refactor: refactor proj structure & configs

Author: ChenZiHong-Gavin

examples/evaluate/evaluate_kg/kg_evaluation_config.yaml

@@ -40,3 +40,6 @@ nodes:
       - build_kg
     params:
       metrics:
+        - kg_accuracy
+        - kg_consistency
+        - kg_structure

examples/evaluate/evaluate_qa/qa_evaluation_config.yaml

@@ -74,6 +74,7 @@ nodes:
     execution_params:
       replicas: 1
       batch_size: 128
+    save_output: true
     params:
       method: aggregated # atomic, aggregated, multi_hop, cot, vqa
       data_format: ChatML # Alpaca, Sharegpt, ChatML
@@ -86,5 +87,10 @@ nodes:
     execution_params:
       replicas: 1
       batch_size: 128
+    save_output: true
     params:
       metrics:
+        - qa_length
+        - qa_mtld
+        - qa_reward_score
+        - qa_uni_score

graphgen/models/evaluator/__init__.py

@@ -1,5 +1,2 @@
-from .kg_quality_evaluator import KGQualityEvaluator
-from .length_evaluator import LengthEvaluator
-from .mtld_evaluator import MTLDEvaluator
-from .reward_evaluator import RewardEvaluator
-from .uni_evaluator import UniEvaluator
+from graphgen.models.evaluator.kg.kg_quality_evaluator import KGQualityEvaluator
+from graphgen.models.evaluator.qa.uni_evaluator import UniEvaluator

…models/evaluator/kg_quality_evaluator.py …els/evaluator/kg/kg_quality_evaluator.pygraphgen/models/evaluator/kg_quality_evaluator.py renamed to graphgen/models/evaluator/kg/kg_quality_evaluator.py graphgen/models/evaluator/kg_quality_evaluator.py renamed to graphgen/models/evaluator/kg/kg_quality_evaluator.py

@@ -76,9 +76,7 @@ def evaluate_all(self) -> Dict[str, Any]:
 
         try:
             logger.info("Starting structural robustness evaluation...")
-            structure_evaluator = StructureEvaluator(
-                graph_storage=self.graph_storage
-            )
+            structure_evaluator = StructureEvaluator(graph_storage=self.graph_storage)
             results["structure"] = structure_evaluator.evaluate()
         except Exception as e:
             logger.error(f"Structural evaluation failed: {e}")
@@ -105,7 +103,5 @@ def evaluate_consistency(self) -> Dict[str, Any]:
         return consistency_evaluator.evaluate()
 
     def evaluate_structure(self) -> Dict[str, Any]:
-        structure_evaluator = StructureEvaluator(
-            graph_storage=self.graph_storage
-        )
+        structure_evaluator = StructureEvaluator(graph_storage=self.graph_storage)
         return structure_evaluator.evaluate()