address coverage gaps

test/data/test_summary.py

@@ -1,29 +1,47 @@
 import torch
-from pytest import approx
+from pytest import approx, raises
 
 from torch_geometric import seed_everything
+from torch_geometric.data import Summary
 from torch_geometric.datasets import FakeDataset, FakeHeteroDataset
 from torch_geometric.testing import withPackage
 
 
-@withPackage('pandas')
-@withPackage('tabulate')
-def test_summary():
-    seed_everything(0)
-    dataset = FakeDataset(num_graphs=10)
-    summary = dataset.summary()
-    assert summary.num_graphs == 10
-    num_nodes = torch.tensor([d.num_nodes for d in dataset]).float()
+def check_summary(summary, num_nodes, num_edges):
     assert summary.mean_num_nodes == approx(float(num_nodes.mean()))
     assert summary.std_num_nodes == approx(float(num_nodes.std()))
     assert summary.min_num_nodes == int(num_nodes.min())
     assert summary.max_num_nodes == int(num_nodes.max())
+    assert summary.median_num_nodes == int(num_nodes.quantile(q=0.5))
 
-    num_edges = torch.tensor([d.num_edges for d in dataset]).float()
     assert summary.mean_num_edges == approx(float(num_edges.mean()))
     assert summary.std_num_edges == approx(float(num_edges.std()))
     assert summary.min_num_edges == int(num_edges.min())
     assert summary.max_num_edges == int(num_edges.max())
+    assert summary.median_num_edges == int(num_edges.quantile(q=0.5))
+
+
+@withPackage('pandas')
+@withPackage('tabulate')
+def test_summary():
+    seed_everything(0)
+    dataset = FakeDataset(num_graphs=10)
+    summary = dataset.summary()
+    assert summary.num_graphs == 10
+    num_nodes = torch.tensor([d.num_nodes for d in dataset]).float()
+    num_edges = torch.tensor([d.num_edges for d in dataset]).float()
+    check_summary(summary, num_nodes, num_edges)
+
+    Summary.progressbar_threshold(0)
+    summary = dataset.summary()
+    assert summary.num_graphs == 10
+    check_summary(summary, num_nodes, num_edges)
+
+    with raises(ValueError, match="threshold must be a positive integer"):
+        Summary.progressbar_threshold(-1)
+
+    with raises(ValueError, match="threshold must be a positive integer"):
+        Summary.progressbar_threshold(10.0)
 
 
 @withPackage('pandas')
@@ -34,13 +52,5 @@ def test_hetero_summary():
     summary = dataset.summary()
     assert summary.num_graphs == 10
     num_nodes = torch.tensor([d.num_nodes for d in dataset]).float()
-    assert summary.mean_num_nodes == approx(float(num_nodes.mean()))
-    assert summary.std_num_nodes == approx(float(num_nodes.std()))
-    assert summary.min_num_nodes == int(num_nodes.min())
-    assert summary.max_num_nodes == int(num_nodes.max())
-
     num_edges = torch.tensor([d.num_edges for d in dataset]).float()
-    assert summary.mean_num_edges == approx(float(num_edges.mean()))
-    assert summary.std_num_edges == approx(float(num_edges.std()))
-    assert summary.min_num_edges == int(num_edges.min())
-    assert summary.max_num_edges == int(num_edges.max())
+    check_summary(summary, num_nodes, num_edges)

torch_geometric/data/summary.py

@@ -43,7 +43,7 @@ def max_num_nodes(self) -> int:
     @property
     def median_num_nodes(self) -> int:
         r"""The median number of nodes"""
-        return int(self._desc['nodes']['median'])
+        return int(self._desc['nodes']['50%'])
 
     @property
     def mean_num_nodes(self) -> float:
@@ -68,7 +68,7 @@ def max_num_edges(self) -> int:
     @property
     def median_num_edges(self) -> int:
         r"""The median number of edges"""
-        return int(self._desc['edges']['median'])
+        return int(self._desc['edges']['50%'])
 
     @property
     def mean_num_edges(self) -> float:
@@ -104,5 +104,7 @@ def progressbar_threshold(value: Optional[int] = None):
             Summary._threshold = Summary._default_threshold
             return
 
-        assert isinstance(value, int) and not value < 0
+        if not isinstance(value, int) or value < 0:
+            raise ValueError("threshold must be a positive integer.")
+
         Summary._threshold = value