diff --git a/README.md b/README.md index 533d3838..fdfd10d0 100644 --- a/README.md +++ b/README.md @@ -38,6 +38,7 @@ **交流群**: 766798517 一群人较多，建议加下面的（开发和建议相关讨论）不一定有空回复，会优先写文档和代码 **交流群**: 571780722 另一个群（开发和建议相关讨论）不一定有空回复，会优先写文档和代码 +**交流群**: 1035228475 另一个群（开发和建议相关讨论）不一定有空回复，会优先写文档和代码 ##
@@ -46,11 +47,13 @@ ### 部署方式 -如果你不知道Docker是什么，建议寻找相关教程或使用手动部署 +如果你不知道Docker是什么，建议寻找相关教程或使用手动部署（现在不建议使用docker，更新慢，可能不适配） - [🐳 Docker部署指南](docs/docker_deploy.md) -- [📦 手动部署指南](docs/manual_deploy.md) +- [📦 手动部署指南 Windows](docs/manual_deploy_windows.md) + +- [📦 手动部署指南 Linux](docs/manual_deploy_linux.md) ### 配置说明 - [🎀 新手配置指南](docs/installation_cute.md) - 通俗易懂的配置教程，适合初次使用的猫娘 diff --git a/bot.py b/bot.py index 51979a5e..f2cc9164 100644 --- a/bot.py +++ b/bot.py @@ -1,88 +1,142 @@ import os - +import shutil import nonebot from dotenv import load_dotenv from loguru import logger from nonebot.adapters.onebot.v11 import Adapter -'''彩蛋''' -from colorama import Fore, init +# 获取没有加载env时的环境变量 +env_mask = {key: os.getenv(key) for key in os.environ} -init() -text = "多年以后，面对AI行刑队，张三将会回想起他2023年在会议上讨论人工智能的那个下午" -rainbow_colors = [Fore.RED, Fore.YELLOW, Fore.GREEN, Fore.CYAN, Fore.BLUE, Fore.MAGENTA] -rainbow_text = "" -for i, char in enumerate(text): - rainbow_text += rainbow_colors[i % len(rainbow_colors)] + char -print(rainbow_text) -'''彩蛋''' +def easter_egg(): + # 彩蛋 + from colorama import init, Fore -# 初次启动检测 -if not os.path.exists("config/bot_config.toml"): - logger.warning("检测到bot_config.toml不存在，正在从模板复制") - import shutil - # 检查config目录是否存在 - if not os.path.exists("config"): - os.makedirs("config") - logger.info("创建config目录") + init() + text = "多年以后，面对AI行刑队，张三将会回想起他2023年在会议上讨论人工智能的那个下午" + rainbow_colors = [Fore.RED, Fore.YELLOW, Fore.GREEN, Fore.CYAN, Fore.BLUE, Fore.MAGENTA] + rainbow_text = "" + for i, char in enumerate(text): + rainbow_text += rainbow_colors[i % len(rainbow_colors)] + char + print(rainbow_text) - shutil.copy("template/bot_config_template.toml", "config/bot_config.toml") - logger.info("复制完成，请修改config/bot_config.toml和.env.prod中的配置后重新启动") +def init_config(): + # 初次启动检测 + if not os.path.exists("config/bot_config.toml"): + logger.warning("检测到bot_config.toml不存在，正在从模板复制") + + # 检查config目录是否存在 + if not os.path.exists("config"): + os.makedirs("config") + logger.info("创建config目录") -# 初始化.env 默认ENVIRONMENT=prod -if not os.path.exists(".env"): - with open(".env", "w") as f: - f.write("ENVIRONMENT=prod") + shutil.copy("template/bot_config_template.toml", "config/bot_config.toml") + logger.info("复制完成，请修改config/bot_config.toml和.env.prod中的配置后重新启动") - # 检测.env.prod文件是否存在 - if not os.path.exists(".env.prod"): - logger.error("检测到.env.prod文件不存在") - shutil.copy("template.env", "./.env.prod") +def init_env(): + # 初始化.env 默认ENVIRONMENT=prod + if not os.path.exists(".env"): + with open(".env", "w") as f: + f.write("ENVIRONMENT=prod") -# 首先加载基础环境变量.env -if os.path.exists(".env"): - load_dotenv(".env") - logger.success("成功加载基础环境变量配置") + # 检测.env.prod文件是否存在 + if not os.path.exists(".env.prod"): + logger.error("检测到.env.prod文件不存在") + shutil.copy("template.env", "./.env.prod") -# 根据 ENVIRONMENT 加载对应的环境配置 -if os.getenv("ENVIRONMENT") == "prod": - logger.success("加载生产环境变量配置") - load_dotenv(".env.prod", override=True) # override=True 允许覆盖已存在的环境变量 -elif os.getenv("ENVIRONMENT") == "dev": - logger.success("加载开发环境变量配置") - load_dotenv(".env.dev", override=True) # override=True 允许覆盖已存在的环境变量 -elif os.path.exists(f".env.{os.getenv('ENVIRONMENT')}"): - logger.success(f"加载{os.getenv('ENVIRONMENT')}环境变量配置") - load_dotenv(f".env.{os.getenv('ENVIRONMENT')}", override=True) # override=True 允许覆盖已存在的环境变量 -else: - logger.error(f"ENVIRONMENT配置错误，请检查.env文件中的ENVIRONMENT变量对应的.env.{os.getenv('ENVIRONMENT')}是否存在") - exit(1) + # 首先加载基础环境变量.env + if os.path.exists(".env"): + load_dotenv(".env") + logger.success("成功加载基础环境变量配置") -# 检测Key是否存在 -if not os.getenv("SILICONFLOW_KEY"): - logger.error("缺失必要的API KEY") - logger.error(f"请至少在.env.{os.getenv('ENVIRONMENT')}文件中填写SILICONFLOW_KEY后重新启动") - exit(1) +def load_env(): + # 使用闭包实现对加载器的横向扩展，避免大量重复判断 + def prod(): + logger.success("加载生产环境变量配置") + load_dotenv(".env.prod", override=True) # override=True 允许覆盖已存在的环境变量 -# 获取所有环境变量 -env_config = {key: os.getenv(key) for key in os.environ} + def dev(): + logger.success("加载开发环境变量配置") + load_dotenv(".env.dev", override=True) # override=True 允许覆盖已存在的环境变量 -# 设置基础配置 -base_config = { - "websocket_port": int(env_config.get("PORT", 8080)), - "host": env_config.get("HOST", "127.0.0.1"), - "log_level": "INFO", -} + fn_map = { + "prod": prod, + "dev": dev + } -# 合并配置 -nonebot.init(**base_config, **env_config) + env = os.getenv("ENVIRONMENT") + logger.info(f"[load_env] 当前的 ENVIRONMENT 变量值：{env}") -# 注册适配器 -driver = nonebot.get_driver() -driver.register_adapter(Adapter) + if env in fn_map: + fn_map[env]() # 根据映射执行闭包函数 -# 加载插件 -nonebot.load_plugins("src/plugins") + elif os.path.exists(f".env.{env}"): + logger.success(f"加载{env}环境变量配置") + load_dotenv(f".env.{env}", override=True) # override=True 允许覆盖已存在的环境变量 + + else: + logger.error(f"ENVIRONMENT 配置错误，请检查 .env 文件中的 ENVIRONMENT 变量及对应 .env.{env} 是否存在") + RuntimeError(f"ENVIRONMENT 配置错误，请检查 .env 文件中的 ENVIRONMENT 变量及对应 .env.{env} 是否存在") + + + +def scan_provider(env_config: dict): + provider = {} + + # 利用未初始化 env 时获取的 env_mask 来对新的环境变量集去重 + # 避免 GPG_KEY 这样的变量干扰检查 + env_config = dict(filter(lambda item: item[0] not in env_mask, env_config.items())) + + # 遍历 env_config 的所有键 + for key in env_config: + # 检查键是否符合 {provider}_BASE_URL 或 {provider}_KEY 的格式 + if key.endswith("_BASE_URL") or key.endswith("_KEY"): + # 提取 provider 名称 + provider_name = key.split("_", 1)[0] # 从左分割一次，取第一部分 + + # 初始化 provider 的字典（如果尚未初始化） + if provider_name not in provider: + provider[provider_name] = {"url": None, "key": None} + + # 根据键的类型填充 url 或 key + if key.endswith("_BASE_URL"): + provider[provider_name]["url"] = env_config[key] + elif key.endswith("_KEY"): + provider[provider_name]["key"] = env_config[key] + + # 检查每个 provider 是否同时存在 url 和 key + for provider_name, config in provider.items(): + if config["url"] is None or config["key"] is None: + logger.error( + f"provider 内容：{config}\n" + f"env_config 内容：{env_config}" + ) + raise ValueError(f"请检查 '{provider_name}' 提供商配置是否丢失 BASE_URL 或 KEY 环境变量") if __name__ == "__main__": + easter_egg() + init_config() + init_env() + load_env() + + env_config = {key: os.getenv(key) for key in os.environ} + scan_provider(env_config) + + # 设置基础配置 + base_config = { + "websocket_port": int(env_config.get("PORT", 8080)), + "host": env_config.get("HOST", "127.0.0.1"), + "log_level": "INFO", + } + + # 合并配置 + nonebot.init(**base_config, **env_config) + + # 注册适配器 + driver = nonebot.get_driver() + driver.register_adapter(Adapter) + + # 加载插件 + nonebot.load_plugins("src/plugins") + nonebot.run() diff --git a/docs/manual_deploy_linux.md b/docs/manual_deploy_linux.md new file mode 100644 index 00000000..09b2cfd0 --- /dev/null +++ b/docs/manual_deploy_linux.md @@ -0,0 +1,116 @@ +# 📦 Linux系统如何手动部署MaiMbot麦麦？ + +## 准备工作 +- 一台联网的Linux设备（本教程以Ubuntu/Debian系为例） +- QQ小号（QQ框架的使用可能导致qq被风控，严重（小概率）可能会导致账号封禁，强烈不推荐使用大号） +- 可用的大模型API +- 一个AI助手，网上随便搜一家打开来用都行，可以帮你解决一些不懂的问题 +- 以下内容假设你对Linux系统有一定的了解，如果觉得难以理解，请直接用Windows系统部署[Windows系统部署指南](./manual_deploy_windows.md) + +## 你需要知道什么？ + +- 如何正确向AI助手提问，来学习新知识 + +- Python是什么 + +- Python的虚拟环境是什么？如何创建虚拟环境 + +- 命令行是什么 + +- 数据库是什么？如何安装并启动MongoDB + +- 如何运行一个QQ机器人，以及NapCat框架是什么 +--- + +## 环境配置 + +### 1️⃣ **确认Python版本** + +需确保Python版本为3.9及以上 + +```bash +python --version +# 或 +python3 --version +``` +如果版本低于3.9，请更新Python版本。 +```bash +# Ubuntu/Debian +sudo apt update +sudo apt install python3.9 +# 如执行了这一步，建议在执行时将python3指向python3.9 +# 更新替代方案，设置 python3.9 为默认的 python3 版本: +sudo update-alternatives --install /usr/bin/python3 python3 /usr/bin/python3.9 1 +sudo update-alternatives --config python3 +``` + +### 2️⃣ **创建虚拟环境** +```bash +# 方法1：使用venv(推荐) +python3 -m venv maimbot +source maimbot/bin/activate # 激活环境 + +# 方法2：使用conda（需先安装Miniconda） +wget https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh +bash Miniconda3-latest-Linux-x86_64.sh +conda create -n maimbot python=3.9 +conda activate maimbot + +# 通过以上方法创建并进入虚拟环境后，再执行以下命令 + +# 安装依赖（任选一种环境） +pip install -r requirements.txt +``` + +--- + +## 数据库配置 +### 3️⃣ **安装并启动MongoDB** +- 安装与启动：Debian参考[官方文档](https://docs.mongodb.com/manual/tutorial/install-mongodb-on-debian/)，Ubuntu参考[官方文档](https://docs.mongodb.com/manual/tutorial/install-mongodb-on-ubuntu/) + +- 默认连接本地27017端口 +--- + +## NapCat配置 +### 4️⃣ **安装NapCat框架** + +- 参考[NapCat官方文档](https://www.napcat.wiki/guide/boot/Shell#napcat-installer-linux%E4%B8%80%E9%94%AE%E4%BD%BF%E7%94%A8%E8%84%9A%E6%9C%AC-%E6%94%AF%E6%8C%81ubuntu-20-debian-10-centos9)安装 + +- 使用QQ小号登录，添加反向WS地址： +`ws://localhost:8080/onebot/v11/ws` + +--- + +## 配置文件设置 +### 5️⃣ **配置文件设置，让麦麦Bot正常工作** +- 修改环境配置文件：`.env.prod` +- 修改机器人配置文件：`bot_config.toml` + + +--- + +## 启动机器人 +### 6️⃣ **启动麦麦机器人** +```bash +# 在项目目录下操作 +nb run +# 或 +python3 bot.py +``` + +--- + +## **其他组件(可选)** +- 直接运行 knowledge.py生成知识库 + + +--- + +## 常见问题 +🔧 权限问题：在命令前加`sudo` +🔌 端口占用：使用`sudo lsof -i :8080`查看端口占用 +🛡️ 防火墙：确保8080/27017端口开放 +```bash +sudo ufw allow 8080/tcp +sudo ufw allow 27017/tcp +``` \ No newline at end of file diff --git a/docs/manual_deploy.md b/docs/manual_deploy_windows.md similarity index 98% rename from docs/manual_deploy.md rename to docs/manual_deploy_windows.md index 6d53beb4..bd9c26f8 100644 --- a/docs/manual_deploy.md +++ b/docs/manual_deploy_windows.md @@ -1,4 +1,4 @@ -# 📦 如何手动部署MaiMbot麦麦？ +# 📦 Windows系统如何手动部署MaiMbot麦麦？ ## 你需要什么？ diff --git a/src/plugins/chat/config.py b/src/plugins/chat/config.py index 180cce56..c027753c 100644 --- a/src/plugins/chat/config.py +++ b/src/plugins/chat/config.py @@ -4,11 +4,15 @@ from typing import Dict, Optional import tomli from loguru import logger - +from packaging import version +from packaging.version import Version, InvalidVersion +from packaging.specifiers import SpecifierSet,InvalidSpecifier @dataclass class BotConfig: """机器人配置类""" + INNER_VERSION: Version = None + BOT_QQ: Optional[int] = 1 BOT_NICKNAME: Optional[str] = None @@ -93,12 +97,284 @@ class BotConfig: if not os.path.exists(config_dir): os.makedirs(config_dir) return config_dir + + @classmethod + def convert_to_specifierset(cls, value: str) -> SpecifierSet: + """将 字符串 版本表达式转换成 SpecifierSet + Args: + value[str]: 版本表达式(字符串) + Returns: + SpecifierSet + """ + try: + converted = SpecifierSet(value) + except InvalidSpecifier as e: + logger.error( + f"{value} 分类使用了错误的版本约束表达式\n", + "请阅读 https://semver.org/lang/zh-CN/ 修改代码" + ) + exit(1) + + return converted + + @classmethod + def get_config_version(cls, toml: dict) -> Version: + """提取配置文件的 SpecifierSet 版本数据 + Args: + toml[dict]: 输入的配置文件字典 + Returns: + Version + """ + + if 'inner' in toml: + try: + config_version : str = toml["inner"]["version"] + except KeyError as e: + logger.error(f"配置文件中 inner 段 不存在 {e}, 这是错误的配置文件") + raise KeyError(f"配置文件中 inner 段 不存在 {e}, 这是错误的配置文件") + else: + toml["inner"] = { "version": "0.0.0" } + config_version = toml["inner"]["version"] + + try: + ver = version.parse(config_version) + except InvalidVersion as e: + logger.error( + "配置文件中 inner段 的 version 键是错误的版本描述\n" + "请阅读 https://semver.org/lang/zh-CN/ 修改配置，并参考本项目指定的模板进行修改\n" + "本项目在不同的版本下有不同的模板，请注意识别" + ) + raise InvalidVersion("配置文件中 inner段 的 version 键是错误的版本描述\n") + + return ver @classmethod def load_config(cls, config_path: str = None) -> "BotConfig": """从TOML配置文件加载配置""" config = cls() + + def personality(parent: dict): + personality_config=parent['personality'] + personality=personality_config.get('prompt_personality') + if len(personality) >= 2: + logger.info(f"载入自定义人格:{personality}") + config.PROMPT_PERSONALITY=personality_config.get('prompt_personality',config.PROMPT_PERSONALITY) + logger.info(f"载入自定义日程prompt:{personality_config.get('prompt_schedule',config.PROMPT_SCHEDULE_GEN)}") + config.PROMPT_SCHEDULE_GEN=personality_config.get('prompt_schedule',config.PROMPT_SCHEDULE_GEN) + + if config.INNER_VERSION in SpecifierSet(">=0.0.2"): + config.PERSONALITY_1=personality_config.get('personality_1_probability',config.PERSONALITY_1) + config.PERSONALITY_2=personality_config.get('personality_2_probability',config.PERSONALITY_2) + config.PERSONALITY_3=personality_config.get('personality_3_probability',config.PERSONALITY_3) + + def emoji(parent: dict): + emoji_config = parent["emoji"] + config.EMOJI_CHECK_INTERVAL = emoji_config.get("check_interval", config.EMOJI_CHECK_INTERVAL) + config.EMOJI_REGISTER_INTERVAL = emoji_config.get("register_interval", config.EMOJI_REGISTER_INTERVAL) + config.EMOJI_CHECK_PROMPT = emoji_config.get('check_prompt',config.EMOJI_CHECK_PROMPT) + config.EMOJI_SAVE = emoji_config.get('auto_save',config.EMOJI_SAVE) + config.EMOJI_CHECK = emoji_config.get('enable_check',config.EMOJI_CHECK) + + def cq_code(parent: dict): + cq_code_config = parent["cq_code"] + config.ENABLE_PIC_TRANSLATE = cq_code_config.get("enable_pic_translate", config.ENABLE_PIC_TRANSLATE) + + def bot(parent: dict): + # 机器人基础配置 + bot_config = parent["bot"] + bot_qq = bot_config.get("qq") + config.BOT_QQ = int(bot_qq) + config.BOT_NICKNAME = bot_config.get("nickname", config.BOT_NICKNAME) + + def response(parent: dict): + response_config = parent["response"] + config.MODEL_R1_PROBABILITY = response_config.get("model_r1_probability", config.MODEL_R1_PROBABILITY) + config.MODEL_V3_PROBABILITY = response_config.get("model_v3_probability", config.MODEL_V3_PROBABILITY) + config.MODEL_R1_DISTILL_PROBABILITY = response_config.get("model_r1_distill_probability", config.MODEL_R1_DISTILL_PROBABILITY) + config.max_response_length = response_config.get("max_response_length", config.max_response_length) + + def model(parent: dict): + # 加载模型配置 + model_config:dict = parent["model"] + + config_list = [ + "llm_reasoning", + "llm_reasoning_minor", + "llm_normal", + "llm_normal_minor", + "llm_topic_judge", + "llm_summary_by_topic", + "llm_emotion_judge", + "vlm", + "embedding", + "moderation" + ] + + for item in config_list: + if item in model_config: + cfg_item:dict = model_config[item] + + # base_url 的例子： SILICONFLOW_BASE_URL + # key 的例子： SILICONFLOW_KEY + cfg_target = { + "name" : "", + "base_url" : "", + "key" : "", + "pri_in" : 0, + "pri_out" : 0 + } + + if config.INNER_VERSION in SpecifierSet("<=0.0.0"): + cfg_target = cfg_item + + elif config.INNER_VERSION in SpecifierSet(">=0.0.1"): + stable_item = ["name","pri_in","pri_out"] + pricing_item = ["pri_in","pri_out"] + # 从配置中原始拷贝稳定字段 + for i in stable_item: + # 如果 字段 属于计费项 且获取不到，那默认值是 0 + if i in pricing_item and i not in cfg_item: + cfg_target[i] = 0 + else: + # 没有特殊情况则原样复制 + try: + cfg_target[i] = cfg_item[i] + except KeyError as e: + logger.error(f"{item} 中的必要字段 {e} 不存在，请检查") + raise KeyError(f"{item} 中的必要字段 {e} 不存在，请检查") + + + provider = cfg_item.get("provider") + if provider == None: + logger.error(f"provider 字段在模型配置 {item} 中不存在，请检查") + raise KeyError(f"provider 字段在模型配置 {item} 中不存在，请检查") + + cfg_target["base_url"] = f"{provider}_BASE_URL" + cfg_target["key"] = f"{provider}_KEY" + + + # 如果 列表中的项目在 model_config 中，利用反射来设置对应项目 + setattr(config,item,cfg_target) + else: + logger.error(f"模型 {item} 在config中不存在，请检查") + raise KeyError(f"模型 {item} 在config中不存在，请检查") + + def message(parent: dict): + msg_config = parent["message"] + config.MIN_TEXT_LENGTH = msg_config.get("min_text_length", config.MIN_TEXT_LENGTH) + config.MAX_CONTEXT_SIZE = msg_config.get("max_context_size", config.MAX_CONTEXT_SIZE) + config.emoji_chance = msg_config.get("emoji_chance", config.emoji_chance) + config.ban_words=msg_config.get("ban_words",config.ban_words) + + if config.INNER_VERSION in SpecifierSet(">=0.0.2"): + config.thinking_timeout = msg_config.get("thinking_timeout", config.thinking_timeout) + config.response_willing_amplifier = msg_config.get("response_willing_amplifier", config.response_willing_amplifier) + config.response_interested_rate_amplifier = msg_config.get("response_interested_rate_amplifier", config.response_interested_rate_amplifier) + config.down_frequency_rate = msg_config.get("down_frequency_rate", config.down_frequency_rate) + + def memory(parent: dict): + memory_config = parent["memory"] + config.build_memory_interval = memory_config.get("build_memory_interval", config.build_memory_interval) + config.forget_memory_interval = memory_config.get("forget_memory_interval", config.forget_memory_interval) + + def mood(parent: dict): + mood_config = parent["mood"] + config.mood_update_interval = mood_config.get("mood_update_interval", config.mood_update_interval) + config.mood_decay_rate = mood_config.get("mood_decay_rate", config.mood_decay_rate) + config.mood_intensity_factor = mood_config.get("mood_intensity_factor", config.mood_intensity_factor) + + def keywords_reaction(parent: dict): + keywords_reaction_config = parent["keywords_reaction"] + if keywords_reaction_config.get("enable", False): + config.keywords_reaction_rules = keywords_reaction_config.get("rules", config.keywords_reaction_rules) + + def chinese_typo(parent: dict): + chinese_typo_config = parent["chinese_typo"] + config.chinese_typo_enable = chinese_typo_config.get("enable", config.chinese_typo_enable) + config.chinese_typo_error_rate = chinese_typo_config.get("error_rate", config.chinese_typo_error_rate) + config.chinese_typo_min_freq = chinese_typo_config.get("min_freq", config.chinese_typo_min_freq) + config.chinese_typo_tone_error_rate = chinese_typo_config.get("tone_error_rate", config.chinese_typo_tone_error_rate) + config.chinese_typo_word_replace_rate = chinese_typo_config.get("word_replace_rate", config.chinese_typo_word_replace_rate) + + def groups(parent: dict): + groups_config = parent["groups"] + config.talk_allowed_groups = set(groups_config.get("talk_allowed", [])) + config.talk_frequency_down_groups = set(groups_config.get("talk_frequency_down", [])) + config.ban_user_id = set(groups_config.get("ban_user_id", [])) + + def others(parent: dict): + others_config = parent["others"] + config.enable_advance_output = others_config.get("enable_advance_output", config.enable_advance_output) + config.enable_kuuki_read = others_config.get("enable_kuuki_read", config.enable_kuuki_read) + + # 版本表达式：>=1.0.0,<2.0.0 + # 允许字段：func: method, support: str, notice: str, necessary: bool + # 如果使用 notice 字段，在该组配置加载时，会展示该字段对用户的警示 + # 例如："notice": "personality 将在 1.3.2 后被移除"，那么在有效版本中的用户就会虽然可以 + # 正常执行程序，但是会看到这条自定义提示 + include_configs = { + "personality": { + "func": personality, + "support": ">=0.0.0" + }, + "emoji": { + "func": emoji, + "support": ">=0.0.0" + }, + "cq_code": { + "func": cq_code, + "support": ">=0.0.0" + }, + "bot": { + "func": bot, + "support": ">=0.0.0" + }, + "response": { + "func": response, + "support": ">=0.0.0" + }, + "model": { + "func": model, + "support": ">=0.0.0" + }, + "message": { + "func": message, + "support": ">=0.0.0" + }, + "memory": { + "func": memory, + "support": ">=0.0.0" + }, + "mood": { + "func": mood, + "support": ">=0.0.0" + }, + "keywords_reaction": { + "func": keywords_reaction, + "support": ">=0.0.2", + "necessary": False + }, + "chinese_typo": { + "func": chinese_typo, + "support": ">=0.0.3", + "necessary": False + }, + "groups": { + "func": groups, + "support": ">=0.0.0" + }, + "others": { + "func": others, + "support": ">=0.0.0" + } + } + + # 原地修改，将 字符串版本表达式 转换成 版本对象 + for key in include_configs: + item_support = include_configs[key]["support"] + include_configs[key]["support"] = cls.convert_to_specifierset(item_support) + if os.path.exists(config_path): with open(config_path, "rb") as f: try: @@ -106,145 +382,60 @@ class BotConfig: except(tomli.TOMLDecodeError) as e: logger.critical(f"配置文件bot_config.toml填写有误，请检查第{e.lineno}行第{e.colno}处：{e.msg}") exit(1) - - if 'personality' in toml_dict: - personality_config=toml_dict['personality'] - personality=personality_config.get('prompt_personality') - if len(personality) >= 2: - logger.info(f"载入自定义人格:{personality}") - config.PROMPT_PERSONALITY=personality_config.get('prompt_personality',config.PROMPT_PERSONALITY) - logger.info(f"载入自定义日程prompt:{personality_config.get('prompt_schedule',config.PROMPT_SCHEDULE_GEN)}") - config.PROMPT_SCHEDULE_GEN=personality_config.get('prompt_schedule',config.PROMPT_SCHEDULE_GEN) - config.PERSONALITY_1=personality_config.get('personality_1_probability',config.PERSONALITY_1) - config.PERSONALITY_2=personality_config.get('personality_2_probability',config.PERSONALITY_2) - config.PERSONALITY_3=personality_config.get('personality_3_probability',config.PERSONALITY_3) + + # 获取配置文件版本 + config.INNER_VERSION = cls.get_config_version(toml_dict) - if "emoji" in toml_dict: - emoji_config = toml_dict["emoji"] - config.EMOJI_CHECK_INTERVAL = emoji_config.get("check_interval", config.EMOJI_CHECK_INTERVAL) - config.EMOJI_REGISTER_INTERVAL = emoji_config.get("register_interval", config.EMOJI_REGISTER_INTERVAL) - config.EMOJI_CHECK_PROMPT = emoji_config.get('check_prompt',config.EMOJI_CHECK_PROMPT) - config.EMOJI_SAVE = emoji_config.get('auto_save',config.EMOJI_SAVE) - config.EMOJI_CHECK = emoji_config.get('enable_check',config.EMOJI_CHECK) - - if "cq_code" in toml_dict: - cq_code_config = toml_dict["cq_code"] - config.ENABLE_PIC_TRANSLATE = cq_code_config.get("enable_pic_translate", config.ENABLE_PIC_TRANSLATE) - - # 机器人基础配置 - if "bot" in toml_dict: - bot_config = toml_dict["bot"] - bot_qq = bot_config.get("qq") - config.BOT_QQ = int(bot_qq) - config.BOT_NICKNAME = bot_config.get("nickname", config.BOT_NICKNAME) - - if "response" in toml_dict: - response_config = toml_dict["response"] - config.MODEL_R1_PROBABILITY = response_config.get("model_r1_probability", config.MODEL_R1_PROBABILITY) - config.MODEL_V3_PROBABILITY = response_config.get("model_v3_probability", config.MODEL_V3_PROBABILITY) - config.MODEL_R1_DISTILL_PROBABILITY = response_config.get("model_r1_distill_probability", config.MODEL_R1_DISTILL_PROBABILITY) - config.max_response_length = response_config.get("max_response_length", config.max_response_length) - - # 加载模型配置 - if "model" in toml_dict: - model_config = toml_dict["model"] - - if "llm_reasoning" in model_config: - config.llm_reasoning = model_config["llm_reasoning"] - - if "llm_reasoning_minor" in model_config: - config.llm_reasoning_minor = model_config["llm_reasoning_minor"] - - if "llm_normal" in model_config: - config.llm_normal = model_config["llm_normal"] - - if "llm_normal_minor" in model_config: - config.llm_normal_minor = model_config["llm_normal_minor"] + # 如果在配置中找到了需要的项，调用对应项的闭包函数处理 + for key in include_configs: + if key in toml_dict: + group_specifierset: SpecifierSet = include_configs[key]["support"] + + # 检查配置文件版本是否在支持范围内 + if config.INNER_VERSION in group_specifierset: + # 如果版本在支持范围内，检查是否存在通知 + if 'notice' in include_configs[key]: + logger.warning(include_configs[key]["notice"]) + + include_configs[key]["func"](toml_dict) + + else: + # 如果版本不在支持范围内，崩溃并提示用户 + logger.error( + f"配置文件中的 '{key}' 字段的版本 ({config.INNER_VERSION}) 不在支持范围内。\n" + f"当前程序仅支持以下版本范围: {group_specifierset}" + ) + raise InvalidVersion(f"当前程序仅支持以下版本范围: {group_specifierset}") - if "llm_topic_judge" in model_config: - config.llm_topic_judge = model_config["llm_topic_judge"] - - if "llm_summary_by_topic" in model_config: - config.llm_summary_by_topic = model_config["llm_summary_by_topic"] - - if "llm_emotion_judge" in model_config: - config.llm_emotion_judge = model_config["llm_emotion_judge"] - - if "vlm" in model_config: - config.vlm = model_config["vlm"] + # 如果 necessary 项目存在，而且显式声明是 False，进入特殊处理 + elif "necessary" in include_configs[key] and include_configs[key].get("necessary") == False: + # 通过 pass 处理的项虽然直接忽略也是可以的，但是为了不增加理解困难，依然需要在这里显式处理 + if key == "keywords_reaction": + pass - if "embedding" in model_config: - config.embedding = model_config["embedding"] - - if "moderation" in model_config: - config.moderation = model_config["moderation"] - - # 消息配置 - if "message" in toml_dict: - msg_config = toml_dict["message"] - config.MIN_TEXT_LENGTH = msg_config.get("min_text_length", config.MIN_TEXT_LENGTH) - config.MAX_CONTEXT_SIZE = msg_config.get("max_context_size", config.MAX_CONTEXT_SIZE) - config.emoji_chance = msg_config.get("emoji_chance", config.emoji_chance) - config.ban_words=msg_config.get("ban_words",config.ban_words) - config.thinking_timeout = msg_config.get("thinking_timeout", config.thinking_timeout) - config.response_willing_amplifier = msg_config.get("response_willing_amplifier", config.response_willing_amplifier) - config.response_interested_rate_amplifier = msg_config.get("response_interested_rate_amplifier", config.response_interested_rate_amplifier) - config.down_frequency_rate = msg_config.get("down_frequency_rate", config.down_frequency_rate) + else: + # 如果用户根本没有需要的配置项，提示缺少配置 + logger.error(f"配置文件中缺少必需的字段: '{key}'") + raise KeyError(f"配置文件中缺少必需的字段: '{key}'") - if "memory" in toml_dict: - memory_config = toml_dict["memory"] - config.build_memory_interval = memory_config.get("build_memory_interval", config.build_memory_interval) - config.forget_memory_interval = memory_config.get("forget_memory_interval", config.forget_memory_interval) - - if "mood" in toml_dict: - mood_config = toml_dict["mood"] - config.mood_update_interval = mood_config.get("mood_update_interval", config.mood_update_interval) - config.mood_decay_rate = mood_config.get("mood_decay_rate", config.mood_decay_rate) - config.mood_intensity_factor = mood_config.get("mood_intensity_factor", config.mood_intensity_factor) - - # print(toml_dict) - if "keywords_reaction" in toml_dict: - # 读取关键词回复配置 - keywords_reaction_config = toml_dict["keywords_reaction"] - if keywords_reaction_config.get("enable", False): - config.keywords_reaction_rules = keywords_reaction_config.get("rules", config.keywords_reaction_rules) - - if "chinese_typo_generator" in toml_dict: - # 读取中文错别字生成器配置 - chinese_typo_generator_config = toml_dict["chinese_typo_generator"] - config.chinese_typo_enable = chinese_typo_generator_config.get("enable", config.chinese_typo_enable) - config.chinese_typo_error_rate = chinese_typo_generator_config.get("error_rate", config.chinese_typo_error_rate) - config.chinese_typo_min_freq = chinese_typo_generator_config.get("min_freq", config.chinese_typo_min_freq) - config.chinese_typo_tone_error_rate = chinese_typo_generator_config.get("tone_error_rate", config.chinese_typo_tone_error_rate) - config.chinese_typo_word_replace_rate = chinese_typo_generator_config.get("word_replace_rate", config.chinese_typo_word_replace_rate) - - # 群组配置 - if "groups" in toml_dict: - groups_config = toml_dict["groups"] - config.talk_allowed_groups = set(groups_config.get("talk_allowed", [])) - config.talk_frequency_down_groups = set(groups_config.get("talk_frequency_down", [])) - config.ban_user_id = set(groups_config.get("ban_user_id", [])) - - if "others" in toml_dict: - others_config = toml_dict["others"] - config.enable_advance_output = others_config.get("enable_advance_output", config.enable_advance_output) - config.enable_kuuki_read = others_config.get("enable_kuuki_read", config.enable_kuuki_read) - - logger.success(f"成功加载配置文件: {config_path}") + logger.success(f"成功加载配置文件: {config_path}") return config # 获取配置文件路径 - bot_config_floder_path = BotConfig.get_config_dir() print(f"正在品鉴配置文件目录: {bot_config_floder_path}") + bot_config_path = os.path.join(bot_config_floder_path, "bot_config.toml") + if os.path.exists(bot_config_path): # 如果开发环境配置文件不存在，则使用默认配置文件 print(f"异常的新鲜，异常的美味: {bot_config_path}") logger.info("使用bot配置文件") else: - logger.info("没有找到美味") + # 配置文件不存在 + logger.error("配置文件不存在，请检查路径: {bot_config_path}") + raise FileNotFoundError(f"配置文件不存在: {bot_config_path}") global_config = BotConfig.load_config(config_path=bot_config_path) diff --git a/src/plugins/chat/utils.py b/src/plugins/chat/utils.py index e0e06118..d166bcd2 100644 --- a/src/plugins/chat/utils.py +++ b/src/plugins/chat/utils.py @@ -12,6 +12,7 @@ from ..models.utils_model import LLM_request from ..utils.typo_generator import ChineseTypoGenerator from .config import global_config from .message import Message +from ..moods.moods import MoodManager driver = get_driver() config = driver.config @@ -326,43 +327,68 @@ def random_remove_punctuation(text: str) -> str: def process_llm_response(text: str) -> List[str]: # processed_response = process_text_with_typos(content) - if len(text) > 300: + if len(text) > 200: print(f"回复过长 ({len(text)} 字符)，返回默认回复") return ['懒得说'] # 处理长消息 - if global_config.chinese_typo_enable: - typo_generator = ChineseTypoGenerator( - error_rate=global_config.chinese_typo_error_rate, - min_freq=global_config.chinese_typo_min_freq, - tone_error_rate=global_config.chinese_typo_tone_error_rate, - word_replace_rate=global_config.chinese_typo_word_replace_rate - ) - typoed_text = typo_generator.create_typo_sentence(text)[0] - else: - typoed_text = text - sentences = split_into_sentences_w_remove_punctuation(typoed_text) + typo_generator = ChineseTypoGenerator( + error_rate=global_config.chinese_typo_error_rate, + min_freq=global_config.chinese_typo_min_freq, + tone_error_rate=global_config.chinese_typo_tone_error_rate, + word_replace_rate=global_config.chinese_typo_word_replace_rate + ) + split_sentences = split_into_sentences_w_remove_punctuation(text) + sentences = [] + for sentence in split_sentences: + if global_config.chinese_typo_enable: + typoed_text, typo_corrections = typo_generator.create_typo_sentence(sentence) + sentences.append(typoed_text) + if typo_corrections: + sentences.append(typo_corrections) + else: + sentences.append(sentence) # 检查分割后的消息数量是否过多（超过3条） - if len(sentences) > 4: + + if len(sentences) > 5: print(f"分割后消息数量过多 ({len(sentences)} 条)，返回默认回复") return [f'{global_config.BOT_NICKNAME}不知道哦'] return sentences -def calculate_typing_time(input_string: str, chinese_time: float = 0.2, english_time: float = 0.1) -> float: +def calculate_typing_time(input_string: str, chinese_time: float = 0.4, english_time: float = 0.2) -> float: """ 计算输入字符串所需的时间，中文和英文字符有不同的输入时间 input_string (str): 输入的字符串 - chinese_time (float): 中文字符的输入时间，默认为0.3秒 - english_time (float): 英文字符的输入时间，默认为0.15秒 + chinese_time (float): 中文字符的输入时间，默认为0.2秒 + english_time (float): 英文字符的输入时间，默认为0.1秒 + + 特殊情况： + - 如果只有一个中文字符，将使用3倍的中文输入时间 + - 在所有输入结束后，额外加上回车时间0.3秒 """ + mood_manager = MoodManager.get_instance() + # 将0-1的唤醒度映射到-1到1 + mood_arousal = mood_manager.current_mood.arousal + # 映射到0.5到2倍的速度系数 + typing_speed_multiplier = 1.5 ** mood_arousal # 唤醒度为1时速度翻倍,为-1时速度减半 + chinese_time *= 1/typing_speed_multiplier + english_time *= 1/typing_speed_multiplier + # 计算中文字符数 + chinese_chars = sum(1 for char in input_string if '\u4e00' <= char <= '\u9fff') + + # 如果只有一个中文字符，使用3倍时间 + if chinese_chars == 1 and len(input_string.strip()) == 1: + return chinese_time * 3 + 0.3 # 加上回车时间 + + # 正常计算所有字符的输入时间 total_time = 0.0 for char in input_string: if '\u4e00' <= char <= '\u9fff': # 判断是否为中文字符 total_time += chinese_time else: # 其他字符（如英文） total_time += english_time - return total_time + return total_time + 0.3 # 加上回车时间 def cosine_similarity(v1, v2): diff --git a/src/plugins/models/utils_model.py b/src/plugins/models/utils_model.py index c70c26ff..e890b4c8 100644 --- a/src/plugins/models/utils_model.py +++ b/src/plugins/models/utils_model.py @@ -23,6 +23,7 @@ class LLM_request: self.api_key = getattr(config, model["key"]) self.base_url = getattr(config, model["base_url"]) except AttributeError as e: + logger.error(f"原始 model dict 信息：{model}") logger.error(f"配置错误：找不到对应的配置项 - {str(e)}") raise ValueError(f"配置错误：找不到对应的配置项 - {str(e)}") from e self.model_name = model["name"] @@ -181,6 +182,13 @@ class LLM_request: continue elif response.status in policy["abort_codes"]: logger.error(f"错误码: {response.status} - {error_code_mapping.get(response.status)}") + if response.status == 403 : + if global_config.llm_normal == "Pro/deepseek-ai/DeepSeek-V3": + logger.error("可能是没有给硅基流动充钱，普通模型自动退化至非Pro模型，反应速度可能会变慢") + global_config.llm_normal = "deepseek-ai/DeepSeek-V3" + if global_config.llm_reasoning == "Pro/deepseek-ai/DeepSeek-R1": + logger.error("可能是没有给硅基流动充钱，推理模型自动退化至非Pro模型，反应速度可能会变慢") + global_config.llm_reasoning = "deepseek-ai/DeepSeek-R1" raise RuntimeError(f"请求被拒绝: {error_code_mapping.get(response.status)}") response.raise_for_status() diff --git a/src/plugins/utils/typo_generator.py b/src/plugins/utils/typo_generator.py index c743ec6e..aa72c387 100644 --- a/src/plugins/utils/typo_generator.py +++ b/src/plugins/utils/typo_generator.py @@ -284,10 +284,13 @@ class ChineseTypoGenerator: 返回: typo_sentence: 包含错别字的句子 - typo_info: 错别字信息列表 + correction_suggestion: 随机选择的一个纠正建议，返回正确的字/词 """ result = [] typo_info = [] + word_typos = [] # 记录词语错误对(错词,正确词) + char_typos = [] # 记录单字错误对(错字,正确字) + current_pos = 0 # 分词 words = self._segment_sentence(sentence) @@ -296,6 +299,7 @@ class ChineseTypoGenerator: # 如果是标点符号或空格，直接添加 if all(not self._is_chinese_char(c) for c in word): result.append(word) + current_pos += len(word) continue # 获取词语的拼音 @@ -316,6 +320,8 @@ class ChineseTypoGenerator: ' '.join(word_pinyin), ' '.join(self._get_word_pinyin(typo_word)), orig_freq, typo_freq)) + word_typos.append((typo_word, word)) # 记录(错词,正确词)对 + current_pos += len(typo_word) continue # 如果不进行整词替换，则进行单字替换 @@ -333,11 +339,15 @@ class ChineseTypoGenerator: result.append(typo_char) typo_py = pinyin(typo_char, style=Style.TONE3)[0][0] typo_info.append((char, typo_char, py, typo_py, orig_freq, typo_freq)) + char_typos.append((typo_char, char)) # 记录(错字,正确字)对 + current_pos += 1 continue result.append(char) + current_pos += 1 else: # 处理多字词的单字替换 word_result = [] + word_start_pos = current_pos for i, (char, py) in enumerate(zip(word, word_pinyin)): # 词中的字替换概率降低 word_error_rate = self.error_rate * (0.7 ** (len(word) - 1)) @@ -353,11 +363,24 @@ class ChineseTypoGenerator: word_result.append(typo_char) typo_py = pinyin(typo_char, style=Style.TONE3)[0][0] typo_info.append((char, typo_char, py, typo_py, orig_freq, typo_freq)) + char_typos.append((typo_char, char)) # 记录(错字,正确字)对 continue word_result.append(char) result.append(''.join(word_result)) + current_pos += len(word) - return ''.join(result), typo_info + # 优先从词语错误中选择，如果没有则从单字错误中选择 + correction_suggestion = None + # 50%概率返回纠正建议 + if random.random() < 0.5: + if word_typos: + wrong_word, correct_word = random.choice(word_typos) + correction_suggestion = correct_word + elif char_typos: + wrong_char, correct_char = random.choice(char_typos) + correction_suggestion = correct_char + + return ''.join(result), correction_suggestion def format_typo_info(self, typo_info): """ @@ -419,16 +442,16 @@ def main(): # 创建包含错别字的句子 start_time = time.time() - typo_sentence, typo_info = typo_generator.create_typo_sentence(sentence) + typo_sentence, correction_suggestion = typo_generator.create_typo_sentence(sentence) # 打印结果 print("\n原句：", sentence) print("错字版：", typo_sentence) - # 打印错别字信息 - if typo_info: - print("\n错别字信息：") - print(typo_generator.format_typo_info(typo_info)) + # 打印纠正建议 + if correction_suggestion: + print("\n随机纠正建议：") + print(f"应该改为：{correction_suggestion}") # 计算并打印总耗时 end_time = time.time() diff --git a/template/bot_config_template.toml b/template/bot_config_template.toml index 9709ebbd..d7c66d3f 100644 --- a/template/bot_config_template.toml +++ b/template/bot_config_template.toml @@ -1,3 +1,6 @@ +[inner] +version = "0.0.3" + [bot] qq = 123 nickname = "麦麦" @@ -100,49 +103,42 @@ ban_user_id = [] #禁止回复消息的QQ号 [model.llm_reasoning] #回复模型1 主要回复模型 name = "Pro/deepseek-ai/DeepSeek-R1" -base_url = "SILICONFLOW_BASE_URL" -key = "SILICONFLOW_KEY" +provider = "SILICONFLOW" pri_in = 0 #模型的输入价格（非必填，可以记录消耗） pri_out = 0 #模型的输出价格（非必填，可以记录消耗） + [model.llm_reasoning_minor] #回复模型3 次要回复模型 name = "deepseek-ai/DeepSeek-R1-Distill-Qwen-32B" -base_url = "SILICONFLOW_BASE_URL" -key = "SILICONFLOW_KEY" +provider = "SILICONFLOW" #非推理模型 [model.llm_normal] #V3 回复模型2 次要回复模型 name = "Pro/deepseek-ai/DeepSeek-V3" -base_url = "SILICONFLOW_BASE_URL" -key = "SILICONFLOW_KEY" +provider = "SILICONFLOW" [model.llm_normal_minor] #V2.5 name = "deepseek-ai/DeepSeek-V2.5" -base_url = "SILICONFLOW_BASE_URL" -key = "SILICONFLOW_KEY" +provider = "SILICONFLOW" [model.llm_emotion_judge] #主题判断 0.7/m name = "Qwen/Qwen2.5-14B-Instruct" -base_url = "SILICONFLOW_BASE_URL" -key = "SILICONFLOW_KEY" +provider = "SILICONFLOW" [model.llm_topic_judge] #主题判断：建议使用qwen2.5 7b name = "Pro/Qwen/Qwen2.5-7B-Instruct" -base_url = "SILICONFLOW_BASE_URL" -key = "SILICONFLOW_KEY" +provider = "SILICONFLOW" [model.llm_summary_by_topic] #建议使用qwen2.5 32b 及以上 name = "Qwen/Qwen2.5-32B-Instruct" -base_url = "SILICONFLOW_BASE_URL" -key = "SILICONFLOW_KEY" +provider = "SILICONFLOW" pri_in = 0 pri_out = 0 [model.moderation] #内容审核 未启用 name = "" -base_url = "SILICONFLOW_BASE_URL" -key = "SILICONFLOW_KEY" +provider = "SILICONFLOW" pri_in = 0 pri_out = 0 @@ -150,8 +146,7 @@ pri_out = 0 [model.vlm] #图像识别 0.35/m name = "Pro/Qwen/Qwen2-VL-7B-Instruct" -base_url = "SILICONFLOW_BASE_URL" -key = "SILICONFLOW_KEY" +provider = "SILICONFLOW" @@ -159,5 +154,4 @@ key = "SILICONFLOW_KEY" [model.embedding] #嵌入 name = "BAAI/bge-m3" -base_url = "SILICONFLOW_BASE_URL" -key = "SILICONFLOW_KEY" +provider = "SILICONFLOW"