@@ -172,8 +173,9 @@ def from_triplets(
             LargeGraphIndexer: Index of unique nodes and edges.
         """
         # NOTE: Right now assumes that all trips can be loaded into memory
-        nodes = set()
-        edges = set()
+        # initialize as a list and then remove duplicates
+        nodes = []
+        edges = []
 
         if pre_transform is not None:
 
@@ -182,16 +184,18 @@ def apply_transform(
                 for trip in trips:
                     yield pre_transform(trip)
 
-            triplets = apply_transform(triplets)
+            triplets = list(apply_transform(triplets))
 
         for h, r, t in triplets:
 
             for node in (h, t):
-                nodes.add(node)
+                nodes.append(node)
 
             edge_idx = (h, r, t)
-            edges.add(edge_idx)
-
+            edges.append(edge_idx)
+        # removes duplicates while mantaining order
+        nodes = ordered_set(nodes)
+        edges = ordered_set(edges)
         return cls(list(nodes), list(edges))
 
     @classmethod