修复：切换embed_model时，FAISS向量库未正确释放，导致`d == self.d assert

error`；
ApiRequest中chat接口增加max_tokens参数；FileDocModel模型字段错误(chatchat-space#1691)

server/db/models/knowledge_file_model.py

@@ -37,4 +37,4 @@ class FileDocModel(Base):
     meta_data = Column(JSON, default={})
 
     def __repr__(self):
-        return f"<FileDoc(id='{self.id}', kb_name='{self.kb_name}', file_name='{self.file_name}', doc_id='{self.doc_id}', metadata='{self.metadata}')>"
+        return f"<FileDoc(id='{self.id}', kb_name='{self.kb_name}', file_name='{self.file_name}', doc_id='{self.doc_id}', metadata='{self.meta_data}')>"

server/knowledge_base/kb_doc_api.py

@@ -337,8 +337,9 @@ def output():
         if not kb.exists() and not allow_empty_kb:
             yield {"code": 404, "msg": f"未找到知识库 ‘{knowledge_base_name}’"}
         else:
+            if kb.exists():
+                kb.clear_vs()
             kb.create_kb()
-            kb.clear_vs()
             files = list_files_from_folder(knowledge_base_name)
             kb_files = [(file, knowledge_base_name) for file in files]
             i = 0

server/knowledge_base/kb_service/faiss_kb_service.py

@@ -89,7 +89,7 @@ def do_delete_doc(self,
 
     def do_clear_vs(self):
         with kb_faiss_pool.atomic:
-            kb_faiss_pool.pop(self.kb_name)
+            kb_faiss_pool.pop((self.kb_name, self.vector_name))
         shutil.rmtree(self.vs_path)
         os.makedirs(self.vs_path)
 

server/knowledge_base/migrate.py

@@ -72,6 +72,7 @@ def files2vs(kb_name: str, kb_files: List[KnowledgeFile]):
         # 清除向量库，从本地文件重建
         if mode == "recreate_vs":
             kb.clear_vs()
+            kb.create_kb()
             kb_files = file_to_kbfile(kb_name, list_files_from_folder(kb_name))
             files2vs(kb_name, kb_files)
             kb.save_vector_store()

webui_pages/utils.py

@@ -275,7 +275,7 @@ def chat_fastchat(
         stream: bool = True,
         model: str = LLM_MODEL,
         temperature: float = TEMPERATURE,
-        max_tokens: int = 1024, # todo:根据message内容自动计算max_tokens
+        max_tokens: int = 1024, # TODO:根据message内容自动计算max_tokens
         no_remote_api: bool = None,
         **kwargs: Any,
     ):
@@ -316,6 +316,7 @@ def chat_chat(
         stream: bool = True,
         model: str = LLM_MODEL,
         temperature: float = TEMPERATURE,
+        max_tokens: int = 1024,
         prompt_name: str = "llm_chat",
         no_remote_api: bool = None,
     ):
@@ -331,6 +332,7 @@ def chat_chat(
             "stream": stream,
             "model_name": model,
             "temperature": temperature,
+            "max_tokens": max_tokens,
             "prompt_name": prompt_name,
         }
 
@@ -346,13 +348,14 @@ def chat_chat(
             return self._httpx_stream2generator(response)
 
     def agent_chat(
-            self,
-            query: str,
-            history: List[Dict] = [],
-            stream: bool = True,
-            model: str = LLM_MODEL,
-            temperature: float = TEMPERATURE,
-            no_remote_api: bool = None,
+        self,
+        query: str,
+        history: List[Dict] = [],
+        stream: bool = True,
+        model: str = LLM_MODEL,
+        temperature: float = TEMPERATURE,
+        max_tokens: int = 1024,
+        no_remote_api: bool = None,
     ):
         '''
         对应api.py/chat/agent_chat 接口
@@ -366,6 +369,7 @@ def agent_chat(
             "stream": stream,
             "model_name": model,
             "temperature": temperature,
+            "max_tokens": max_tokens,
         }
 
         print(f"received input message:")
@@ -389,6 +393,7 @@ def knowledge_base_chat(
         stream: bool = True,
         model: str = LLM_MODEL,
         temperature: float = TEMPERATURE,
+        max_tokens: int = 1024,
         prompt_name: str = "knowledge_base_chat",
         no_remote_api: bool = None,
     ):
@@ -407,6 +412,7 @@ def knowledge_base_chat(
             "stream": stream,
             "model_name": model,
             "temperature": temperature,
+            "max_tokens": max_tokens,
             "local_doc_url": no_remote_api,
             "prompt_name": prompt_name,
         }
@@ -435,6 +441,7 @@ def search_engine_chat(
         stream: bool = True,
         model: str = LLM_MODEL,
         temperature: float = TEMPERATURE,
+        max_tokens: int = 1024,
         prompt_name: str = "knowledge_base_chat",
         no_remote_api: bool = None,
     ):
@@ -452,6 +459,7 @@ def search_engine_chat(
             "stream": stream,
             "model_name": model,
             "temperature": temperature,
+            "max_tokens": max_tokens,
             "prompt_name": prompt_name,
         }
 
@@ -475,7 +483,7 @@ def search_engine_chat(
     def _check_httpx_json_response(
             self,
             response: httpx.Response,
-            errorMsg: str = f"无法连接API服务器，请确认已执行python server\\api.py",
+            errorMsg: str = f"无法连接API服务器，请确认 ‘api.py’ 已正常启动。",
         ) -> Dict:
         '''
         check whether httpx returns correct data with normal Response.