fufa/chatgpt

1import markdown, mdtex2html, threading, importlib, traceback2from show_math import convert as convert_math3from functools import wraps4 5def predict_no_ui_but_counting_down(i_say, i_say_show_user, chatbot, top_p, temperature, history=[], sys_prompt=''):6    """7        调用简单的predict_no_ui接口，但是依然保留了些许界面心跳功能，当对话太长时，会自动采用二分法截断8    """9    import time10    from predict import predict_no_ui11    from toolbox import get_conf12    TIMEOUT_SECONDS, MAX_RETRY = get_conf('TIMEOUT_SECONDS', 'MAX_RETRY')13    # 多线程的时候，需要一个mutable结构在不同线程之间传递信息14    # list就是最简单的mutable结构，我们第一个位置放gpt输出，第二个位置传递报错信息15    mutable = [None, '']16    # multi-threading worker17    def mt(i_say, history):18        while True:19            try:20                mutable[0] = predict_no_ui(inputs=i_say, top_p=top_p, temperature=temperature, history=history, sys_prompt=sys_prompt)21                break22            except ConnectionAbortedError as e:23                if len(history) > 0:24                    history = [his[len(his)//2:] for his in history if his is not None]25                    mutable[1] = 'Warning! History conversation is too long, cut into half. '26                else:27                    i_say = i_say[:len(i_say)//2]28                    mutable[1] = 'Warning! Input file is too long, cut into half. '29            except TimeoutError as e:30                mutable[0] = '[Local Message] Failed with timeout.'31                raise TimeoutError32    # 创建新线程发出http请求33    thread_name = threading.Thread(target=mt, args=(i_say, history)); thread_name.start()34    # 原来的线程则负责持续更新UI，实现一个超时倒计时，并等待新线程的任务完成35    cnt = 036    while thread_name.is_alive():37        cnt += 138        chatbot[-1] = (i_say_show_user, f"[Local Message] {mutable[1]}waiting gpt response {cnt}/{TIMEOUT_SECONDS*2*(MAX_RETRY+1)}"+''.join(['.']*(cnt%4)))39        yield chatbot, history, '正常'40        time.sleep(1)41    # 把gpt的输出从mutable中取出来42    gpt_say = mutable[0]43    if gpt_say=='[Local Message] Failed with timeout.': raise TimeoutError44    return gpt_say45 46def write_results_to_file(history, file_name=None):47    """48        将对话记录history以Markdown格式写入文件中。如果没有指定文件名，则使用当前时间生成文件名。49    """50    import os, time51    if file_name is None:52        # file_name = time.strftime("chatGPT分析报告%Y-%m-%d-%H-%M-%S", time.localtime()) + '.md'53        file_name = 'chatGPT分析报告' + time.strftime("%Y-%m-%d-%H-%M-%S", time.localtime()) + '.md'54    os.makedirs('./gpt_log/', exist_ok=True)55    with open(f'./gpt_log/{file_name}', 'w', encoding = 'utf8') as f:56        f.write('# chatGPT 分析报告\n')57        for i, content in enumerate(history):58            if i%2==0: f.write('## ')59            f.write(content)60            f.write('\n\n')61    res = '以上材料已经被写入' + os.path.abspath(f'./gpt_log/{file_name}')62    print(res)63    return res64 65def regular_txt_to_markdown(text):66    """67        将普通文本转换为Markdown格式的文本。68    """69    text = text.replace('\n', '\n\n')70    text = text.replace('\n\n\n', '\n\n')71    text = text.replace('\n\n\n', '\n\n')72    return text73 74def CatchException(f):75    """76        装饰器函数，捕捉函数f中的异常并封装到一个生成器中返回，并显示到聊天当中。77    """78    @wraps(f)79    def decorated(txt, top_p, temperature, chatbot, history, systemPromptTxt, WEB_PORT):80        try:81            yield from f(txt, top_p, temperature, chatbot, history, systemPromptTxt, WEB_PORT)82        except Exception as e:83            from check_proxy import check_proxy84            from toolbox import get_conf85            proxies, = get_conf('proxies')86            tb_str = regular_txt_to_markdown(traceback.format_exc())87            chatbot[-1] = (chatbot[-1][0], f"[Local Message] 实验性函数调用出错: \n\n {tb_str} \n\n 当前代理可用性: \n\n {check_proxy(proxies)}")88            yield chatbot, history, f'异常 {e}'89    return decorated90 91def report_execption(chatbot, history, a, b):92    """93        向chatbot中添加错误信息94    """95    chatbot.append((a, b))96    history.append(a); history.append(b)97 98def text_divide_paragraph(text):99    """100        将文本按照段落分隔符分割开，生成带有段落标签的HTML代码。101    """102    if '```' in text:103        # careful input104        return text105    else:106        # wtf input107        lines = text.split("\n")108        for i, line in enumerate(lines):109            lines[i] = lines[i].replace(" ", "&nbsp;")110        text = "</br>".join(lines)111        return text112 113def markdown_convertion(txt):114    """115        将Markdown格式的文本转换为HTML格式。如果包含数学公式，则先将公式转换为HTML格式。116    """117    if ('$' in txt) and ('```' not in txt):118        return markdown.markdown(txt,extensions=['fenced_code','tables']) + '<br><br>' + \119            markdown.markdown(convert_math(txt, splitParagraphs=False),extensions=['fenced_code','tables'])120    else:121        return markdown.markdown(txt,extensions=['fenced_code','tables'])122 123 124def format_io(self, y):125    """126        将输入和输出解析为HTML格式。将y中最后一项的输入部分段落化，并将输出部分的Markdown和数学公式转换为HTML格式。127    """128    if y is None or y == []: return []129    i_ask, gpt_reply = y[-1]130    i_ask = text_divide_paragraph(i_ask) # 输入部分太自由，预处理一波131    y[-1] = (132        None if i_ask is None else markdown.markdown(i_ask, extensions=['fenced_code','tables']),133        None if gpt_reply is None else markdown_convertion(gpt_reply)134    )135    return y136 137 138def find_free_port():139    """140        返回当前系统中可用的未使用端口。141    """142    import socket143    from contextlib import closing144    with closing(socket.socket(socket.AF_INET, socket.SOCK_STREAM)) as s:145        s.bind(('', 0))146        s.setsockopt(socket.SOL_SOCKET, socket.SO_REUSEADDR, 1)147        return s.getsockname()[1]148 149 150def extract_archive(file_path, dest_dir):151    import zipfile152    import tarfile153    import os154    # Get the file extension of the input file155    file_extension = os.path.splitext(file_path)[1]156 157    # Extract the archive based on its extension158    if file_extension == '.zip':159        with zipfile.ZipFile(file_path, 'r') as zipobj:160            zipobj.extractall(path=dest_dir)161            print("Successfully extracted zip archive to {}".format(dest_dir))162 163    elif file_extension in ['.tar', '.gz', '.bz2']:164        with tarfile.open(file_path, 'r:*') as tarobj:165            tarobj.extractall(path=dest_dir)166            print("Successfully extracted tar archive to {}".format(dest_dir))167    else:168        return169 170def find_recent_files(directory):171    """172        me: find files that is created with in one minutes under a directory with python, write a function173        gpt: here it is!174    """175    import os176    import time177    current_time = time.time()178    one_minute_ago = current_time - 60179    recent_files = []180 181    for filename in os.listdir(directory):182        file_path = os.path.join(directory, filename)183        if file_path.endswith('.log'): continue184        created_time = os.path.getctime(file_path)185        if created_time >= one_minute_ago:186            if os.path.isdir(file_path): continue187            recent_files.append(file_path)188 189    return recent_files190 191 192def on_file_uploaded(files, chatbot, txt):193    if len(files) == 0: return chatbot, txt194    import shutil, os, time, glob195    from toolbox import extract_archive196    try: shutil.rmtree('./private_upload/')197    except: pass198    time_tag = time.strftime("%Y-%m-%d-%H-%M-%S", time.localtime())199    os.makedirs(f'private_upload/{time_tag}', exist_ok=True)200    for file in files:201        file_origin_name = os.path.basename(file.orig_name)202        shutil.copy(file.name, f'private_upload/{time_tag}/{file_origin_name}')203        extract_archive(f'private_upload/{time_tag}/{file_origin_name}', 204                        dest_dir=f'private_upload/{time_tag}/{file_origin_name}.extract')205    moved_files = [fp for fp in glob.glob('private_upload/**/*', recursive=True)]206    txt = f'private_upload/{time_tag}'207    moved_files_str = '\t\n\n'.join(moved_files)208    chatbot.append(['我上传了文件，请查收', 209                    f'[Local Message] 收到以下文件: \n\n{moved_files_str}\n\n调用路径参数已自动修正到: \n\n{txt}\n\n现在您点击任意实验功能时，以上文件将被作为输入参数'])210    return chatbot, txt211 212 213def on_report_generated(files, chatbot):214    from toolbox import find_recent_files215    report_files = find_recent_files('gpt_log')216    if len(report_files) == 0: return report_files, chatbot217    # files.extend(report_files)218    chatbot.append(['汇总报告如何远程获取？', '汇总报告已经添加到右侧文件上传区，请查收。'])219    return report_files, chatbot220 221def get_conf(*args):222    # 建议您复制一个config_private.py放自己的秘密, 如API和代理网址, 避免不小心传github被别人看到223    res = []224    for arg in args:225        try: r = getattr(importlib.import_module('config_private'), arg)226        except: r = getattr(importlib.import_module('config'), arg)227        res.append(r)228        # 在读取API_KEY时，检查一下是不是忘了改config229        if arg=='API_KEY' and len(r) != 51:230            assert False, "正确的API_KEY密钥是51位，请在config文件中修改API密钥, 添加海外代理之后再运行。" + \231                        "（如果您刚更新过代码，请确保旧版config_private文件中没有遗留任何新增键值）"232    return res233 234def clear_line_break(txt):235    txt = txt.replace('\n', ' ')236    txt = txt.replace('  ', ' ')237    txt = txt.replace('  ', ' ')238    return txt