Add unit test for graph extending data.

Committed-by: SighingSnow from Dev container . Committed-by: SighingSnow from Dev container Committed-by: SighingSnow from Dev container Committed-by: SighingSnow from Dev container
alibaba · Nov 6, 2023 · 193d2f1 · 193d2f1
1 parent a16359d
commit 193d2f1
Show file tree

Hide file tree

Showing 2 changed files with 52 additions and 3 deletions.
diff --git a/python/graphscope/framework/graph.py b/python/graphscope/framework/graph.py
@@ -307,7 +307,7 @@ def __init__(
         self._compact_edges = compact_edges
         self._use_perfect_hash = use_perfect_hash
         # for need to extend label in 'eager mode' when add_vertices and add_edges
-        # 0 - not extending label 
+        # 0 - not extending label
         # 1 - extend vertex label
         # 2 - extend edge label
         self._extend_label_data = 0
@@ -517,8 +517,10 @@ def add_vertices(
         # currently not support local_vertex_map
         if label in self._v_labels:
             self._extend_label_data = 1
-            warnings.warn(f"Label {label} already existed in graph"
-                            ", origin label data will be extend.")
+            warnings.warn(
+                f"Label {label} already existed in graph"
+                ", origin label data will be extend."
+            )
         unsealed_vertices_and_edges = deepcopy(self._unsealed_vertices_and_edges)
         vertex_label = VertexLabel(
             label=label,

diff --git a/python/graphscope/tests/unittest/test_graph.py b/python/graphscope/tests/unittest/test_graph.py
@@ -20,6 +20,7 @@
 import os
 
 import numpy as np
+import pandas as pd
 import pytest
 
 import graphscope
@@ -298,6 +299,52 @@ def test_add_vertices_edges(graphscope_session):
     assert graph.schema.edge_labels == ["knows", "created"]
 
 
+@pytest.mark.skip("unlock this when update")
+def test_extend_vertices_edges(graphscope_session):
+    prefix = os.path.expandvars("${GS_TEST_DIR}/")
+    verts = pd.read_csv(f"{prefix}/p2p_v.csv")
+    edges = pd.read_csv(f"{prefix}/p2p_e.csv")
+    test_list = ["v11308", "v50089", "v60129"]
+
+    g1 = graphscope_session.g(oid_type="std::string")
+    g1 = g1.add_vertices(Loader(verts), "person")
+    g1 = g1.add_edges(Loader(edges), "knows", src_label="person", dst_label="person")
+
+    g2 = graphscope_session.g(oid_type="std::string")
+    g2 = g2.add_vertices(Loader(verts[:12980]), "person")
+    g2 = g2.add_vertices(Loader(verts[12980:31530]), "person")
+    g2 = g2.add_vertices(Loader(verts[31530:]), "person")
+
+    g2 = g2.add_edges(
+        Loader(edges[:2302]), "knows", src_label="person", dst_label="person"
+    )
+    g2 = g2.add_edges(
+        Loader(edges[2302:40021]), "knows", src_label="person", dst_label="person"
+    )
+    g2 = g2.add_edges(
+        Loader(edges[40021:]), "knows", src_label="person", dst_label="person"
+    )
+
+    sg1 = g1.project(vertices={"person": ["id"]}, edges={"knows": ["dist"]})
+    sg2 = g2.project(vertices={"person": ["id"]}, edges={"knows": ["dist"]})
+    for src in test_list:
+        res1 = graphscope.sssp(sg1, src=src, weight="dist")
+        res2 = graphscope.sssp(sg2, src=src, weight="dist")
+        df1 = res1.to_dataframe(selector={"id": "v.id", "r": "r"}).sort_values(
+            by=["id"], ignore_index=True
+        )
+        df2 = res2.to_dataframe(selector={"id": "v.id", "r": "r"}).sort_values(
+            by=["id"], ignore_index=True
+        )
+        if not df1.equals(df2):
+            pytest.raises(
+                AssertionError, "different sssp result got after extending graph data"
+            )
+
+    del g1, g2, sg1, sg2
+    print("pass graph extending test")
+
+
 def test_complicated_add_edges(graphscope_session):
     prefix = os.path.expandvars("${GS_TEST_DIR}/modern_graph")
     graph = graphscope_session.g()