fix(storage): Fix storage connector error (#1547)

Co-authored-by: dong <dongzhancai@iie2.com>
Co-authored-by: Zhancai Dong <dongzhancai1@jd.com>
Co-authored-by: aries_ckt <916701291@qq.com>

dbgpt/rag/index/base.py

@@ -109,6 +109,10 @@ class IndexStoreBase(ABC):
             index_name(str): The name of index to delete.
         """
 
+    def vector_name_exists(self) -> bool:
+        """Whether name exists."""
+        return True
+
     def load_document_with_limit(
         self, chunks: List[Chunk], max_chunks_once_load: int = 10, max_threads: int = 1
     ) -> List[str]:

dbgpt/storage/vector_store/connector.py

@@ -87,6 +87,9 @@ class VectorStoreConnector:
             - vector_store_type: vector store type Milvus, Chroma, Weaviate
             - ctx: vector store config params.
         """
+        if vector_store_config is None:
+            raise Exception("vector_store_config is required")
+
         self._index_store_config = vector_store_config
         self._register()
 
@@ -96,36 +99,25 @@ class VectorStoreConnector:
             raise Exception(f"Vector store {vector_store_type} not supported")
 
         logger.info(f"VectorStore:{self.connector_class}")
+
         self._vector_store_type = vector_store_type
-        self._embeddings = (
-            vector_store_config.embedding_fn if vector_store_config else None
-        )
+        self._embeddings = vector_store_config.embedding_fn
 
+        config_dict = {}
+        for key in vector_store_config.to_dict().keys():
+            value = getattr(vector_store_config, key)
+            if value is not None:
+                config_dict[key] = value
+        for key, value in vector_store_config.model_extra.items():
+            if value is not None:
+                config_dict[key] = value
+        config = self.config_class(**config_dict)
         try:
-            if vector_store_config is not None:
-                config: IndexStoreConfig = self.config_class()
-                config.name = getattr(vector_store_config, "name", "default_name")
-                config.embedding_fn = getattr(vector_store_config, "embedding_fn", None)
-                config.max_chunks_once_load = getattr(
-                    vector_store_config, "max_chunks_once_load", 5
-                )
-                config.max_threads = getattr(vector_store_config, "max_threads", 4)
-                config.user = getattr(vector_store_config, "user", None)
-                config.password = getattr(vector_store_config, "password", None)
-
-                # extra
-                config_dict = vector_store_config.dict()
-                config.llm_client = config_dict.get("llm_client", None)
-                config.model_name = config_dict.get("model_name", None)
-                if (
-                    vector_store_type in pools
-                    and config.name in pools[vector_store_type]
-                ):
-                    self.client = pools[vector_store_type][config.name]
-                else:
-                    client = self.connector_class(config)
-                    pools[vector_store_type][config.name] = self.client = client
-                    self.client = client
+            if vector_store_type in pools and config.name in pools[vector_store_type]:
+                self.client = pools[vector_store_type][config.name]
+            else:
+                client = self.connector_class(config)
+                pools[vector_store_type][config.name] = self.client = client
         except Exception as e:
             logger.error("connect vector store failed: %s", e)
             raise e

examples/client/client_chat_example.py

@@ -57,12 +57,12 @@ async def main():
     # initialize client
     DBGPT_API_KEY = "dbgpt"
     client = Client(api_key=DBGPT_API_KEY)
-
-    async for data in client.chat_stream(
-        model="chatgpt_proxyllm",
-        messages="hello",
-    ):
-        print(data)
+    data = await client.chat(model="chatgpt_proxyllm", messages="hello")
+    # async for data in client.chat_stream(
+    #     model="chatgpt_proxyllm",
+    #     messages="hello",
+    # ):
+    print(data)
 
     # res = await client.chat(model="chatgpt_proxyllm" ,messages="hello")
     # print(res)

examples/rag/graph_rag_example.py

@@ -6,6 +6,7 @@ from dbgpt.model.proxy.llms.chatgpt import OpenAILLMClient
 from dbgpt.rag import ChunkParameters
 from dbgpt.rag.assembler import EmbeddingAssembler
 from dbgpt.rag.knowledge import KnowledgeFactory
+from dbgpt.storage.knowledge_graph.knowledge_graph import BuiltinKnowledgeGraphConfig
 from dbgpt.storage.vector_store.base import VectorStoreConfig
 from dbgpt.storage.vector_store.connector import VectorStoreConnector