From e6cf5532a95e8d406e3cc59873c20a6d2dcf6291 Mon Sep 17 00:00:00 2001 From: qingxu fu <505030475@qq.com> Date: Sun, 9 Apr 2023 18:59:43 +0800 Subject: [PATCH] =?UTF-8?q?=E6=8F=90=E5=8D=87=E7=A8=B3=E5=AE=9A=E6=80=A7?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- crazy_functional.py | 2 +- crazy_functions/Latex全文润色.py | 70 ++++++ crazy_functions/crazy_utils.py | 249 +++++++++++++++++++-- crazy_functions/代码重写为全英文_多线程.py | 9 +- crazy_functions/批量翻译PDF文档_多线程.py | 3 +- crazy_functions/解析项目源代码.py | 152 +++++++------ objdump.tmp | Bin 0 -> 26593 bytes request_llm/bridge_chatgpt.py | 7 +- toolbox.py | 2 + version | 4 +- 10 files changed, 387 insertions(+), 111 deletions(-) create mode 100644 crazy_functions/Latex全文润色.py create mode 100644 objdump.tmp diff --git a/crazy_functional.py b/crazy_functional.py index 6bb11b9..7be9355 100644 --- a/crazy_functional.py +++ b/crazy_functional.py @@ -29,7 +29,7 @@ def get_crazy_functions(): "Color": "stop", # 按钮颜色 "Function": HotReload(解析一个C项目的头文件) }, - "解析整个C++项目（.cpp/.h）": { + "解析整个C++项目（.cpp/.hpp/.c/.h）": { "Color": "stop", # 按钮颜色 "AsButton": False, # 加入下拉菜单中 "Function": HotReload(解析一个C项目) diff --git a/crazy_functions/Latex全文润色.py b/crazy_functions/Latex全文润色.py new file mode 100644 index 0000000..b669864 --- /dev/null +++ b/crazy_functions/Latex全文润色.py @@ -0,0 +1,70 @@ +from request_llm.bridge_chatgpt import predict_no_ui +from toolbox import CatchException, report_execption, write_results_to_file, predict_no_ui_but_counting_down +fast_debug = False + + +def 解析Paper(file_manifest, project_folder, top_p, temperature, chatbot, history, systemPromptTxt): + import time, glob, os + print('begin analysis on:', file_manifest) + for index, fp in enumerate(file_manifest): + with open(fp, 'r', encoding='utf-8') as f: + file_content = f.read() + + prefix = "接下来请你逐文件分析下面的论文文件，概括其内容" if index==0 else "" + i_say = prefix + f'请对下面的文章片段用中文做一个概述，文件名是{os.path.relpath(fp, project_folder)}，文章内容是 ```{file_content}```' + i_say_show_user = prefix + f'[{index}/{len(file_manifest)}] 请对下面的文章片段做一个概述: {os.path.abspath(fp)}' + chatbot.append((i_say_show_user, "[Local Message] waiting gpt response.")) + print('[1] yield chatbot, history') + yield chatbot, history, '正常' + + if not fast_debug: + msg = '正常' + # ** gpt request ** + gpt_say = yield from predict_no_ui_but_counting_down(i_say, i_say_show_user, chatbot, top_p, temperature, history=[]) # 带超时倒计时 + + print('[2] end gpt req') + chatbot[-1] = (i_say_show_user, gpt_say) + history.append(i_say_show_user); history.append(gpt_say) + print('[3] yield chatbot, history') + yield chatbot, history, msg + print('[4] next') + if not fast_debug: time.sleep(2) + + all_file = ', '.join([os.path.relpath(fp, project_folder) for index, fp in enumerate(file_manifest)]) + i_say = f'根据以上你自己的分析，对全文进行概括，用学术性语言写一段中文摘要，然后再写一段英文摘要（包括{all_file}）。' + chatbot.append((i_say, "[Local Message] waiting gpt response.")) + yield chatbot, history, '正常' + + if not fast_debug: + msg = '正常' + # ** gpt request ** + gpt_say = yield from predict_no_ui_but_counting_down(i_say, i_say, chatbot, top_p, temperature, history=history) # 带超时倒计时 + + chatbot[-1] = (i_say, gpt_say) + history.append(i_say); history.append(gpt_say) + yield chatbot, history, msg + res = write_results_to_file(history) + chatbot.append(("完成了吗？", res)) + yield chatbot, history, msg + + + +@CatchException +def 读文章写摘要(txt, top_p, temperature, chatbot, history, systemPromptTxt, WEB_PORT): + history = [] # 清空历史，以免输入溢出 + import glob, os + if os.path.exists(txt): + project_folder = txt + else: + if txt == "": txt = '空空如也的输入栏' + report_execption(chatbot, history, a = f"解析项目: {txt}", b = f"找不到本地项目或无权访问: {txt}") + yield chatbot, history, '正常' + return + file_manifest = [f for f in glob.glob(f'{project_folder}/**/*.tex', recursive=True)] # + \ + # [f for f in glob.glob(f'{project_folder}/**/*.cpp', recursive=True)] + \ + # [f for f in glob.glob(f'{project_folder}/**/*.c', recursive=True)] + if len(file_manifest) == 0: + report_execption(chatbot, history, a = f"解析项目: {txt}", b = f"找不到任何.tex文件: {txt}") + yield chatbot, history, '正常' + return + yield from 解析Paper(file_manifest, project_folder, top_p, temperature, chatbot, history, systemPromptTxt) diff --git a/crazy_functions/crazy_utils.py b/crazy_functions/crazy_utils.py index ac96896..58d9776 100644 --- a/crazy_functions/crazy_utils.py +++ b/crazy_functions/crazy_utils.py @@ -1,19 +1,115 @@ import traceback +from toolbox import update_ui -def request_gpt_model_in_new_thread_with_ui_alive(inputs, inputs_show_user, top_p, temperature, chatbot, history, sys_prompt, refresh_interval=0.2): +def input_clipping(inputs, history, max_token_limit): + import tiktoken + import numpy as np + from toolbox import get_conf + enc = tiktoken.encoding_for_model(*get_conf('LLM_MODEL')) + def get_token_num(txt): return len(enc.encode(txt)) + + mode = 'input-and-history' + # 当 输入部分的token占比 小于 全文的一半时，只裁剪历史 + input_token_num = get_token_num(inputs) + if input_token_num < max_token_limit//2: + mode = 'only-history' + max_token_limit = max_token_limit - input_token_num + + everything = [inputs] if mode == 'input-and-history' else [''] + everything.extend(history) + n_token = get_token_num('\n'.join(everything)) + everything_token = [get_token_num(e) for e in everything] + delta = max(everything_token) // 16 # 截断时的颗粒度 + + while n_token > max_token_limit: + where = np.argmax(everything_token) + encoded = enc.encode(everything[where]) + clipped_encoded = encoded[:len(encoded)-delta] + everything[where] = enc.decode(clipped_encoded)[:-1] # -1 to remove the may-be illegal char + everything_token[where] = get_token_num(everything[where]) + n_token = get_token_num('\n'.join(everything)) + + if mode == 'input-and-history': + inputs = everything[0] + else: + pass + history = everything[1:] + return inputs, history + +def request_gpt_model_in_new_thread_with_ui_alive( + inputs, inputs_show_user, top_p, temperature, + chatbot, history, sys_prompt, refresh_interval=0.2, + handle_token_exceed=True, + retry_times_at_unknown_error=2, + ): + """ + Request GPT model，请求GPT模型同时维持用户界面活跃。 + + 输入参数 Args （以_array结尾的输入变量都是列表，列表长度为子任务的数量，执行时，会把列表拆解，放到每个子线程中分别执行）: + inputs (string): List of inputs （输入） + inputs_show_user (string): List of inputs to show user（展现在报告中的输入，借助此参数，在汇总报告中隐藏啰嗦的真实输入，增强报告的可读性） + top_p (float): Top p value for sampling from model distribution （GPT参数，浮点数） + temperature (float): Temperature value for sampling from model distribution（GPT参数，浮点数） + chatbot: chatbot inputs and outputs （用户界面对话窗口句柄，用于数据流可视化） + history (list): List of chat history （历史，对话历史列表） + sys_prompt (string): List of system prompts （系统输入，列表，用于输入给GPT的前提提示，比如你是翻译官怎样怎样） + refresh_interval (float, optional): Refresh interval for UI (default: 0.2) （刷新时间间隔频率，建议低于1，不可高于3，仅仅服务于视觉效果） + handle_token_exceed：是否自动处理token溢出的情况，如果选择自动处理，则会在溢出时暴力截断，默认开启 + retry_times_at_unknown_error：失败时的重试次数 + + 输出 Returns: + future: 输出，GPT返回的结果 + """ import time from concurrent.futures import ThreadPoolExecutor from request_llm.bridge_chatgpt import predict_no_ui_long_connection # 用户反馈 chatbot.append([inputs_show_user, ""]) msg = '正常' - yield chatbot, [], msg + yield from update_ui(chatbot=chatbot, history=[]) executor = ThreadPoolExecutor(max_workers=16) mutable = ["", time.time()] - future = executor.submit(lambda: - predict_no_ui_long_connection( - inputs=inputs, top_p=top_p, temperature=temperature, history=history, sys_prompt=sys_prompt, observe_window=mutable) - ) + def _req_gpt(inputs, history, sys_prompt): + retry_op = retry_times_at_unknown_error + exceeded_cnt = 0 + while True: + try: + # 【第一种情况】：顺利完成 + result = predict_no_ui_long_connection( + inputs=inputs, top_p=top_p, temperature=temperature, + history=history, sys_prompt=sys_prompt, observe_window=mutable) + return result + except ConnectionAbortedError as token_exceeded_error: + # 【第二种情况】：Token溢出， + if handle_token_exceed: + exceeded_cnt += 1 + # 【选择处理】 尝试计算比例，尽可能多地保留文本 + from toolbox import get_reduce_token_percent + p_ratio, n_exceed = get_reduce_token_percent(str(token_exceeded_error)) + MAX_TOKEN = 4096 + EXCEED_ALLO = 512 + 512 * exceeded_cnt + inputs, history = input_clipping(inputs, history, max_token_limit=MAX_TOKEN-EXCEED_ALLO) + mutable[0] += f'[Local Message] 警告，文本过长将进行截断，Token溢出数：{n_exceed}。\n\n' + continue # 返回重试 + else: + # 【选择放弃】 + tb_str = '```\n' + traceback.format_exc() + '```' + mutable[0] += f"[Local Message] 警告，在执行过程中遭遇问题, Traceback：\n\n{tb_str}\n\n" + return mutable[0] # 放弃 + except: + # 【第三种情况】：其他错误 + tb_str = '```\n' + traceback.format_exc() + '```' + mutable[0] += f"[Local Message] 警告，在执行过程中遭遇问题, Traceback：\n\n{tb_str}\n\n" + if retry_op > 0: + retry_op -= 1 + mutable[0] += f"[Local Message] 重试中 {retry_times_at_unknown_error-retry_op}/{retry_times_at_unknown_error}：\n\n" + time.sleep(5) + continue # 返回重试 + else: + time.sleep(5) + return mutable[0] # 放弃 + + future = executor.submit(_req_gpt, inputs, history, sys_prompt) while True: # yield一次以刷新前端页面 time.sleep(refresh_interval) @@ -27,8 +123,42 @@ def request_gpt_model_in_new_thread_with_ui_alive(inputs, inputs_show_user, top_ return future.result() -def request_gpt_model_multi_threads_with_very_awesome_ui_and_high_efficiency(inputs_array, inputs_show_user_array, top_p, temperature, chatbot, history_array, sys_prompt_array, refresh_interval=0.2, max_workers=10, scroller_max_len=30): - import time +def request_gpt_model_multi_threads_with_very_awesome_ui_and_high_efficiency( + inputs_array, inputs_show_user_array, top_p, temperature, + chatbot, history_array, sys_prompt_array, + refresh_interval=0.2, max_workers=10, scroller_max_len=30, + handle_token_exceed=True, show_user_at_complete=False, + retry_times_at_unknown_error=2, + ): + """ + Request GPT model using multiple threads with UI and high efficiency + 请求GPT模型的[多线程]版。 + 具备以下功能： + 实时在UI上反馈远程数据流 + 使用线程池，可调节线程池的大小避免openai的流量限制错误 + 处理中途中止的情况 + 网络等出问题时，会把traceback和已经接收的数据转入输出 + + 输入参数 Args （以_array结尾的输入变量都是列表，列表长度为子任务的数量，执行时，会把列表拆解，放到每个子线程中分别执行）: + inputs_array (list): List of inputs （每个子任务的输入） + inputs_show_user_array (list): List of inputs to show user（每个子任务展现在报告中的输入，借助此参数，在汇总报告中隐藏啰嗦的真实输入，增强报告的可读性） + top_p (float): Top p value for sampling from model distribution （GPT参数，浮点数） + temperature (float): Temperature value for sampling from model distribution（GPT参数，浮点数） + chatbot: chatbot （用户界面对话窗口句柄，用于数据流可视化） + history_array (list): List of chat history （历史对话输入，双层列表，第一层列表是子任务分解，第二层列表是对话历史） + sys_prompt_array (list): List of system prompts （系统输入，列表，用于输入给GPT的前提提示，比如你是翻译官怎样怎样） + refresh_interval (float, optional): Refresh interval for UI (default: 0.2) （刷新时间间隔频率，建议低于1，不可高于3，仅仅服务于视觉效果） + max_workers (int, optional): Maximum number of threads (default: 10) （最大线程数，如果子任务非常多，需要用此选项防止高频地请求openai导致错误） + scroller_max_len (int, optional): Maximum length for scroller (default: 30)（数据流的显示最后收到的多少个字符，仅仅服务于视觉效果） + handle_token_exceed (bool, optional): （是否在输入过长时，自动缩减文本） + handle_token_exceed：是否自动处理token溢出的情况，如果选择自动处理，则会在溢出时暴力截断，默认开启 + show_user_at_complete (bool, optional): (在结束时，把完整输入-输出结果显示在聊天框) + retry_times_at_unknown_error：子任务失败时的重试次数 + + 输出 Returns: + list: List of GPT model responses （每个子任务的输出汇总，如果某个子任务出错，response中会携带traceback报错信息，方便调试和定位问题。） + """ + import time, random from concurrent.futures import ThreadPoolExecutor from request_llm.bridge_chatgpt import predict_no_ui_long_connection assert len(inputs_array) == len(history_array) @@ -40,20 +170,61 @@ def request_gpt_model_multi_threads_with_very_awesome_ui_and_high_efficiency(inp msg = '正常' yield chatbot, [], msg # 异步原子 - mutable = [["", time.time()] for _ in range(n_frag)] + mutable = [["", time.time(), "等待中"] for _ in range(n_frag)] def _req_gpt(index, inputs, history, sys_prompt): - try: - gpt_say = predict_no_ui_long_connection( - inputs=inputs, top_p=top_p, temperature=temperature, history=history, sys_prompt=sys_prompt, observe_window=mutable[index] - ) - except: - # 收拾残局 - tb_str = '```\n' + traceback.format_exc() + '```' - gpt_say = f"[Local Message] 线程{index}在执行过程中遭遇问题, Traceback：\n\n{tb_str}\n\n" - if len(mutable[index][0]) > 0: - gpt_say += "此线程失败前收到的回答：" + mutable[index][0] - return gpt_say + gpt_say = "" + retry_op = retry_times_at_unknown_error + exceeded_cnt = 0 + mutable[index][2] = "执行中" + while True: + try: + # 【第一种情况】：顺利完成 + # time.sleep(10); raise RuntimeError("测试") + gpt_say = predict_no_ui_long_connection( + inputs=inputs, top_p=top_p, temperature=temperature, history=history, + sys_prompt=sys_prompt, observe_window=mutable[index], console_slience=True + ) + mutable[index][2] = "已成功" + return gpt_say + except ConnectionAbortedError as token_exceeded_error: + # 【第二种情况】：Token溢出， + if handle_token_exceed: + exceeded_cnt += 1 + # 【选择处理】 尝试计算比例，尽可能多地保留文本 + from toolbox import get_reduce_token_percent + p_ratio, n_exceed = get_reduce_token_percent(str(token_exceeded_error)) + MAX_TOKEN = 4096 + EXCEED_ALLO = 512 + 512 * exceeded_cnt + inputs, history = input_clipping(inputs, history, max_token_limit=MAX_TOKEN-EXCEED_ALLO) + gpt_say += f'[Local Message] 警告，文本过长将进行截断，Token溢出数：{n_exceed}。\n\n' + mutable[index][2] = f"截断重试" + continue # 返回重试 + else: + # 【选择放弃】 + tb_str = '```\n' + traceback.format_exc() + '```' + gpt_say += f"[Local Message] 警告，线程{index}在执行过程中遭遇问题, Traceback：\n\n{tb_str}\n\n" + if len(mutable[index][0]) > 0: gpt_say += "此线程失败前收到的回答：\n\n" + mutable[index][0] + mutable[index][2] = "输入过长已放弃" + return gpt_say # 放弃 + except: + # 【第三种情况】：其他错误 + tb_str = '```\n' + traceback.format_exc() + '```' + gpt_say += f"[Local Message] 警告，线程{index}在执行过程中遭遇问题, Traceback：\n\n{tb_str}\n\n" + if len(mutable[index][0]) > 0: gpt_say += "此线程失败前收到的回答：\n\n" + mutable[index][0] + if retry_op > 0: + retry_op -= 1 + wait = random.randint(5, 20) + for i in range(wait):# 也许等待十几秒后，情况会好转 + mutable[index][2] = f"等待重试 {wait-i}"; time.sleep(1) + mutable[index][2] = f"重试中 {retry_times_at_unknown_error-retry_op}/{retry_times_at_unknown_error}" + continue # 返回重试 + else: + mutable[index][2] = "已失败" + wait = 5 + time.sleep(5) + return gpt_say # 放弃 + # 异步任务开始 futures = [executor.submit(_req_gpt, index, inputs, history, sys_prompt) for index, inputs, history, sys_prompt in zip( range(len(inputs_array)), inputs_array, history_array, sys_prompt_array)] @@ -68,6 +239,7 @@ def request_gpt_model_multi_threads_with_very_awesome_ui_and_high_efficiency(inp break # 更好的UI视觉效果 observe_win = [] + # print([mutable[thread_index][2] for thread_index, _ in enumerate(worker_done)]) # 每个线程都要“喂狗”（看门狗） for thread_index, _ in enumerate(worker_done): mutable[thread_index][1] = time.time() @@ -77,10 +249,10 @@ def request_gpt_model_multi_threads_with_very_awesome_ui_and_high_efficiency(inp replace('\n', '').replace('```', '...').replace( ' ', '.').replace('
', '.....').replace('$', '.')+"`... ]" observe_win.append(print_something_really_funny) - stat_str = ''.join([f'执行中: {obs}\n\n' if not done else '已完成\n\n' for done, obs in zip( - worker_done, observe_win)]) - chatbot[-1] = [chatbot[-1][0], - f'多线程操作已经开始，完成情况: \n\n{stat_str}' + ''.join(['.']*(cnt % 10+1))] + stat_str = ''.join([f'`{mutable[thread_index][2]}`: {obs}\n\n' + if not done else f'`{mutable[thread_index][2]}`\n\n' + for thread_index, done, obs in zip(range(len(worker_done)), worker_done, observe_win)]) + chatbot[-1] = [chatbot[-1][0], f'多线程操作已经开始，完成情况: \n\n{stat_str}' + ''.join(['.']*(cnt % 10+1))] msg = "正常" yield chatbot, [], msg # 异步任务结束 @@ -88,9 +260,38 @@ def request_gpt_model_multi_threads_with_very_awesome_ui_and_high_efficiency(inp for inputs_show_user, f in zip(inputs_show_user_array, futures): gpt_res = f.result() gpt_response_collection.extend([inputs_show_user, gpt_res]) + + if show_user_at_complete: + for inputs_show_user, f in zip(inputs_show_user_array, futures): + gpt_res = f.result() + chatbot.append([inputs_show_user, gpt_res]) + yield chatbot, [], msg + time.sleep(1) return gpt_response_collection +def WithRetry(f): + """ + 装饰器函数，用于自动重试。 + """ + def decorated(retry, res_when_fail, *args, **kwargs): + assert retry >= 0 + while True: + try: + res = yield from f(*args, **kwargs) + return res + except: + retry -= 1 + if retry<0: + print("达到最大重试次数") + break + else: + print("重试中……") + continue + return res_when_fail + return decorated + + def breakdown_txt_to_satisfy_token_limit(txt, get_token_fn, limit): def cut(txt_tocut, must_break_at_empty_line): # 递归 if get_token_fn(txt_tocut) <= limit: diff --git a/crazy_functions/代码重写为全英文_多线程.py b/crazy_functions/代码重写为全英文_多线程.py index ad23fea..028573e 100644 --- a/crazy_functions/代码重写为全英文_多线程.py +++ b/crazy_functions/代码重写为全英文_多线程.py @@ -58,11 +58,10 @@ def 全项目切换英文(txt, top_p, temperature, chatbot, history, sys_prompt, # 第5步：Token限制下的截断与处理 MAX_TOKEN = 3000 - from transformers import GPT2TokenizerFast - print('加载tokenizer中') - tokenizer = GPT2TokenizerFast.from_pretrained("gpt2") - get_token_fn = lambda txt: len(tokenizer(txt)["input_ids"]) - print('加载tokenizer结束') + import tiktoken + from toolbox import get_conf + enc = tiktoken.encoding_for_model(*get_conf('LLM_MODEL')) + def get_token_fn(txt): return len(enc.encode(txt)) # 第6步：任务函数 diff --git a/crazy_functions/批量翻译PDF文档_多线程.py b/crazy_functions/批量翻译PDF文档_多线程.py index 348ebb9..4dae32a 100644 --- a/crazy_functions/批量翻译PDF文档_多线程.py +++ b/crazy_functions/批量翻译PDF文档_多线程.py @@ -148,7 +148,8 @@ def 解析PDF(file_manifest, project_folder, top_p, temperature, chatbot, histor file_content, page_one = read_and_clean_pdf_text(fp) # 递归地切割PDF文件 from .crazy_utils import breakdown_txt_to_satisfy_token_limit_for_pdf - enc = tiktoken.get_encoding("gpt2") + from toolbox import get_conf + enc = tiktoken.encoding_for_model(*get_conf('LLM_MODEL')) def get_token_num(txt): return len(enc.encode(txt)) # 分解文本 paper_fragments = breakdown_txt_to_satisfy_token_limit_for_pdf( diff --git a/crazy_functions/解析项目源代码.py b/crazy_functions/解析项目源代码.py index ced0f26..a39ef08 100644 --- a/crazy_functions/解析项目源代码.py +++ b/crazy_functions/解析项目源代码.py @@ -2,92 +2,96 @@ from request_llm.bridge_chatgpt import predict_no_ui from toolbox import CatchException, report_execption, write_results_to_file, predict_no_ui_but_counting_down fast_debug = False -def 解析源代码(file_manifest, project_folder, top_p, temperature, chatbot, history, systemPromptTxt): - import time, glob, os - print('begin analysis on:', file_manifest) + +def 解析源代码新(file_manifest, project_folder, top_p, temperature, chatbot, history, systemPromptTxt): + import os, copy + from .crazy_utils import request_gpt_model_multi_threads_with_very_awesome_ui_and_high_efficiency + from .crazy_utils import request_gpt_model_in_new_thread_with_ui_alive, WithRetry + msg = '正常' + inputs_array = [] + inputs_show_user_array = [] + history_array = [] + sys_prompt_array = [] + report_part_1 = [] + + ############################## <第一步，逐个文件分析，多线程> ################################## for index, fp in enumerate(file_manifest): with open(fp, 'r', encoding='utf-8') as f: file_content = f.read() - prefix = "接下来请你逐文件分析下面的工程" if index==0 else "" i_say = prefix + f'请对下面的程序文件做一个概述文件名是{os.path.relpath(fp, project_folder)}，文件代码是 ```{file_content}```' i_say_show_user = prefix + f'[{index}/{len(file_manifest)}] 请对下面的程序文件做一个概述: {os.path.abspath(fp)}' - chatbot.append((i_say_show_user, "[Local Message] waiting gpt response.")) - yield chatbot, history, '正常' + # 装载请求内容 + inputs_array.append(i_say) + inputs_show_user_array.append(i_say_show_user) + history_array.append([]) + sys_prompt_array.append("你是一个程序架构分析师，正在分析一个源代码项目。你的回答必须简单明了。") - if not fast_debug: - msg = '正常' + gpt_response_collection = yield from request_gpt_model_multi_threads_with_very_awesome_ui_and_high_efficiency( + inputs_array = inputs_array, + inputs_show_user_array = inputs_show_user_array, + history_array = history_array, + sys_prompt_array = sys_prompt_array, + top_p = top_p, + temperature = temperature, + chatbot = chatbot, + show_user_at_complete = True + ) - # ** gpt request ** - gpt_say = yield from predict_no_ui_but_counting_down(i_say, i_say_show_user, chatbot, top_p, temperature, history=[]) # 带超时倒计时 + report_part_1 = copy.deepcopy(gpt_response_collection) + history_to_return = report_part_1 + res = write_results_to_file(report_part_1) + chatbot.append(("完成？", "逐个文件分析已完成。" + res + "\n\n正在开始汇总。")) + yield chatbot, history_to_return, msg - chatbot[-1] = (i_say_show_user, gpt_say) - history.append(i_say_show_user); history.append(gpt_say) - yield chatbot, history, msg - if not fast_debug: time.sleep(2) - - all_file = ', '.join([os.path.relpath(fp, project_folder) for index, fp in enumerate(file_manifest)]) - i_say = f'根据以上你自己的分析，对程序的整体功能和构架做出概括。然后用一张markdown表格整理每个文件的功能（包括{all_file}）。' - chatbot.append((i_say, "[Local Message] waiting gpt response.")) - yield chatbot, history, '正常' - - if not fast_debug: - msg = '正常' - # ** gpt request ** - gpt_say = yield from predict_no_ui_but_counting_down(i_say, i_say, chatbot, top_p, temperature, history=history) # 带超时倒计时 - - chatbot[-1] = (i_say, gpt_say) - history.append(i_say); history.append(gpt_say) - yield chatbot, history, msg - res = write_results_to_file(history) - chatbot.append(("完成了吗？", res)) - yield chatbot, history, msg + ############################## <第二步，综合，单线程，分组+迭代处理> ################################## + batchsize = 16 # 10个文件为一组 + report_part_2 = [] + previous_iteration_files = [] + while True: + if len(file_manifest) == 0: break + this_iteration_file_manifest = file_manifest[:batchsize] + this_iteration_gpt_response_collection = gpt_response_collection[:batchsize*2] + file_rel_path = [os.path.relpath(fp, project_folder) for index, fp in enumerate(this_iteration_file_manifest)] + # 把“请对下面的程序文件做一个概述” 替换成 精简的 "文件名：{all_file[index]}" + for index, content in enumerate(this_iteration_gpt_response_collection): + if index%2==0: this_iteration_gpt_response_collection[index] = f"文件名：{file_rel_path[index//2]}" + previous_iteration_files.extend([os.path.relpath(fp, project_folder) for index, fp in enumerate(this_iteration_file_manifest)]) + previous_iteration_files_string = ', '.join(previous_iteration_files) + current_iteration_focus = ', '.join([os.path.relpath(fp, project_folder) for index, fp in enumerate(this_iteration_file_manifest)]) + i_say = f'根据以上分析，对程序的整体功能和构架重新做出概括。然后用一张markdown表格整理每个文件的功能（包括{previous_iteration_files_string}）。' + inputs_show_user = f'根据以上分析，对程序的整体功能和构架重新做出概括，由于输入长度限制，可能需要分组处理，本组文件为 {current_iteration_focus} + 已经汇总的文件组。' + this_iteration_history = copy.deepcopy(this_iteration_gpt_response_collection) + this_iteration_history.extend(report_part_2) + result = yield from request_gpt_model_in_new_thread_with_ui_alive( + inputs=i_say, inputs_show_user=inputs_show_user, top_p=top_p, temperature=temperature, chatbot=chatbot, + history=this_iteration_history, # 迭代之前的分析 + sys_prompt="你是一个程序架构分析师，正在分析一个源代码项目。") + report_part_2.extend([i_say, result]) + file_manifest = file_manifest[batchsize:] + gpt_response_collection = gpt_response_collection[batchsize*2:] + ############################## ################################## + history_to_return.extend(report_part_2) + res = write_results_to_file(history_to_return) + chatbot.append(("完成了吗？", res)) + yield chatbot, history_to_return, msg @CatchException def 解析项目本身(txt, top_p, temperature, chatbot, history, systemPromptTxt, WEB_PORT): history = [] # 清空历史，以免输入溢出 - import time, glob, os + import glob file_manifest = [f for f in glob.glob('./*.py') if ('test_project' not in f) and ('gpt_log' not in f)] + \ - [f for f in glob.glob('./crazy_functions/*.py') if ('test_project' not in f) and ('gpt_log' not in f)] - for index, fp in enumerate(file_manifest): - # if 'test_project' in fp: continue - with open(fp, 'r', encoding='utf-8') as f: - file_content = f.read() - - prefix = "接下来请你分析自己的程序构成，别紧张，" if index==0 else "" - i_say = prefix + f'请对下面的程序文件做一个概述文件名是{fp}，文件代码是 ```{file_content}```' - i_say_show_user = prefix + f'[{index}/{len(file_manifest)}] 请对下面的程序文件做一个概述: {os.path.abspath(fp)}' - chatbot.append((i_say_show_user, "[Local Message] waiting gpt response.")) - yield chatbot, history, '正常' - - if not fast_debug: - # ** gpt request ** - # gpt_say = predict_no_ui(inputs=i_say, top_p=top_p, temperature=temperature) - gpt_say = yield from predict_no_ui_but_counting_down(i_say, i_say_show_user, chatbot, top_p, temperature, history=[], long_connection=True) # 带超时倒计时 - - chatbot[-1] = (i_say_show_user, gpt_say) - history.append(i_say_show_user); history.append(gpt_say) - yield chatbot, history, '正常' - time.sleep(2) - - i_say = f'根据以上你自己的分析，对程序的整体功能和构架做出概括。然后用一张markdown表格整理每个文件的功能（包括{file_manifest}）。' - chatbot.append((i_say, "[Local Message] waiting gpt response.")) - yield chatbot, history, '正常' - - if not fast_debug: - # ** gpt request ** - # gpt_say = predict_no_ui(inputs=i_say, top_p=top_p, temperature=temperature, history=history) - gpt_say = yield from predict_no_ui_but_counting_down(i_say, i_say, chatbot, top_p, temperature, history=history, long_connection=True) # 带超时倒计时 - - chatbot[-1] = (i_say, gpt_say) - history.append(i_say); history.append(gpt_say) - yield chatbot, history, '正常' - res = write_results_to_file(history) - chatbot.append(("完成了吗？", res)) + [f for f in glob.glob('./crazy_functions/*.py') if ('test_project' not in f) and ('gpt_log' not in f)]+ \ + [f for f in glob.glob('./request_llm/*.py') if ('test_project' not in f) and ('gpt_log' not in f)] + project_folder = './' + if len(file_manifest) == 0: + report_execption(chatbot, history, a = f"解析项目: {txt}", b = f"找不到任何python文件: {txt}") yield chatbot, history, '正常' + return + yield from 解析源代码新(file_manifest, project_folder, top_p, temperature, chatbot, history, systemPromptTxt) @CatchException def 解析一个Python项目(txt, top_p, temperature, chatbot, history, systemPromptTxt, WEB_PORT): @@ -105,7 +109,7 @@ def 解析一个Python项目(txt, top_p, temperature, chatbot, history, systemPr report_execption(chatbot, history, a = f"解析项目: {txt}", b = f"找不到任何python文件: {txt}") yield chatbot, history, '正常' return - yield from 解析源代码(file_manifest, project_folder, top_p, temperature, chatbot, history, systemPromptTxt) + yield from 解析源代码新(file_manifest, project_folder, top_p, temperature, chatbot, history, systemPromptTxt) @CatchException @@ -126,7 +130,7 @@ def 解析一个C项目的头文件(txt, top_p, temperature, chatbot, history, s report_execption(chatbot, history, a = f"解析项目: {txt}", b = f"找不到任何.h头文件: {txt}") yield chatbot, history, '正常' return - yield from 解析源代码(file_manifest, project_folder, top_p, temperature, chatbot, history, systemPromptTxt) + yield from 解析源代码新(file_manifest, project_folder, top_p, temperature, chatbot, history, systemPromptTxt) @CatchException def 解析一个C项目(txt, top_p, temperature, chatbot, history, systemPromptTxt, WEB_PORT): @@ -147,7 +151,7 @@ def 解析一个C项目(txt, top_p, temperature, chatbot, history, systemPromptT report_execption(chatbot, history, a = f"解析项目: {txt}", b = f"找不到任何.h头文件: {txt}") yield chatbot, history, '正常' return - yield from 解析源代码(file_manifest, project_folder, top_p, temperature, chatbot, history, systemPromptTxt) + yield from 解析源代码新(file_manifest, project_folder, top_p, temperature, chatbot, history, systemPromptTxt) @CatchException @@ -169,7 +173,7 @@ def 解析一个Java项目(txt, top_p, temperature, chatbot, history, systemProm report_execption(chatbot, history, a=f"解析项目: {txt}", b=f"找不到任何java文件: {txt}") yield chatbot, history, '正常' return - yield from 解析源代码(file_manifest, project_folder, top_p, temperature, chatbot, history, systemPromptTxt) + yield from 解析源代码新(file_manifest, project_folder, top_p, temperature, chatbot, history, systemPromptTxt) @CatchException @@ -192,7 +196,7 @@ def 解析一个Rect项目(txt, top_p, temperature, chatbot, history, systemProm report_execption(chatbot, history, a=f"解析项目: {txt}", b=f"找不到任何Rect文件: {txt}") yield chatbot, history, '正常' return - yield from 解析源代码(file_manifest, project_folder, top_p, temperature, chatbot, history, systemPromptTxt) + yield from 解析源代码新(file_manifest, project_folder, top_p, temperature, chatbot, history, systemPromptTxt) @CatchException @@ -211,4 +215,4 @@ def 解析一个Golang项目(txt, top_p, temperature, chatbot, history, systemPr report_execption(chatbot, history, a=f"解析项目: {txt}", b=f"找不到任何golang文件: {txt}") yield chatbot, history, '正常' return - yield from 解析源代码(file_manifest, project_folder, top_p, temperature, chatbot, history, systemPromptTxt) + yield from 解析源代码新(file_manifest, project_folder, top_p, temperature, chatbot, history, systemPromptTxt) diff --git a/objdump.tmp b/objdump.tmp new file mode 100644 index 0000000000000000000000000000000000000000..d5186a03ae4c6f5b581c7d04b82141904122fc06 GIT binary patch literal 26593 zcmc&-?QdJxdGEG<@-Sf7_W>70H#lhG)lR#0#lCcD)^v-PAaSr1azW7&EwhnCRiu)* zz&6UZBvPVYY*TNRWJ|In%P*qj_(c*aN*MA#><`#@FE57y`7-SDe$Vqe_wbUUlMS1_ zqD`rL&pqdPJJ0($Nq>9#pS!-L|M_M9WNyiuSg-7jc}wg5_HL!TTuV=RGb5GKb2l^W z-JZg$+R|$E)-`u`y}CKp_WeJ3^Ykwr40X5n?K`}1i<9QF!CRHRbY<^>w{gWk+WE2b z){oDBl#KN!&tDw)b-cIh^MUg}jU_)xCHl^H^`*{tT#Baf|B+}%v@_Nd@4$mt$0w1# z{>0~(PxW2SfBf3-eCu2O_Bto`=C-T(%}Q~?yPNjztXE2Rt5d`N(QV-1-pF|iPraEP zdAK|Xi2pTx#T^1lHx;+1UmfJ^JZ^5t8+%kK3Fhv?0-$wA%HH(OR|mIFoH%{TX%9hh zQecV4lCQPn?Tc2&jQe|G9sZcNcC~ zOEe2_^h`(5$!zLG0@*1>rF*I1s&sw-q)S}SWhAq z3;qs)=B|NvWpDOgHJkAkAN^~35GeZPG6=&bBoChZ_wKqAV@=Q9%#N2EaSOMq3kA14 ztw6mAp!%Yzj!Ozte=M1b_NPwb$)(V`{>E^1VaBe@Re)k`WZa$Frho|2h1=DM9SBG* zUm&}5i63Bz=s+qF8R+YbreY^U6NAnE_Nsrhh&jxx#haDAEl}25Ok41U0V)$fS}3mA z%YqwMKwNiHy5*a$Y?%^Q^N*(dm7EN~^a^Z96dM2v$-eh&%-n;`nHg;%bZ@jjc_BRGAM+GmxFEcN0TDo8fNE0_Es6%K07|G#nuDG|y zWaiSkJ2?u<0LNPTR^?~~V?O%eT_KConj#4jyF*W)L`4Iw03N};T0FuA17m<@_2D)U zcPDr8+b^v7+k-&Xo7p#pK)HfBEqGW3@+-iX7X?|6v~a7kpEckqgM^IAe!}UMR{ZBr zfVDf71LEG}D>8{f0-AX9+n7I}OSYYPHNp&g6aN&7M}_*#Z-s;!679TyIdv(~>(o~E zp%pD6R(vS~B4CO&7$gE3`J1t@p9&-GV#Yo>?xdNRQ&O6%QTwxzV2nFFc{#;_dh)4kRCG2*yBp^21bM@0xLL?mzm*bFXBwo>=G8RhN@gD7NDvl<4GgT=rH z+v3EDx|+?7|4{U(zM=`K?2ox?@U%5FFnd8v3#827OkbxEFdhgq`HO(t6ERj4s;Mx8lA|^ z3^QtUxbhP`41jQpH!(51!QcUhrKk~*`VaeV(clO=pcaVdTVwmOv<6@R- z9)`GbI2Kxp(93G-=g|xhEb!#7qn}2t?21<18Rj_5V9dzUEF}ck7>%dSBUx#|z=uRS zz({b(-gh5Qa~cpX(HXlG?~bQp?dnvtUySv3#5!STI%9}SRj2O^cO;YE@-FlPzOOzm zf$G(JJ2d=a*xQq-%iXbLFdku9epA%QQbtwyB+H&kpoX&Wb%jq4>C!|?MIA83=R~$AfgZ)iVJ}&$4!L10C0598sT=hIj#6O zf|^4C0;>>fSjlt(MmZ2hNCIaU=9gyt!y$L{KAy|;8t!^qYjhlCmXIKcE3aUAPd1@i zyosz89AW9?Har5^SLH=NJtQhFJMQi*A=1Mzt_7zY+{Qv)_8vG{M+I>_!N*V{gc*bQ zKqx_P#5>4C=IyG`#R8%_I2J|7STFJp;&GjP&Cz`>zP_E9dQPTS_Z_K;uGr^}AYaTQ zkaE83R3B|X#<;rkZuDI2V+Vl-bfP=-2;gH$8~(~pC{1|FBkssBb+)L*?Glu-;b5Ix zY6$@b*@pect!B}!UW4A55GmFK3tR@60IF(y#fRMBfFPgy$F@Ix^^zMnKH zSWP}Os{?Uz?>hm)_NEP*TZst_f7$R|d>!ByLug$}6feQ}ILSPF7w-($1W##+@l&>I&U z#}SHo5Ce;zP`xQ1G3l~61~!n3mI~q&s$#NodNH2*jjBwcJd`&aJ0!X-3uHe8T)=d- zksR|LVUEa{o#F5kcWcheAIf5y{B|?52&@1ZMp6$2L|W%7pd!ZdN3U`Rz-35u1{>h_ zS{O?v@aUuIo(X+HP8lf;at{Vs{Vv|yH-Nxbb{W=y7+FtMIIK1W4{eI5Hb3QNo(olm z)}fBPam}nuA0g9*rM%^RJYhzyFVY`{^#P}PBeBmrVzEwe&%Zn9W_9BQpM#=C#RY3s zA3drRA>1aBcqyJtCHgNTVt_MLyQH1jV5r=QVR!P2FhIth=t^o{BwJ6Exjf8^ z;HkW}%iCpwHK1HW%lxkv(0Q13Hl!cAS}ng;KIv+g2$1x(C2dlZGZj>#PK(m zn=?3n!4gkc{aVu(cx)6rLG*grL=y-mn&%QGgxoYkqsfQJusiQ0QXj;mnxc8O-F4K4 zss)7pL3UYPwd&O^mL>4qq(YSD-sBApVk%*y zKb1o0kGi{GR7;co;UU~hXlS>P zs~%3Upe3ks#fZ#56$S01n`zz3`f7}3xO37!f6w9&SC8&Fzoc5wj^iKf^?=1t?p;zDh8 z3HexA@bPUgeNT7-CrdN8VjeVtr&WcY4n*zXuMZlr=EuG8kBh4JMc3moS(sZ#1j%0 zdAF~*C5bKk!BKbZzVoy9&pC)-q__YiP;t27YF|<=5P)# zzwPA)IR}d2$OnmX)wE8S)tHhB7(b}NDtn_9WX>=+q*tW8(ojJ(KFriZh?+&&2!~d& z%uXjIDSZOr-5sQGK@Vo-s}tMq>Q$!q@lZ+(?)CxftstlYqmXXlHhe-tgFma&;_e_H z;1^aftWp#X)Y90X*YOs#udU88TYx`}H=t6k@q}e6nZR752_oxa1eeOrBA%YA>Em*Q8Lb3B_JW)dq>4-;B0?iorBG3#&&By0 zmTX@M&rydfhN%IQ-zb!v11p7PhhQVbCrsp=UoLq|hfQJyr<-CGD%efHeI`nrT%vgu zJ)yn2tynIfOSGMS>1ACToarY2L%N5C!|LGuOL{3h=3ml_*t}q(8N*Bqy%c+t z7LJ&hG}cCoA5FlE+0z{kWiP{!smbnpg;DS7kURWLsHnS5rro&0-{1YNrII|>_vhQ3 z^S^`)opfG8dfxL4LQqkXlT)?f8x(yN zO0*ih^zv-#Qmlu`z2ALNyb=rn0yGUUz<^Tf+l5;QF{nz6%~A6K6~;6v(5P+9AoYP* zGL=YlUr2lo;iry-f1{oo+8M0if(k4Uy!GBs-;TWV^Pj!*F8+5;xRO5j>R^G99yLOj z8p2ZwUGmgGD&Czu)sg5<^d}{H(PAvbAAsU(<9BErfoa5lTuQS98SVVlKr$84X^h(l zM7cz&|FRMB}10Pb@N zn75Pno~%GDk;?)s^rda5g^Muf>9R4>Lgf-xc~(Hdi3jS)7^k;=({uevD{BL=sq9In~4riVmQH@#NJr6u4o}l#%$* zn>$1S!4%omQUIiUZm{k2t0+x|!vExXrUR&GrE&Dv3{$8{xIw`lnqIr{(A5fuf(pLA ziGvh_)GM(@FsGC$ZJtH?M^@-)h+2 z-XHt*05a7P>4)t>(z81f?~U}vK7;m`Eg$(Tp1Kqnh)1H`@lU1dXSz|jgk98dXmOLm zMTZ_?2lE+cphNXiyz5dV zcJX4oBaU3*WdVx11%x5S{|C_h9q|4j25YC&t6yg&+S+E4UM$ThifOuM>mgOOH0+Op z=A*K6lv&Ud0ofZ9mnZelf{Nw9RkbczeR4?;D`Z<}uWL3XW3NWCa$45ut zN}+PJWVs}}Y;v|XePmRp&Lhy3A7g79wp;+)wMi$E`W!w$B9cTRGgE#|ctD-cjkKM96&2uyt^cl+ncN{oM_Zyag!xbD;|m#V zrVSgjhRO4-z6BqsA;QW7*tHIJlCM4dx>vJfbVkWX8 z^byU5B*|`1S)oFRicj%RP#=Byn3Z4fA>Ai;c*rg6V*tB_`}$*@@s3oa7roT+$OXyB zCkA>`a2K?joa>6>4onj*6~3owA*k%^z(d9^td&aMV&S&g7Ihqsp(n;&E4IV~#Ds_v zWNE3<#&S+wOJ}{Y`&6kIb_!hNIFotT!VHJ9CBgP1RJVg}wDdjDflW@9y7}Bcw4Htx z)!YUtj6hw-Q;~=Y-|xcuG>ckj01{+*k)}q2BvxZ-R zE2QBu?sc@dWhuPk?co>(`ykf#?dN>UwZKM^s>~zCm&jdVt?+>3z);beL=s#9jr#?c z0C>15(5@yM4qP;dklho8XKEZ050?!EMf9k^TXxz*Zawv5kWr4$eO|0 zaG}su^Jw~;2qR@Fe4KrtiMX!FrE8A}i6ag#bD!*-Xs0{xG|5f7Hd7o-i#^|3Bbk8` zgH4k!GoDJ&8k;;7{EGBOd*nQY%s~wirlCkaD&!_m`-^F5pu8C@Zf zZP}$zA=M+Ui;}61p)`U?8!*5y=>Y^V+8ApZY|!lf{!jl(+o)FVvfYNKakzy6DZ3y@ z8a}M--?6lIN0y>d%z{938kzyn| zoC4XiTM+gZ?Ns;*;U)!->YW8ZtB4>EI6xT*(5cjtP;^UOTE zbKOD~(Ev(KenE~mXp6Zjd6-Gj>Y9r3^3S1*OwlvwR+Jua_wa#~Ntl>mF)r}9lwcz8 z-$)V>XpNxRh$Ux!fP^v^_FRCHZgC`Z^%0#L=bZTP4%;9=HGIL40%LS=@O+gMGnA1R7$>e8} zB-SHlC;l9q9UTq#)Y^|~V9Qt>5Eg^}N zn(}l8$Zts+?XG+CMf89&Nx{6lJ2_t45Q_?>PpFu|W@D&HUF-Hi)(tZog+8^2awcoZ z6d(~d6?E)U)T9LlEezlofP|)~@0UuI^0Yx29=*yp6FuxoGcXS>dRwbxCZMvDutR{K z%jB7P;)N&DqOG1i@nWoC77BtHg6(GH?8BszW3=QfUyd_X_Cc6~`;Hi~NbZUY1XAi!ZTY4iUIcr&m&S>Uq}<P__bAoup0SOkG;3XdTJ_@bjb7VQll$sW)aZvcjOm?|0L^jz64 zNcv$EoEWjZp{`FL1PRA96mPeD1Js34K_VQx)q6TCvZvk8cu0DnWb-1}xvq(OvdHaD zz)7!NDbc=*JZGiB1_b62Wt1q& ztKtxQ6u8pUg7m6hxY>5*r5Ex4TY3Slu08SISikCre-EE_wa#S_4bfw?SgHL&40j&HFAj7k69aueP_RUOD6OCpSksXM@!AIvj*B2^6i-=bZAK20 zv%nX0W8k?tHiFE`+XLvz?R?{}<7ml;e>L59=B3x|TgxG%MXS4;!kmBNLEf;hW#y6a z9>t|;7NE*~X3s~2rSyLe2xvqQOlpZ|a03C>>p`>{BNuc~13|D<{Uh)poD@^yDdr39 zWCekyNJ!yAB6-YQGGC;~Yc1>H2V z9A=PUb8FVy9+%)^?Ww>M`RkYj*;8cwXMx zBkpE1Sl2S;NomIvU1LYmRDqG~u;saSgORpuhqVRCg2M^eRP3CU#x?Tm_~4+|TkmRV zDUVhPTE4X#oNl|Pj`?vlL$+H90_yaL>CL=A{)~sAI#PI`IB+vq(Xo9DUMCB+&ST;#%KFFv@S_|2?qa+Dt<&XN_)^1g!H}sIONLO&M ziB(`j1b4$7K?li?*B{gOKH|>3`wkFB{S8${~N2| zMkR(%L!#kam73Gu@@uuw)G1-h-s*DlO0P8HTaKY3}8b z&W&)s9US5jNoag7z!?H73yBaW2H_#3yQ#BN*oOprC|byZ5BtW=dd^i zLt3#KiQqlY%AE%pxy6Ali>Dc_AeY5`XT2hAP{dzn2S9UiCt$AerPD6raVmjZY9!@u zrq&Z)Aw;0`^iBnn^q2gk#1opNXCsbke*J>;f`Z~38|)V(#YI2cX>WAH>#IVU;nc8= zOzJP76`CRl>nd{@0S}2Wbgx&3P+OBeG@wSaB+8nwWJaJ9a5fm$MV=I%i9!&=AVyAJ zA^A3Wy@rebQ&c!FuGh@ zDORV4#5EfkMnwri=Aed^F2PE1g8+8OF~}`c}&bT4w+uqP49wwx0o zH^oenWv%*+jmgz)Ko}l;2zv1e7)0H|L|)6a%=oP-egami0iuT~l-InT)mXAEc!YZw z$<<|1BU)rJId;_z#-i(0LIB~d?|qNrU^uR-Sl5AB1x#Oq31QM*lnZCH-L-w$sA@Q6 z??9g^jZ738S`)*#I7k3}fczhhFS=JY#i9qT+ZW^AF`U)yjbB7dlA(uQFocBxC*2Xs zwE;-=*^9^-09`cxn{I{UOb$`t_)9<$Oe5ziEeb7w?`wkCiVZg`pp<4RM>I7?n)X>$ z)Mi%1mGq-DnIpoZ+Bb|jr?Lc#qdVtvcv-AT94Kx^73pCZzp#Vb^#~MIrq?JrSP(Eu zn&Co1GflyHQUCzYO-$F4HlGI#rA;FJ4Q(I>Ocgpz#KM?u2|?g%fK@OB`QWCRdf|_q ziy>?v)b$y3yv50!I@H;C-jLEugF!1ml8E-?IZPdl7#7-`Ia>Piz+YXgSKebpuf*)P z;94owbAQybOkmYKoyowdk={zVn&bAEQ4U%X5)zvT)A*Uj%Tjg0k?=O(YW^kPt>My3 zu`75l7n0Z^AH<>^DgFhj{w&cQ?d=L))^n6x$EJfZA^IesuWKW6HXK&UdhnEGtYwpO z_Dl`!kBD=quL8=)-m%%}Old1q(3qkfPxzDz+JGXMQeCBL=>TCaH7Jo4#Pv(cM*VxV zCTOE2sDT0KNcf(Av?Wg5m}4bPsD@(@!{g>F-ny2)2ZO0_D>&N9q5tISYSe%qzR&g-&%{SO8G<#^HKe9(?f0Ao(^a?}+ z{;@PEu38mRg>G9F!^8^)hIvZS_=2<>VuwSSFp5NoLAnx z<^K|mQa`y&hwu^$Z9^fdz{6ddK!}*0m~yAo69w# zNX&}{XVaUWy$~)2-LuSH8342ah$;OX(LgSOD?!THAB322OY{V8?S-SexCvtTW4Lxi zexo!EJzgINtho@j22!J>C#i&4>|2sTR*(R`Czgt${qv$!2u0*g?2iwyl!OPWfrX%2 z<|5{D_X{vbwoUJil#ziO?%+~Skw~Y>T>_*{L9=6=LY-mr2X-(E8V5D3sJY^N3-crO41Q!O@{tNAsMTP{y|Id*jpcsUDLAv<*xRYHt6n zqY#GJz#d*g$h_43SWP~{BSZQ}01jQ?XMj9w?*P-fDpMk1lmH-tc|8>;suTyD^&W3X zC5evr`P@X?nOELJhBI>2;yvT4(4pU4@s^HMQM}Pq{HKNg&eweN9mpsH#;|7~qKVUf zt&06JS#8c-)QG|vmZosYOvY@^Z}Jg{Wb=eCb`K?Q9rTScDnK;bx^+x@$F#2vY9CCK z3;guTB=DtD8mpGZ%}xuVt8lCV=~u4817rtO1$Q_;S^=(b;P8V^i#NNDdoyqvzFL?C z+ZksujR^UGdB%7K!xSYGhy{spBHnZe#*jrbSPFtPqU{WGR)7wlo9QjubTg%1^}3>iE7VxwOkP1PW@zy30wD& z!EdhJ^WThMyfrzBr2*muELwF%x;CM1ambORq&Sk1{>T%tQGj0<@(fmek|u+Sl&c%o zXⅅSw+?)kjJ0kI1=Wk3(J{JVUPE;W9i^Dq$q^;28Xf^sL^_8ZGV6cFQYQq^3L== zrbh3Z-0{Lga(cEA7TzAAleM*7yIM$_raA2?vr(&TLA|{nu_jKX($zx2(_51J%}E+w zrtytf#IXaL5a-S(oJsr*@+XaCC#YUJd6}Xt*U1 z!LgP2opa~jx3Qn-maq-OX%9U#3^iv@wc`RSzTt_o@$K zJ4W=fef@>o-x;>h83jVdJOsz9;VBz6ki-JNN+x>of2nwnUYnTc>cWXFfW=z`K$(2* z@7vyd>1~BA6jQ2eAg)?zmB;Cz7s8$aQ_Nb;kP-9E)7xp zXR!;(M8_wwRI=TOmR>AHwa0(w2YYe&Z+Q^O0Sgy5IlGzz$TKISwqT?OsH^}&n}p{u zDw7R^YQXIwfJRt}ig?wbL)vhNYv~(EHiRGmRPbHVVcxI||J5#m3uIjdfnS>AjdiTxkk+WiteqPKB$nk`AB*2ZyZku~YvAM@GTR literal 0 HcmV?d00001 diff --git a/request_llm/bridge_chatgpt.py b/request_llm/bridge_chatgpt.py index 3ffbd73..b3bba97 100644 --- a/request_llm/bridge_chatgpt.py +++ b/request_llm/bridge_chatgpt.py @@ -72,7 +72,7 @@ def predict_no_ui(inputs, top_p, temperature, history=[], sys_prompt=""): raise ConnectionAbortedError("Json解析不合常规，可能是文本过长" + response.text) -def predict_no_ui_long_connection(inputs, top_p, temperature, history=[], sys_prompt="", observe_window=None): +def predict_no_ui_long_connection(inputs, top_p, temperature, history=[], sys_prompt="", observe_window=None, console_slience=False): """ 发送至chatGPT，等待回复，一次性完成，不显示中间过程。但内部用stream的方法避免中途网线被掐。 inputs： @@ -121,7 +121,7 @@ def predict_no_ui_long_connection(inputs, top_p, temperature, history=[], sys_pr if "role" in delta: continue if "content" in delta: result += delta["content"] - print(delta["content"], end='') + if not console_slience: print(delta["content"], end='') if observe_window is not None: # 观测窗，把已经获取的数据显示出去 if len(observe_window) >= 1: observe_window[0] += delta["content"] @@ -264,8 +264,7 @@ def generate_payload(inputs, top_p, temperature, history, system_prompt, stream) "presence_penalty": 0, "frequency_penalty": 0, } - - print(f" {LLM_MODEL} : {conversation_cnt} : {inputs}") + print(f" {LLM_MODEL} : {conversation_cnt} : {inputs[:100]}") return headers,payload diff --git a/toolbox.py b/toolbox.py index 7c329ed..ffd44ea 100644 --- a/toolbox.py +++ b/toolbox.py @@ -21,6 +21,8 @@ def ArgsGeneralWrapper(f): yield from f(txt_passon, *args, **kwargs) return decorated +def update_ui(chatbot, history, msg='正常', *args, **kwargs): + yield chatbot, history, msg def get_reduce_token_percent(text): try: diff --git a/version b/version index 2f1e653..9916cb3 100644 --- a/version +++ b/version @@ -1,5 +1,5 @@ { - "version": 2.5, + "version": 2.6, "show_feature": true, - "new_feature": "新增一键更新程序<->高亮代码<->高亮公式<->新增垂直布局选项" + "new_feature": "增强多线程稳定性（涉及代码解析、PDF翻译等）<->修复Token计数错误（解决PDF翻译的分割不合理的问题）" }