Spaces:
				
			
			
	
			
			
					
		Running
		
	
	
	
			
			
	
	
	
	
		
		
					
		Running
		
	强调完整翻译
Browse files
    	
        crazy_functions/批量翻译PDF文档_多线程.py
    CHANGED
    
    | @@ -59,7 +59,7 @@ def 批量翻译PDF文档(txt, llm_kwargs, plugin_kwargs, chatbot, history, sys_ | |
| 59 | 
             
            def 解析PDF(file_manifest, project_folder, llm_kwargs, plugin_kwargs, chatbot, history, sys_prompt):
         | 
| 60 | 
             
                import os
         | 
| 61 | 
             
                import tiktoken
         | 
| 62 | 
            -
                TOKEN_LIMIT_PER_FRAGMENT =  | 
| 63 | 
             
                generated_conclusion_files = []
         | 
| 64 | 
             
                for index, fp in enumerate(file_manifest):
         | 
| 65 |  | 
| @@ -91,13 +91,13 @@ def 解析PDF(file_manifest, project_folder, llm_kwargs, plugin_kwargs, chatbot, | |
| 91 | 
             
                    # 多线,翻译
         | 
| 92 | 
             
                    gpt_response_collection = yield from request_gpt_model_multi_threads_with_very_awesome_ui_and_high_efficiency(
         | 
| 93 | 
             
                        inputs_array=[
         | 
| 94 | 
            -
                            f" | 
| 95 | 
             
                        inputs_show_user_array=[f"\n---\n 原文: \n\n {frag.replace('#', '')}  \n---\n 翻译:\n " for frag in paper_fragments],
         | 
| 96 | 
             
                        llm_kwargs=llm_kwargs,
         | 
| 97 | 
             
                        chatbot=chatbot,
         | 
| 98 | 
             
                        history_array=[[paper_meta] for _ in paper_fragments],
         | 
| 99 | 
             
                        sys_prompt_array=[
         | 
| 100 | 
            -
                            " | 
| 101 | 
             
                        # max_workers=5  # OpenAI所允许的最大并行过载
         | 
| 102 | 
             
                    )
         | 
| 103 |  | 
|  | |
| 59 | 
             
            def 解析PDF(file_manifest, project_folder, llm_kwargs, plugin_kwargs, chatbot, history, sys_prompt):
         | 
| 60 | 
             
                import os
         | 
| 61 | 
             
                import tiktoken
         | 
| 62 | 
            +
                TOKEN_LIMIT_PER_FRAGMENT = 1280
         | 
| 63 | 
             
                generated_conclusion_files = []
         | 
| 64 | 
             
                for index, fp in enumerate(file_manifest):
         | 
| 65 |  | 
|  | |
| 91 | 
             
                    # 多线,翻译
         | 
| 92 | 
             
                    gpt_response_collection = yield from request_gpt_model_multi_threads_with_very_awesome_ui_and_high_efficiency(
         | 
| 93 | 
             
                        inputs_array=[
         | 
| 94 | 
            +
                            f"你需要翻译以下内容:\n{frag}" for frag in paper_fragments],
         | 
| 95 | 
             
                        inputs_show_user_array=[f"\n---\n 原文: \n\n {frag.replace('#', '')}  \n---\n 翻译:\n " for frag in paper_fragments],
         | 
| 96 | 
             
                        llm_kwargs=llm_kwargs,
         | 
| 97 | 
             
                        chatbot=chatbot,
         | 
| 98 | 
             
                        history_array=[[paper_meta] for _ in paper_fragments],
         | 
| 99 | 
             
                        sys_prompt_array=[
         | 
| 100 | 
            +
                            "请你作为一个学术翻译,负责把学术论文准确翻译成中文。注意文章中的每一句话都要翻译。" for _ in paper_fragments],
         | 
| 101 | 
             
                        # max_workers=5  # OpenAI所允许的最大并行过载
         | 
| 102 | 
             
                    )
         | 
| 103 |  |