feat: add evaluation service module for RAG and Agent (#2070)

dbgpt/rag/evaluation/answer.py

@@ -287,7 +287,7 @@ class AnswerEvaluatorOperator(JoinOperator[List[EvaluationResult]]):
                     contexts=contexts,
                     passing=result.passing,
                     raw_dataset=raw_dataset,
-                    metric_name=metric.name,
+                    metric_name=metric.name(),
                     feedback=result.feedback,
                 )
             )

dbgpt/rag/index/base.py

@@ -184,6 +184,7 @@ class IndexStoreBase(ABC):
             max_threads,
         )
 
+    @abstractmethod
     def similar_search(
         self, text: str, topk: int, filters: Optional[MetadataFilters] = None
     ) -> List[Chunk]:
@@ -196,16 +197,26 @@ class IndexStoreBase(ABC):
         Return:
             List[Chunk]: The similar documents.
         """
-        return self.similar_search_with_scores(text, topk, 0.0, filters)
+
+    async def asimilar_search(
+        self,
+        query: str,
+        topk: int,
+        filters: Optional[MetadataFilters] = None,
+    ) -> List[Chunk]:
+        """Async similar_search in vector database."""
+        return await blocking_func_to_async_no_executor(
+            self.similar_search, query, topk, filters
+        )
 
     async def asimilar_search_with_scores(
         self,
-        doc: str,
+        query: str,
         topk: int,
         score_threshold: float,
         filters: Optional[MetadataFilters] = None,
     ) -> List[Chunk]:
-        """Aynsc similar_search_with_score in vector database."""
+        """Async similar_search_with_score in vector database."""
         return await blocking_func_to_async_no_executor(
-            self.similar_search_with_scores, doc, topk, score_threshold, filters
+            self.similar_search_with_scores, query, topk, score_threshold, filters
         )

dbgpt/rag/operators/evaluation.py

@@ -54,7 +54,7 @@ class RetrieverEvaluatorOperator(JoinOperator[List[EvaluationResult]]):
                     contexts=contexts,
                     passing=result.passing,
                     raw_dataset=raw_dataset,
-                    metric_name=metric.name,
+                    metric_name=metric.name(),
                 )
             )
         return results

dbgpt/rag/retriever/embedding.py

@@ -10,7 +10,6 @@ from dbgpt.rag.retriever.rerank import DefaultRanker, Ranker
 from dbgpt.rag.retriever.rewrite import QueryRewrite
 from dbgpt.storage.vector_store.filters import MetadataFilters
 from dbgpt.util.chat_util import run_async_tasks
-from dbgpt.util.executor_utils import blocking_func_to_async_no_executor
 from dbgpt.util.tracer import root_tracer
 
 
@@ -241,9 +240,7 @@ class EmbeddingRetriever(BaseRetriever):
                 "query": query,
             },
         ):
-            return await blocking_func_to_async_no_executor(
-                self._index_store.similar_search, query, self._top_k, filters
-            )
+            return await self._index_store.asimilar_search(query, self._top_k, filters)
 
     async def _run_async_tasks(self, tasks) -> List[Chunk]:
         """Run async tasks."""