Merge pull request #683 from FlorentinD/cypher-proj-empty-graph

Check for empty aggregation result

Author: FlorentinD

changelog.md

@@ -22,5 +22,7 @@
 
 ## Improvements
 
+* Improved the error message if `gds.graph.project.cypher` produces an empty graph.
+
 
 ## Other changes

graphdatascience/graph/graph_cypher_runner.py

@@ -1,6 +1,10 @@
+from __future__ import annotations
+
 import re
 from itertools import chain, zip_longest
-from typing import Any, Optional
+from typing import Any, Dict, Optional
+
+from pandas import Series
 
 from ..caller_base import CallerBase
 from ..query_runner.query_runner import QueryRunner
@@ -41,7 +45,10 @@ def project(
 
         GraphCypherRunner._verify_query_ends_with_return_clause(self._namespace, query)
 
-        result = self._query_runner.run_cypher(query, params, database, False).squeeze()
+        result: Optional[Dict[str, Any]] = self._query_runner.run_cypher(query, params, database, False).squeeze()
+
+        if not result:
+            raise ValueError("Projected graph cannot be empty.")
 
         try:
             graph_name = str(result["graphName"])
@@ -50,7 +57,7 @@ def project(
                 f"Invalid query, the query must end with the `RETURN {self._namespace}(...)` call: {query}"
             )
 
-        return GraphCreateResult(Graph(graph_name, self._query_runner, self._server_version), result)
+        return GraphCreateResult(Graph(graph_name, self._query_runner, self._server_version), Series(data=result))
 
     __separators = re.compile(r"[,(.]")
 

graphdatascience/tests/integration/test_graph_ops.py

@@ -94,6 +94,12 @@ def test_cypher_projection(gds: GraphDataScience) -> None:
     assert result["exists"]
 
 
+@pytest.mark.filterwarnings("ignore: One of the labels in your query is not available in the database")
+def test_cypher_projection_empty_graph(gds: GraphDataScience) -> None:
+    with pytest.raises(ValueError, match="Projected graph cannot be empty"):
+        gds.graph.cypher.project("MATCH (n:MISSING_LABEL) RETURN gds.graph.project('some-graph', n, null)")
+
+
 def test_beta_project_subgraph(runner: QueryRunner, gds: GraphDataScience) -> None:
     from_G, _ = gds.graph.project(GRAPH_NAME, {"Node": {"properties": "x"}}, "*")
 

graphdatascience/tests/unit/test_graph_cypher.py

@@ -9,7 +9,7 @@
 
 @pytest.mark.parametrize("server_version", [ServerVersion(2, 4, 0)])
 def test_simple(runner: CollectingQueryRunner, gds: GraphDataScience) -> None:
-    runner.set__mock_result(DataFrame([{"graphName": "g", "don't squeeze": "me now"}]))
+    runner.set__mock_result(DataFrame([[{"graphName": "g", "don't squeeze": "me now"}]]))
 
     G, _ = gds.graph.cypher.project("MATCH (s)-->(t) RETURN gds.graph.project('g', s, t)")
 
@@ -21,7 +21,7 @@ def test_simple(runner: CollectingQueryRunner, gds: GraphDataScience) -> None:
 
 @pytest.mark.parametrize("server_version", [ServerVersion(2, 4, 0)])
 def test_fstring(runner: CollectingQueryRunner, gds: GraphDataScience) -> None:
-    runner.set__mock_result(DataFrame([{"graphName": "g", "don't squeeze": "me now"}]))
+    runner.set__mock_result(DataFrame([[{"graphName": "g", "don't squeeze": "me now"}]]))
 
     graph_name = "g"
     G, _ = gds.graph.cypher.project(f"MATCH (s)-->(t) RETURN gds.graph.project('{graph_name}', s, t)")
@@ -34,7 +34,7 @@ def test_fstring(runner: CollectingQueryRunner, gds: GraphDataScience) -> None:
 
 @pytest.mark.parametrize("server_version", [ServerVersion(2, 4, 0)])
 def test_expression(runner: CollectingQueryRunner, gds: GraphDataScience) -> None:
-    runner.set__mock_result(DataFrame([{"graphName": "gg", "don't squeeze": "me now"}]))
+    runner.set__mock_result(DataFrame([[{"graphName": "gg", "don't squeeze": "me now"}]]))
 
     G, _ = gds.graph.cypher.project("WITH 'g' AS suffix MATCH (s)-->(t) RETURN gds.graph.project('g' + suffix, s, t)")
 
@@ -46,7 +46,7 @@ def test_expression(runner: CollectingQueryRunner, gds: GraphDataScience) -> Non
 
 @pytest.mark.parametrize("server_version", [ServerVersion(2, 4, 0)])
 def test_with_parameter(runner: CollectingQueryRunner, gds: GraphDataScience) -> None:
-    runner.set__mock_result(DataFrame([{"graphName": "g", "don't squeeze": "me now"}]))
+    runner.set__mock_result(DataFrame([[{"graphName": "g", "don't squeeze": "me now"}]]))
 
     G, _ = gds.graph.cypher.project("MATCH (s)-->(t) RETURN gds.graph.project($the_graph, s, t)", the_graph="g")
 
@@ -58,7 +58,7 @@ def test_with_parameter(runner: CollectingQueryRunner, gds: GraphDataScience) ->
 
 @pytest.mark.parametrize("server_version", [ServerVersion(2, 4, 0)])
 def test_with_lots_of_whitespace(runner: CollectingQueryRunner, gds: GraphDataScience) -> None:
-    runner.set__mock_result(DataFrame([{"graphName": "g", "don't squeeze": "me now"}]))
+    runner.set__mock_result(DataFrame([[{"graphName": "g", "don't squeeze": "me now"}]]))
 
     G, _ = gds.graph.cypher.project("MATCH (s)-->(t) RETURN gds .graph. project\n(\t'g'  ,s, t)")
 
@@ -70,7 +70,7 @@ def test_with_lots_of_whitespace(runner: CollectingQueryRunner, gds: GraphDataSc
 
 @pytest.mark.parametrize("server_version", [ServerVersion(2, 4, 0)])
 def test_extracting_graph_name(runner: CollectingQueryRunner, gds: GraphDataScience) -> None:
-    runner.set__mock_result(DataFrame([{"graphName": "the graph", "don't squeeze": "me now"}]))
+    runner.set__mock_result(DataFrame([[{"graphName": "the graph", "don't squeeze": "me now"}]]))
 
     G, _ = gds.graph.cypher.project("MATCH (s)-->(t) RETURN gds.graph.project('g', s, t)")