上传文件时自动总结

ChuanhuChatbot.py

@@ -333,7 +333,7 @@ with gr.Blocks(css=customCSS, theme=small_and_beautiful_theme) as demo:
     submitBtn.click(**transfer_input_args).then(**chatgpt_predict_args, api_name="predict").then(**end_outputing_args)
     submitBtn.click(**get_usage_args)
 
-    index_files.change(handle_file_upload, [current_model, index_files, chatbot], [index_files, chatbot, status_display])
+    index_files.change(handle_file_upload, [current_model, index_files, chatbot, language_select_dropdown], [index_files, chatbot, status_display])
 
     emptyBtn.click(
         reset,

locale/en_US.json

@@ -70,5 +70,7 @@
     "您的IP区域：未知。": "Your IP region: Unknown.",
     "获取IP地理位置失败。原因：": "Failed to get IP location. Reason: ",
     "。你仍然可以使用聊天功能。": ". You can still use the chat function.",
-    "您的IP区域：": "Your IP region: "
+    "您的IP区域：": "Your IP region: ",
+    "总结": "Summarize",
+    "生成内容总结中……": "Generating content summary..."
 }

locale/ja_JP.json

@@ -70,5 +70,7 @@
     "您的IP区域：未知。": "あなたのIPアドレス地域：不明",
     "获取IP地理位置失败。原因：": "IPアドレス地域の取得に失敗しました。理由：",
     "。你仍然可以使用聊天功能。": "。あなたはまだチャット機能を使用できます。",
-    "您的IP区域：": "あなたのIPアドレス地域："
+    "您的IP区域：": "あなたのIPアドレス地域：",
+    "总结": "要約する",
+    "生成内容总结中……": "コンテンツ概要を生成しています..."
 }

modules/models/base_model.py

@@ -178,12 +178,27 @@ class BaseLLMModel:
         status_text = self.token_message()
         return chatbot, status_text
 
-    def handle_file_upload(self, files, chatbot):
+    def handle_file_upload(self, files, chatbot, language):
         """if the model accepts multi modal input, implement this function"""
         status = gr.Markdown.update()
         if files:
-            construct_index(self.api_key, file_src=files)
-            status = "索引构建完成"
+            index = construct_index(self.api_key, file_src=files)
+            status = i18n("索引构建完成")
+            # Summarize the document
+            logging.info(i18n("生成内容总结中……"))
+            os.environ["OPENAI_API_KEY"] = self.api_key
+            from langchain.chains.summarize import load_summarize_chain
+            from langchain.prompts import PromptTemplate
+            from langchain.chat_models import ChatOpenAI
+            from langchain.callbacks import StdOutCallbackHandler
+            prompt_template = "Write a concise summary of the following:\n\n{text}\n\nCONCISE SUMMARY IN " + language + ":"
+            PROMPT = PromptTemplate(template=prompt_template, input_variables=["text"])
+            handler = StdOutCallbackHandler()
+            llm = ChatOpenAI(callbacks=[handler])
+            chain = load_summarize_chain(llm, chain_type="map_reduce", return_intermediate_steps=True, map_prompt=PROMPT, combine_prompt=PROMPT)
+            summary = chain({"input_documents": list(index.docstore.__dict__["_dict"].values())}, return_only_outputs=True)["output_text"]
+            print(i18n("总结") + f": {summary}")
+            chatbot.append([i18n("总结"), summary])
         return gr.Files.update(), chatbot, status
 
     def prepare_inputs(self, real_inputs, use_websearch, files, reply_language, chatbot):

modules/models/models.py

@@ -494,7 +494,7 @@ class XMChat(BaseLLMModel):
         limited_context = False
         return limited_context, fake_inputs, display_append, real_inputs, chatbot
 
-    def handle_file_upload(self, files, chatbot):
+    def handle_file_upload(self, files, chatbot, language):
         """if the model accepts multi modal input, implement this function"""
         if files:
             for file in files: