Merge branch 'dev' into dbgpt_doc

.env.template

@@ -108,4 +108,4 @@ PROXY_SERVER_URL=http://127.0.0.1:3000/proxy_address
 #*******************************************************************#
 # **    SUMMARY_CONFIG
 #*******************************************************************#
-SUMMARY_CONFIG=VECTOR
+SUMMARY_CONFIG=FAST

pilot/scene/chat_knowledge/custom/chat.py

@@ -56,8 +56,9 @@ class ChatNewKnowledge(BaseChat):
         docs = self.knowledge_embedding_client.similar_search(
             self.current_user_input, VECTOR_SEARCH_TOP_K
         )
-        docs = docs[:2000]
-        input_values = {"context": docs, "question": self.current_user_input}
+        context = [d.page_content for d in docs]
+        context = context[:2000]
+        input_values = {"context": context, "question": self.current_user_input}
         return input_values
 
     def do_with_prompt_response(self, prompt_response):

pilot/scene/chat_knowledge/default/chat.py

@@ -17,7 +17,7 @@ from pilot.configs.model_config import (
     VECTOR_SEARCH_TOP_K,
 )
 
-from pilot.scene.chat_normal.prompt import prompt
+from pilot.scene.chat_knowledge.default.prompt import prompt
 from pilot.source_embedding.knowledge_embedding import KnowledgeEmbedding
 
 CFG = Config()
@@ -52,8 +52,9 @@ class ChatDefaultKnowledge(BaseChat):
         docs = self.knowledge_embedding_client.similar_search(
             self.current_user_input, VECTOR_SEARCH_TOP_K
         )
-        docs = docs[:2000]
-        input_values = {"context": docs, "question": self.current_user_input}
+        context = [d.page_content for d in docs]
+        context = context[:2000]
+        input_values = {"context": context, "question": self.current_user_input}
         return input_values
 
     def do_with_prompt_response(self, prompt_response):

pilot/scene/chat_knowledge/url/chat.py

@@ -52,14 +52,17 @@ class ChatUrlKnowledge(BaseChat):
         )
 
         # url soruce in vector
-        self.knowledge_embedding_client.knowledge_embedding()
+        if not self.knowledge_embedding_client.vector_exist():
+            self.knowledge_embedding_client.knowledge_embedding()
+        logger.info("url embedding success")
 
     def generate_input_values(self):
         docs = self.knowledge_embedding_client.similar_search(
             self.current_user_input, VECTOR_SEARCH_TOP_K
         )
-        docs = docs[:2000]
-        input_values = {"context": docs, "question": self.current_user_input}
+        context = [d.page_content for d in docs]
+        context = context[:2000]
+        input_values = {"context": context, "question": self.current_user_input}
         return input_values
 
     def do_with_prompt_response(self, prompt_response):

pilot/scene/chat_knowledge/url/prompt.py

@@ -11,11 +11,10 @@ from pilot.scene.chat_normal.out_parser import NormalChatOutputParser
 
 CFG = Config()
 
-_DEFAULT_TEMPLATE = """ 基于以下已知的信息, 专业、简要的回答用户的问题,
-            如果无法从提供的内容中获取答案, 请说: "知识库中提供的内容不足以回答此问题" 禁止胡乱编造。 
-            已知内容: 
+_DEFAULT_TEMPLATE = """ Based on the known information, provide professional and concise answers to the user's questions. If the answer cannot be obtained from the provided content, please say: 'The information provided in the knowledge base is not sufficient to answer this question.' Fabrication is prohibited.。 
+            known information: 
             {context}
-            问题:
+            question:
             {question}
 """
 

pilot/server/webserver.py

@@ -643,7 +643,7 @@ def knowledge_embedding_store(vs_id, files):
         knowledge_embedding_client.knowledge_embedding()
 
     logger.info("knowledge embedding success")
-    return os.path.join(KNOWLEDGE_UPLOAD_ROOT_PATH, vs_id, vs_id + ".vectordb")
+    return vs_id
 
 
 if __name__ == "__main__":

pilot/source_embedding/knowledge_embedding.py

@@ -82,6 +82,9 @@ class KnowledgeEmbedding:
     def similar_search(self, text, topk):
         return self.knowledge_embedding_client.similar_search(text, topk)
 
+    def vector_exist(self):
+        return self.knowledge_embedding_client.vector_name_exist()
+
     def knowledge_persist_initialization(self, append_mode):
         documents = self._load_knownlege(self.file_path)
         self.vector_client = VectorStoreConnector(