Warn when doing sampling in from_gds

Co-Authored-By: Florentin Dörre <[email protected]>

Author: adamnsch

python-wrapper/src/neo4j_viz/gds.py

@@ -1,5 +1,6 @@
 from __future__ import annotations
 
+import warnings
 from itertools import chain
 from typing import Optional
 from uuid import uuid4
@@ -99,6 +100,9 @@ def from_gds(
 
     node_count = G.node_count()
     if node_count > max_node_count:
+        warnings.warn(
+            f"The '{G.name()}' projection's node count ({G.node_count()}) exceeds `max_node_count` ({max_node_count}), so subsampling will be applied. Increase `max_node_count` if needed"
+        )
         sampling_ratio = float(max_node_count) / node_count
         sample_name = f"neo4j-viz_sample_{uuid4()}"
         G_fetched, _ = gds.graph.sample.rwr(sample_name, G, samplingRatio=sampling_ratio, nodeLabelStratification=True)

python-wrapper/tests/test_gds.py

@@ -1,3 +1,4 @@
+import re
 from typing import Any
 
 import pandas as pd
@@ -267,7 +268,13 @@ def test_from_gds_sample(gds: Any) -> None:
     from neo4j_viz.gds import from_gds
 
     with gds.graph.generate("hello", node_count=11_000, average_degree=1) as G:
-        VG = from_gds(gds, G)
+        with pytest.warns(
+            UserWarning,
+            match=re.escape(
+                "The 'hello' projection's node count (11000) exceeds `max_node_count` (10000), so subsampling will be applied. Increase `max_node_count` if needed"
+            ),
+        ):
+            VG = from_gds(gds, G)
 
         assert len(VG.nodes) >= 9_500
         assert len(VG.nodes) <= 10_500