Remove Graph from prediction

Author: orazve

examples/kge-distmult-nations.py

@@ -158,15 +158,11 @@ def project_graphs(gds):
     gds.graph.drop("testGraph", failIfMissing=False)
 
     G_full, _ = gds.graph.project("fullGraph", ["Entity"], all_rels)
-    inspect_graph(G_full)
 
     G_train, _ = gds.graph.filter("trainGraph", G_full, "*", "r.split = 0.0")
     G_valid, _ = gds.graph.filter("validGraph", G_full, "*", "r.split = 1.0")
     G_test, _ = gds.graph.filter("testGraph", G_full, "*", "r.split = 2.0")
 
-    inspect_graph(G_train)
-    inspect_graph(G_valid)
-    inspect_graph(G_test)
 
     gds.graph.drop("fullGraph", failIfMissing=False)
 
@@ -190,14 +186,9 @@ def inspect_graph(G):
     create_constraint(gds)
     put_data_in_db(gds)
     G_train, G_valid, G_test = project_graphs(gds)
-    inspect_graph(G_train)
-    inspect_graph(G_valid)
-    inspect_graph(G_test)
 
     gds.set_compute_cluster_ip("localhost")
 
-    print(gds.debug.arrow())
-
     model_name = "dummyModelName_" + str(time.time())
 
     gds.kge.model.train(
@@ -210,7 +201,6 @@ def inspect_graph(G):
     )
 
     df = gds.kge.model.predict(
-        G_train,
         model_name=model_name,
         top_k=3,
         node_ids=[
@@ -221,7 +211,7 @@ def inspect_graph(G):
         rel_types=["REL_RELDIPLOMACY", "REL_RELNGO"],
     )
 
-    print(df)
+    print(df.to_string())
     #
     # gds.kge.model.predict_tail(
     #     G_train,
@@ -240,4 +230,3 @@ def inspect_graph(G):
     #     ],
     # )
 
-    print("Finished training")

graphdatascience/model/kge_runner.py

@@ -34,8 +34,6 @@ def __init__(
         self._compute_cluster_mlflow_uri = f"http://{compute_cluster_ip}:8080"
         self._encrypted_db_password = encrypted_db_password
         self._arrow_uri = arrow_uri
-        print("KgeRunner __dict__:")
-        print(self.__dict__)
 
     @property
     def model(self):
@@ -89,14 +87,12 @@ def train(
     @client_only_endpoint("gds.kge.model")
     def predict(
         self,
-        G: Graph,
         model_name: str,
         top_k: int,
         node_ids: list[int],
         rel_types: list[str],
         mlflow_experiment_name: Optional[str] = None,
     ) -> DataFrame:
-        graph_config = {"name": G.name()}
 
         algo_config = {
             "top_k": top_k,
@@ -108,7 +104,6 @@ def predict(
             "user_name": "DUMMY_USER",
             "task": "KGE_PREDICT_PYG",
             "task_config": {
-                "graph_config": graph_config,
                 "modelname": model_name,
                 "task_config": algo_config,
             },
@@ -122,8 +117,6 @@ def predict(
                 "config": {"tracking_uri": self._compute_cluster_mlflow_uri, "experiment_name": mlflow_experiment_name}
             }
 
-        print("predict config")
-        print(config)
         job_id = self._start_job(config)
 
         self._wait_for_job(job_id)
@@ -146,8 +139,6 @@ def _stream_results(self, user_name: str, model_name: str, job_id: str) -> DataF
         return df
 
     def _start_job(self, config: Dict[str, Any]) -> str:
-        print("_start_job")
-        print(config)
         url = f"{self._compute_cluster_web_uri}/api/machine-learning/start"
         print(url)
         res = requests.post(url, json=config)