Merge branch 'refactor' into msg_buffer
This commit is contained in:
meng_xi_pan
4
.gitignore
vendored
4
.gitignore
vendored
@@ -6,6 +6,10 @@ log/
|
||||
logs/
|
||||
/test
|
||||
/src/test
|
||||
nonebot-maibot-adapter/
|
||||
*.zip
|
||||
run.bat
|
||||
run.py
|
||||
message_queue_content.txt
|
||||
message_queue_content.bat
|
||||
message_queue_window.bat
|
||||
|
||||
47
README.md
47
README.md
@@ -1,4 +1,4 @@
|
||||
# 麦麦！MaiMBot-MaiCore (编辑中)
|
||||
# 麦麦！MaiCore-MaiMBot (编辑中)
|
||||
|
||||
<div align="center">
|
||||
|
||||
@@ -13,10 +13,12 @@
|
||||
**🍔MaiCore是一个基于大语言模型的可交互智能体**
|
||||
|
||||
- LLM 提供对话能力
|
||||
- 动态Prompt构建器
|
||||
- 实时的思维系统
|
||||
- MongoDB 提供数据持久化支持
|
||||
- 可扩展，可支持多种平台和多种功能
|
||||
|
||||
**最新版本: v0.6.0** ([查看更新日志](changelog.md))
|
||||
**最新版本: v0.6.0** ([查看更新日志](changelogs/changelog.md))
|
||||
> [!WARNING]
|
||||
> 次版本MaiBot将基于MaiCore运行，不再依赖于nonebot相关组件运行。
|
||||
> MaiBot将通过nonebot的插件与nonebot建立联系，然后nonebot与QQ建立联系，实现MaiBot与QQ的交互
|
||||
@@ -58,46 +60,57 @@
|
||||
### 最新版本部署教程(MaiCore版本)
|
||||
- [🚀 最新版本部署教程](https://docs.mai-mai.org/manual/deployment/refactor_deploy.html) - 基于MaiCore的新版本部署方式（与旧版本不兼容）
|
||||
|
||||
|
||||
## 🎯 功能介绍
|
||||
|
||||
### 💬 聊天功能
|
||||
|
||||
- 提供思维流（心流）聊天和推理聊天两种对话逻辑
|
||||
- 支持关键词检索主动发言：对消息的话题topic进行识别，如果检测到麦麦存储过的话题就会主动进行发言
|
||||
- 支持bot名字呼唤发言：检测到"麦麦"会主动发言，可配置
|
||||
- 支持多模型，多厂商自定义配置
|
||||
- 动态的prompt构建器，更拟人
|
||||
- 支持图片，转发消息，回复消息的识别
|
||||
- 支持私聊功能，包括消息处理和回复
|
||||
- 支持私聊功能，可使用PFC模式的有目的多轮对话（实验性）
|
||||
|
||||
### 🧠 思维流系统（实验性功能）
|
||||
- 思维流能够生成实时想法，增加回复的拟人性
|
||||
### 🧠 思维流系统
|
||||
- 思维流能够在回复前后进行思考，生成实时想法
|
||||
- 思维流自动启停机制，提升资源利用效率
|
||||
- 思维流与日程系统联动，实现动态日程生成
|
||||
|
||||
### 🧠 记忆系统
|
||||
### 🧠 记忆系统 2.0
|
||||
- 优化记忆抽取策略和prompt结构
|
||||
- 改进海马体记忆提取机制，提升自然度
|
||||
- 对聊天记录进行概括存储，在需要时调用
|
||||
|
||||
### 😊 表情包功能
|
||||
### 😊 表情包系统
|
||||
- 支持根据发言内容发送对应情绪的表情包
|
||||
- 支持识别和处理gif表情包
|
||||
- 会自动偷群友的表情包
|
||||
- 表情包审查功能
|
||||
- 表情包文件完整性自动检查
|
||||
- 自动清理缓存图片
|
||||
|
||||
### 📅 日程功能
|
||||
- 麦麦会自动生成一天的日程，实现更拟人的回复
|
||||
- 支持动态日程生成
|
||||
- 优化日程文本解析功能
|
||||
### 📅 日程系统
|
||||
- 动态更新的日程生成
|
||||
- 可自定义想象力程度
|
||||
- 与聊天情况交互（思维流模式下）
|
||||
|
||||
### 👥 关系系统
|
||||
- 针对每个用户创建"关系"，可以对不同用户进行个性化回复
|
||||
### 👥 关系系统 2.0
|
||||
- 优化关系管理系统，适用于新版本
|
||||
- 提供更丰富的关系接口
|
||||
- 针对每个用户创建"关系"，实现个性化回复
|
||||
|
||||
### 📊 统计系统
|
||||
- 详细统计系统
|
||||
- LLM使用统计
|
||||
- 详细的使用数据统计
|
||||
- LLM调用统计
|
||||
- 在控制台显示统计信息
|
||||
|
||||
### 🔧 系统功能
|
||||
- 支持优雅的shutdown机制
|
||||
- 自动保存功能，定期保存聊天记录和关系数据
|
||||
- 完善的异常处理机制
|
||||
- 可自定义时区设置
|
||||
- 优化的日志输出格式
|
||||
- 配置自动更新功能
|
||||
|
||||
## 开发计划TODO：LIST
|
||||
|
||||
|
||||
@@ -1,26 +1,24 @@
|
||||
# Changelog
|
||||
|
||||
## [0.6.0] - 2025-3-30
|
||||
## [0.6.0] - 2025-4-4
|
||||
|
||||
### 摘要
|
||||
- MaiBot 0.6.0 重磅升级！ 核心重构为独立智能体MaiCore，新增思维流对话系统，支持拟真思考过程。记忆与关系系统2.0让交互更自然，动态日程引擎实现智能调整。优化部署流程，修复30+稳定性问题，隐私政策同步更新，推荐所有用户升级体验全新AI交互！（V3激烈生成）
|
||||
|
||||
### 🌟 核心功能增强
|
||||
#### 架构重构
|
||||
- 将MaiBot重构为MaiCore独立智能体
|
||||
- 移除NoneBot相关代码，改为插件方式与NoneBot对接
|
||||
- 精简代码结构，优化文件夹组织
|
||||
- 新增详细统计系统
|
||||
|
||||
#### 思维流系统
|
||||
- 新增思维流作为实验功能
|
||||
- 思维流大核+小核架构
|
||||
- 思维流回复意愿模式
|
||||
- 优化思维流自动启停机制，提升资源利用效率
|
||||
- 提供两种聊天逻辑，思维流（心流）聊天（ThinkFlowChat）和推理聊天（ReasoningChat）
|
||||
- 思维流聊天能够在回复前后进行思考
|
||||
- 思维流自动启停机制，提升资源利用效率
|
||||
- 思维流与日程系统联动，实现动态日程生成
|
||||
- 优化心流运行逻辑和思考时间计算
|
||||
- 添加错误检测机制
|
||||
- 修复心流无法观察群消息的问题
|
||||
|
||||
#### 回复系统
|
||||
- 优化回复逻辑，添加回复前思考机制
|
||||
- 移除推理模型在回复中的使用
|
||||
- 更改了回复引用的逻辑，从基于时间改为基于新消息
|
||||
- 提供私聊的PFC模式，可以进行有目的，自由多轮对话（实验性）
|
||||
|
||||
#### 记忆系统优化
|
||||
- 优化记忆抽取策略
|
||||
@@ -28,41 +26,33 @@
|
||||
- 改进海马体记忆提取机制，提升自然度
|
||||
|
||||
#### 关系系统优化
|
||||
- 修复relationship_value类型错误
|
||||
- 优化关系管理系统
|
||||
- 改进关系值计算方式
|
||||
- 优化关系管理系统，适用于新版本
|
||||
- 改进关系值计算方式，提供更丰富的关系接口
|
||||
|
||||
#### 表情包系统
|
||||
- 可以识别gif表情包
|
||||
- 表情包增加存储上限
|
||||
- 自动清理缓存图片
|
||||
|
||||
## 日程系统优化
|
||||
- 日程现在动态更新
|
||||
- 日程可以自定义想象力程度
|
||||
- 日程会与聊天情况交互（思维流模式下）
|
||||
|
||||
### 💻 系统架构优化
|
||||
#### 配置系统改进
|
||||
- 优化配置文件整理
|
||||
- 新增分割器功能
|
||||
- 新增表情惩罚系数自定义
|
||||
- 新增更多项目的配置项
|
||||
- 修复配置文件保存问题
|
||||
- 优化配置项管理
|
||||
- 新增配置项：
|
||||
- `schedule`: 日程表生成功能配置
|
||||
- `response_spliter`: 回复分割控制
|
||||
- `experimental`: 实验性功能开关
|
||||
- `llm_observation`和`llm_sub_heartflow`: 思维流模型配置
|
||||
- `llm_heartflow`: 思维流核心模型配置
|
||||
- `prompt_schedule_gen`: 日程生成提示词配置
|
||||
- `memory_ban_words`: 记忆过滤词配置
|
||||
- 优化配置结构：
|
||||
- 调整模型配置组织结构
|
||||
- 优化配置项默认值
|
||||
- 调整配置项顺序
|
||||
- 移除冗余配置
|
||||
|
||||
#### WebUI改进
|
||||
- 新增回复意愿模式选择功能
|
||||
- 优化WebUI界面
|
||||
- 优化WebUI配置保存机制
|
||||
|
||||
#### 部署支持扩展
|
||||
- 优化Docker构建流程
|
||||
- 完善Windows脚本支持
|
||||
- 优化Linux一键安装脚本
|
||||
- 新增macOS教程支持
|
||||
|
||||
### 🐛 问题修复
|
||||
#### 功能稳定性
|
||||
@@ -75,43 +65,24 @@
|
||||
- 修复自定义API提供商识别问题
|
||||
- 修复人格设置保存问题
|
||||
- 修复EULA和隐私政策编码问题
|
||||
- 修复cfg变量引用问题
|
||||
|
||||
#### 性能优化
|
||||
- 提高topic提取效率
|
||||
- 优化logger输出格式
|
||||
- 优化cmd清理功能
|
||||
- 改进LLM使用统计
|
||||
- 优化记忆处理效率
|
||||
|
||||
### 📚 文档更新
|
||||
- 更新README.md内容
|
||||
- 添加macOS部署教程
|
||||
- 优化文档结构
|
||||
- 更新EULA和隐私政策
|
||||
- 完善部署文档
|
||||
|
||||
### 🔧 其他改进
|
||||
- 新增神秘小测验功能
|
||||
- 新增人格测评模型
|
||||
- 新增详细统计系统
|
||||
- 优化表情包审查功能
|
||||
- 改进消息转发处理
|
||||
- 优化代码风格和格式
|
||||
- 完善异常处理机制
|
||||
- 可以自定义时区
|
||||
- 优化日志输出格式
|
||||
- 版本硬编码，新增配置自动更新功能
|
||||
- 更新日程生成器功能
|
||||
- 优化了统计信息，会在控制台显示统计信息
|
||||
|
||||
### 主要改进方向
|
||||
1. 完善思维流系统功能
|
||||
2. 优化记忆系统效率
|
||||
3. 改进关系系统稳定性
|
||||
4. 提升配置系统可用性
|
||||
5. 加强WebUI功能
|
||||
6. 完善部署文档
|
||||
|
||||
|
||||
|
||||
## [0.5.15] - 2025-3-17
|
||||
### 🌟 核心功能增强
|
||||
|
||||
@@ -1,8 +1,18 @@
|
||||
这里放置了测试版本的细节更新
|
||||
## [test-0.6.0-snapshot-9] - 2025-4-4
|
||||
- 可以识别gif表情包
|
||||
|
||||
## [test-0.6.0-snapshot-8] - 2025-4-3
|
||||
- 修复了表情包的注册，获取和发送逻辑
|
||||
- 表情包增加存储上限
|
||||
- 更改了回复引用的逻辑，从基于时间改为基于新消息
|
||||
- 增加了调试信息
|
||||
- 自动清理缓存图片
|
||||
- 修复并重启了关系系统
|
||||
|
||||
## [test-0.6.0-snapshot-7] - 2025-4-2
|
||||
- 修改版本号命名：test-前缀为测试版，无前缀为正式版
|
||||
- 提供私聊的PFC模式
|
||||
- 提供私聊的PFC模式，可以进行有目的，自由多轮对话
|
||||
|
||||
## [0.6.0-mmc-4] - 2025-4-1
|
||||
- 提供两种聊天逻辑，思维流聊天（ThinkFlowChat 和 推理聊天（ReasoningChat）
|
||||
|
||||
@@ -8,9 +8,10 @@ services:
|
||||
ports:
|
||||
- "18002:18002"
|
||||
volumes:
|
||||
- ./docker-config/adapters/plugins:/adapters/src/plugins # 持久化adapters
|
||||
- ./docker-config/adapters/config.py:/adapters/src/plugins/nonebot_plugin_maibot_adapters/config.py # 持久化adapters配置文件
|
||||
- ./docker-config/adapters/.env:/adapters/.env # 持久化adapters配置文件
|
||||
- ./data/qq:/app/.config/QQ # 持久化QQ本体并同步qq表情和图片到adapters
|
||||
- ./data/MaiMBot:/adapters/data
|
||||
restart: always
|
||||
depends_on:
|
||||
- mongodb
|
||||
@@ -61,7 +62,7 @@ services:
|
||||
volumes:
|
||||
- ./docker-config/napcat:/app/napcat/config # 持久化napcat配置文件
|
||||
- ./data/qq:/app/.config/QQ # 持久化QQ本体并同步qq表情和图片到adapters
|
||||
- ./data/MaiMBot:/MaiMBot/data # NapCat 和 NoneBot 共享此卷，否则发送图片会有问题
|
||||
- ./data/MaiMBot:/adapters/data # NapCat 和 NoneBot 共享此卷，否则发送图片会有问题
|
||||
container_name: maim-bot-napcat
|
||||
restart: always
|
||||
image: mlikiowa/napcat-docker:latest
|
||||
|
||||
BIN
requirements.txt
BIN
requirements.txt
Binary file not shown.
@@ -192,7 +192,7 @@ class SubHeartflow:
|
||||
logger.info(f"麦麦的思考前脑内状态：{self.current_mind}")
|
||||
|
||||
async def do_thinking_after_reply(self, reply_content, chat_talking_prompt):
|
||||
print("麦麦回复之后脑袋转起来了")
|
||||
# print("麦麦回复之后脑袋转起来了")
|
||||
current_thinking_info = self.current_mind
|
||||
mood_info = self.current_state.mood
|
||||
|
||||
|
||||
@@ -107,8 +107,8 @@ class MainSystem:
|
||||
self.forget_memory_task(),
|
||||
self.print_mood_task(),
|
||||
self.remove_recalled_message_task(),
|
||||
emoji_manager.start_periodic_check(),
|
||||
emoji_manager.start_periodic_register(),
|
||||
emoji_manager.start_periodic_check_register(),
|
||||
# emoji_manager.start_periodic_register(),
|
||||
self.app.run(),
|
||||
]
|
||||
await asyncio.gather(*tasks)
|
||||
|
||||
@@ -75,25 +75,48 @@ class ChatBot:
|
||||
- 表情包处理
|
||||
- 性能计时
|
||||
"""
|
||||
try:
|
||||
message = MessageRecv(message_data)
|
||||
groupinfo = message.message_info.group_info
|
||||
logger.debug(f"处理消息:{str(message_data)[:50]}...")
|
||||
|
||||
message = MessageRecv(message_data)
|
||||
groupinfo = message.message_info.group_info
|
||||
|
||||
if global_config.enable_pfc_chatting:
|
||||
try:
|
||||
if global_config.enable_pfc_chatting:
|
||||
try:
|
||||
if groupinfo is None and global_config.enable_friend_chat:
|
||||
userinfo = message.message_info.user_info
|
||||
messageinfo = message.message_info
|
||||
# 创建聊天流
|
||||
chat = await chat_manager.get_or_create_stream(
|
||||
platform=messageinfo.platform,
|
||||
user_info=userinfo,
|
||||
group_info=groupinfo,
|
||||
)
|
||||
message.update_chat_stream(chat)
|
||||
await self.only_process_chat.process_message(message)
|
||||
await self._create_PFC_chat(message)
|
||||
else:
|
||||
if groupinfo.group_id in global_config.talk_allowed_groups:
|
||||
logger.debug(f"开始群聊模式{message_data}")
|
||||
if global_config.response_mode == "heart_flow":
|
||||
await self.think_flow_chat.process_message(message_data)
|
||||
elif global_config.response_mode == "reasoning":
|
||||
logger.debug(f"开始推理模式{message_data}")
|
||||
await self.reasoning_chat.process_message(message_data)
|
||||
else:
|
||||
logger.error(f"未知的回复模式，请检查配置文件！！: {global_config.response_mode}")
|
||||
except Exception as e:
|
||||
logger.error(f"处理PFC消息失败: {e}")
|
||||
else:
|
||||
if groupinfo is None and global_config.enable_friend_chat:
|
||||
userinfo = message.message_info.user_info
|
||||
messageinfo = message.message_info
|
||||
# 创建聊天流
|
||||
chat = await chat_manager.get_or_create_stream(
|
||||
platform=messageinfo.platform,
|
||||
user_info=userinfo,
|
||||
group_info=groupinfo,
|
||||
)
|
||||
message.update_chat_stream(chat)
|
||||
await self.only_process_chat.process_message(message)
|
||||
await self._create_PFC_chat(message)
|
||||
else:
|
||||
# 私聊处理流程
|
||||
# await self._handle_private_chat(message)
|
||||
if global_config.response_mode == "heart_flow":
|
||||
await self.think_flow_chat.process_message(message_data)
|
||||
elif global_config.response_mode == "reasoning":
|
||||
await self.reasoning_chat.process_message(message_data)
|
||||
else:
|
||||
logger.error(f"未知的回复模式，请检查配置文件！！: {global_config.response_mode}")
|
||||
else: # 群聊处理
|
||||
if groupinfo.group_id in global_config.talk_allowed_groups:
|
||||
if global_config.response_mode == "heart_flow":
|
||||
await self.think_flow_chat.process_message(message_data)
|
||||
@@ -101,26 +124,8 @@ class ChatBot:
|
||||
await self.reasoning_chat.process_message(message_data)
|
||||
else:
|
||||
logger.error(f"未知的回复模式，请检查配置文件！！: {global_config.response_mode}")
|
||||
except Exception as e:
|
||||
logger.error(f"处理PFC消息失败: {e}")
|
||||
else:
|
||||
if groupinfo is None and global_config.enable_friend_chat:
|
||||
# 私聊处理流程
|
||||
# await self._handle_private_chat(message)
|
||||
if global_config.response_mode == "heart_flow":
|
||||
await self.think_flow_chat.process_message(message_data)
|
||||
elif global_config.response_mode == "reasoning":
|
||||
await self.reasoning_chat.process_message(message_data)
|
||||
else:
|
||||
logger.error(f"未知的回复模式，请检查配置文件！！: {global_config.response_mode}")
|
||||
else: # 群聊处理
|
||||
if groupinfo.group_id in global_config.talk_allowed_groups:
|
||||
if global_config.response_mode == "heart_flow":
|
||||
await self.think_flow_chat.process_message(message_data)
|
||||
elif global_config.response_mode == "reasoning":
|
||||
await self.reasoning_chat.process_message(message_data)
|
||||
else:
|
||||
logger.error(f"未知的回复模式，请检查配置文件！！: {global_config.response_mode}")
|
||||
except Exception as e:
|
||||
logger.error(f"预处理消息失败: {e}")
|
||||
|
||||
|
||||
# 创建全局ChatBot实例
|
||||
|
||||
@@ -39,12 +39,28 @@ class EmojiManager:
|
||||
model=global_config.llm_emotion_judge, max_tokens=600, temperature=0.8, request_type="emoji"
|
||||
) # 更高的温度，更少的token（后续可以根据情绪来调整温度）
|
||||
|
||||
self.emoji_num = 0
|
||||
self.emoji_num_max = global_config.max_emoji_num
|
||||
self.emoji_num_max_reach_deletion = global_config.max_reach_deletion
|
||||
|
||||
logger.info("启动表情包管理器")
|
||||
|
||||
def _ensure_emoji_dir(self):
|
||||
"""确保表情存储目录存在"""
|
||||
os.makedirs(self.EMOJI_DIR, exist_ok=True)
|
||||
|
||||
def _update_emoji_count(self):
|
||||
"""更新表情包数量统计
|
||||
|
||||
检查数据库中的表情包数量并更新到 self.emoji_num
|
||||
"""
|
||||
try:
|
||||
self._ensure_db()
|
||||
self.emoji_num = db.emoji.count_documents({})
|
||||
logger.info(f"[统计] 当前表情包数量: {self.emoji_num}")
|
||||
except Exception as e:
|
||||
logger.error(f"[错误] 更新表情包数量失败: {str(e)}")
|
||||
|
||||
def initialize(self):
|
||||
"""初始化数据库连接和表情目录"""
|
||||
if not self._initialized:
|
||||
@@ -52,6 +68,8 @@ class EmojiManager:
|
||||
self._ensure_emoji_collection()
|
||||
self._ensure_emoji_dir()
|
||||
self._initialized = True
|
||||
# 更新表情包数量
|
||||
self._update_emoji_count()
|
||||
# 启动时执行一次完整性检查
|
||||
self.check_emoji_file_integrity()
|
||||
except Exception:
|
||||
@@ -340,12 +358,6 @@ class EmojiManager:
|
||||
except Exception:
|
||||
logger.exception("[错误] 扫描表情包失败")
|
||||
|
||||
async def start_periodic_register(self):
|
||||
"""定期扫描新表情包"""
|
||||
while True:
|
||||
logger.info("[扫描] 开始扫描新表情包...")
|
||||
await self.scan_new_emojis()
|
||||
await asyncio.sleep(global_config.EMOJI_CHECK_INTERVAL * 60)
|
||||
|
||||
def check_emoji_file_integrity(self):
|
||||
"""检查表情包文件完整性
|
||||
@@ -418,12 +430,136 @@ class EmojiManager:
|
||||
logger.error(f"[错误] 检查表情包完整性失败: {str(e)}")
|
||||
logger.error(traceback.format_exc())
|
||||
|
||||
async def start_periodic_check(self):
|
||||
def check_emoji_file_full(self):
|
||||
"""检查表情包文件是否完整，如果数量超出限制且允许删除，则删除多余的表情包
|
||||
|
||||
删除规则：
|
||||
1. 优先删除创建时间更早的表情包
|
||||
2. 优先删除使用次数少的表情包，但使用次数多的也有小概率被删除
|
||||
"""
|
||||
try:
|
||||
self._ensure_db()
|
||||
# 更新表情包数量
|
||||
self._update_emoji_count()
|
||||
|
||||
# 检查是否超出限制
|
||||
if self.emoji_num <= self.emoji_num_max:
|
||||
return
|
||||
|
||||
# 如果超出限制但不允许删除，则只记录警告
|
||||
if not global_config.max_reach_deletion:
|
||||
logger.warning(f"[警告] 表情包数量({self.emoji_num})超出限制({self.emoji_num_max})，但未开启自动删除")
|
||||
return
|
||||
|
||||
# 计算需要删除的数量
|
||||
delete_count = self.emoji_num - self.emoji_num_max
|
||||
logger.info(f"[清理] 需要删除 {delete_count} 个表情包")
|
||||
|
||||
# 获取所有表情包，按时间戳升序（旧的在前）排序
|
||||
all_emojis = list(db.emoji.find().sort([("timestamp", 1)]))
|
||||
|
||||
# 计算权重：使用次数越多，被删除的概率越小
|
||||
weights = []
|
||||
max_usage = max((emoji.get("usage_count", 0) for emoji in all_emojis), default=1)
|
||||
for emoji in all_emojis:
|
||||
usage_count = emoji.get("usage_count", 0)
|
||||
# 使用指数衰减函数计算权重，使用次数越多权重越小
|
||||
weight = 1.0 / (1.0 + usage_count / max(1, max_usage))
|
||||
weights.append(weight)
|
||||
|
||||
# 根据权重随机选择要删除的表情包
|
||||
to_delete = []
|
||||
remaining_indices = list(range(len(all_emojis)))
|
||||
|
||||
while len(to_delete) < delete_count and remaining_indices:
|
||||
# 计算当前剩余表情包的权重
|
||||
current_weights = [weights[i] for i in remaining_indices]
|
||||
# 归一化权重
|
||||
total_weight = sum(current_weights)
|
||||
if total_weight == 0:
|
||||
break
|
||||
normalized_weights = [w/total_weight for w in current_weights]
|
||||
|
||||
# 随机选择一个表情包
|
||||
selected_idx = random.choices(remaining_indices, weights=normalized_weights, k=1)[0]
|
||||
to_delete.append(all_emojis[selected_idx])
|
||||
remaining_indices.remove(selected_idx)
|
||||
|
||||
# 删除选中的表情包
|
||||
deleted_count = 0
|
||||
for emoji in to_delete:
|
||||
try:
|
||||
# 删除文件
|
||||
if "path" in emoji and os.path.exists(emoji["path"]):
|
||||
os.remove(emoji["path"])
|
||||
logger.info(f"[删除] 文件: {emoji['path']} (使用次数: {emoji.get('usage_count', 0)})")
|
||||
|
||||
# 删除数据库记录
|
||||
db.emoji.delete_one({"_id": emoji["_id"]})
|
||||
deleted_count += 1
|
||||
|
||||
# 同时从images集合中删除
|
||||
if "hash" in emoji:
|
||||
db.images.delete_one({"hash": emoji["hash"]})
|
||||
|
||||
except Exception as e:
|
||||
logger.error(f"[错误] 删除表情包失败: {str(e)}")
|
||||
continue
|
||||
|
||||
# 更新表情包数量
|
||||
self._update_emoji_count()
|
||||
logger.success(f"[清理] 已删除 {deleted_count} 个表情包，当前数量: {self.emoji_num}")
|
||||
|
||||
except Exception as e:
|
||||
logger.error(f"[错误] 检查表情包数量失败: {str(e)}")
|
||||
|
||||
async def start_periodic_check_register(self):
|
||||
"""定期检查表情包完整性和数量"""
|
||||
while True:
|
||||
logger.info("[扫描] 开始检查表情包完整性...")
|
||||
self.check_emoji_file_integrity()
|
||||
logger.info("[扫描] 开始删除所有图片缓存...")
|
||||
await self.delete_all_images()
|
||||
logger.info("[扫描] 开始扫描新表情包...")
|
||||
if self.emoji_num < self.emoji_num_max:
|
||||
await self.scan_new_emojis()
|
||||
if (self.emoji_num > self.emoji_num_max):
|
||||
logger.warning(f"[警告] 表情包数量超过最大限制: {self.emoji_num} > {self.emoji_num_max},跳过注册")
|
||||
if not global_config.max_reach_deletion:
|
||||
logger.warning("表情包数量超过最大限制，终止注册")
|
||||
break
|
||||
else:
|
||||
logger.warning("表情包数量超过最大限制，开始删除表情包")
|
||||
self.check_emoji_file_full()
|
||||
await asyncio.sleep(global_config.EMOJI_CHECK_INTERVAL * 60)
|
||||
|
||||
async def delete_all_images(self):
|
||||
"""删除 data/image 目录下的所有文件"""
|
||||
try:
|
||||
image_dir = os.path.join("data", "image")
|
||||
if not os.path.exists(image_dir):
|
||||
logger.warning(f"[警告] 目录不存在: {image_dir}")
|
||||
return
|
||||
|
||||
deleted_count = 0
|
||||
failed_count = 0
|
||||
|
||||
# 遍历目录下的所有文件
|
||||
for filename in os.listdir(image_dir):
|
||||
file_path = os.path.join(image_dir, filename)
|
||||
try:
|
||||
if os.path.isfile(file_path):
|
||||
os.remove(file_path)
|
||||
deleted_count += 1
|
||||
logger.debug(f"[删除] 文件: {file_path}")
|
||||
except Exception as e:
|
||||
failed_count += 1
|
||||
logger.error(f"[错误] 删除文件失败 {file_path}: {str(e)}")
|
||||
|
||||
logger.success(f"[清理] 已删除 {deleted_count} 个文件，失败 {failed_count} 个")
|
||||
|
||||
except Exception as e:
|
||||
logger.error(f"[错误] 删除图片目录失败: {str(e)}")
|
||||
|
||||
# 创建全局单例
|
||||
|
||||
emoji_manager = EmojiManager()
|
||||
|
||||
@@ -31,7 +31,7 @@ class Message(MessageBase):
|
||||
def __init__(
|
||||
self,
|
||||
message_id: str,
|
||||
time: int,
|
||||
time: float,
|
||||
chat_stream: ChatStream,
|
||||
user_info: UserInfo,
|
||||
message_segment: Optional[Seg] = None,
|
||||
|
||||
@@ -9,7 +9,7 @@ from .message import MessageSending, MessageThinking, MessageSet
|
||||
|
||||
from ..storage.storage import MessageStorage
|
||||
from ..config.config import global_config
|
||||
from .utils import truncate_message, calculate_typing_time
|
||||
from .utils import truncate_message, calculate_typing_time, count_messages_between
|
||||
|
||||
from src.common.logger import LogConfig, SENDER_STYLE_CONFIG
|
||||
|
||||
@@ -69,9 +69,14 @@ class Message_Sender:
|
||||
if end_point:
|
||||
# logger.info(f"发送消息到{end_point}")
|
||||
# logger.info(message_json)
|
||||
await global_api.send_message(end_point, message_json)
|
||||
await global_api.send_message_REST(end_point, message_json)
|
||||
else:
|
||||
raise ValueError(f"未找到平台：{message.message_info.platform} 的url配置，请检查配置文件")
|
||||
try:
|
||||
await global_api.send_message(message)
|
||||
except Exception as e:
|
||||
raise ValueError(
|
||||
f"未找到平台：{message.message_info.platform} 的url配置，请检查配置文件"
|
||||
) from e
|
||||
logger.success(f"发送消息“{message_preview}”成功")
|
||||
except Exception as e:
|
||||
logger.error(f"发送消息“{message_preview}”失败: {str(e)}")
|
||||
@@ -85,16 +90,16 @@ class MessageContainer:
|
||||
self.max_size = max_size
|
||||
self.messages = []
|
||||
self.last_send_time = 0
|
||||
self.thinking_timeout = 10 # 思考等待超时时间（秒）
|
||||
self.thinking_wait_timeout = 20 # 思考等待超时时间（秒）
|
||||
|
||||
def get_timeout_messages(self) -> List[MessageSending]:
|
||||
"""获取所有超时的Message_Sending对象（思考时间超过30秒），按thinking_start_time排序"""
|
||||
"""获取所有超时的Message_Sending对象（思考时间超过20秒），按thinking_start_time排序"""
|
||||
current_time = time.time()
|
||||
timeout_messages = []
|
||||
|
||||
for msg in self.messages:
|
||||
if isinstance(msg, MessageSending):
|
||||
if current_time - msg.thinking_start_time > self.thinking_timeout:
|
||||
if current_time - msg.thinking_start_time > self.thinking_wait_timeout:
|
||||
timeout_messages.append(msg)
|
||||
|
||||
# 按thinking_start_time排序，时间早的在前面
|
||||
@@ -172,6 +177,7 @@ class MessageManager:
|
||||
message_earliest = container.get_earliest_message()
|
||||
|
||||
if isinstance(message_earliest, MessageThinking):
|
||||
"""取得了思考消息"""
|
||||
message_earliest.update_thinking_time()
|
||||
thinking_time = message_earliest.thinking_time
|
||||
# print(thinking_time)
|
||||
@@ -187,14 +193,20 @@ class MessageManager:
|
||||
container.remove_message(message_earliest)
|
||||
|
||||
else:
|
||||
# print(message_earliest.is_head)
|
||||
# print(message_earliest.update_thinking_time())
|
||||
# print(message_earliest.is_private_message())
|
||||
"""取得了发送消息"""
|
||||
thinking_time = message_earliest.update_thinking_time()
|
||||
print(thinking_time)
|
||||
thinking_start_time = message_earliest.thinking_start_time
|
||||
now_time = time.time()
|
||||
thinking_messages_count, thinking_messages_length = count_messages_between(
|
||||
start_time=thinking_start_time, end_time=now_time, stream_id=message_earliest.chat_stream.stream_id
|
||||
)
|
||||
# print(thinking_time)
|
||||
# print(thinking_messages_count)
|
||||
# print(thinking_messages_length)
|
||||
|
||||
if (
|
||||
message_earliest.is_head
|
||||
and message_earliest.update_thinking_time() > 18
|
||||
and (thinking_messages_count > 4 or thinking_messages_length > 250)
|
||||
and not message_earliest.is_private_message() # 避免在私聊时插入reply
|
||||
):
|
||||
logger.debug(f"设置回复消息{message_earliest.processed_plain_text}")
|
||||
@@ -216,12 +228,18 @@ class MessageManager:
|
||||
continue
|
||||
|
||||
try:
|
||||
# print(msg.is_head)
|
||||
print(msg.update_thinking_time())
|
||||
# print(msg.is_private_message())
|
||||
thinking_time = msg.update_thinking_time()
|
||||
thinking_start_time = msg.thinking_start_time
|
||||
now_time = time.time()
|
||||
thinking_messages_count, thinking_messages_length = count_messages_between(
|
||||
start_time=thinking_start_time, end_time=now_time, stream_id=msg.chat_stream.stream_id
|
||||
)
|
||||
# print(thinking_time)
|
||||
# print(thinking_messages_count)
|
||||
# print(thinking_messages_length)
|
||||
if (
|
||||
msg.is_head
|
||||
and msg.update_thinking_time() > 18
|
||||
and (thinking_messages_count > 4 or thinking_messages_length > 250)
|
||||
and not msg.is_private_message() # 避免在私聊时插入reply
|
||||
):
|
||||
logger.debug(f"设置回复消息{msg.processed_plain_text}")
|
||||
|
||||
@@ -487,3 +487,108 @@ def is_western_char(char):
|
||||
def is_western_paragraph(paragraph):
|
||||
"""检测是否为西文字符段落"""
|
||||
return all(is_western_char(char) for char in paragraph if char.isalnum())
|
||||
|
||||
|
||||
def count_messages_between(start_time: float, end_time: float, stream_id: str) -> tuple[int, int]:
|
||||
"""计算两个时间点之间的消息数量和文本总长度
|
||||
|
||||
Args:
|
||||
start_time (float): 起始时间戳
|
||||
end_time (float): 结束时间戳
|
||||
stream_id (str): 聊天流ID
|
||||
|
||||
Returns:
|
||||
tuple[int, int]: (消息数量, 文本总长度)
|
||||
- 消息数量：包含起始时间的消息，不包含结束时间的消息
|
||||
- 文本总长度：所有消息的processed_plain_text长度之和
|
||||
"""
|
||||
try:
|
||||
# 获取开始时间之前最新的一条消息
|
||||
start_message = db.messages.find_one(
|
||||
{
|
||||
"chat_id": stream_id,
|
||||
"time": {"$lte": start_time}
|
||||
},
|
||||
sort=[("time", -1), ("_id", -1)] # 按时间倒序，_id倒序（最后插入的在前）
|
||||
)
|
||||
|
||||
# 获取结束时间最近的一条消息
|
||||
# 先找到结束时间点的所有消息
|
||||
end_time_messages = list(db.messages.find(
|
||||
{
|
||||
"chat_id": stream_id,
|
||||
"time": {"$lte": end_time}
|
||||
},
|
||||
sort=[("time", -1)] # 先按时间倒序
|
||||
).limit(10)) # 限制查询数量，避免性能问题
|
||||
|
||||
if not end_time_messages:
|
||||
logger.warning(f"未找到结束时间 {end_time} 之前的消息")
|
||||
return 0, 0
|
||||
|
||||
# 找到最大时间
|
||||
max_time = end_time_messages[0]["time"]
|
||||
# 在最大时间的消息中找最后插入的（_id最大的）
|
||||
end_message = max(
|
||||
[msg for msg in end_time_messages if msg["time"] == max_time],
|
||||
key=lambda x: x["_id"]
|
||||
)
|
||||
|
||||
if not start_message:
|
||||
logger.warning(f"未找到开始时间 {start_time} 之前的消息")
|
||||
return 0, 0
|
||||
|
||||
# 调试输出
|
||||
# print("\n=== 消息范围信息 ===")
|
||||
# print("Start message:", {
|
||||
# "message_id": start_message.get("message_id"),
|
||||
# "time": start_message.get("time"),
|
||||
# "text": start_message.get("processed_plain_text", ""),
|
||||
# "_id": str(start_message.get("_id"))
|
||||
# })
|
||||
# print("End message:", {
|
||||
# "message_id": end_message.get("message_id"),
|
||||
# "time": end_message.get("time"),
|
||||
# "text": end_message.get("processed_plain_text", ""),
|
||||
# "_id": str(end_message.get("_id"))
|
||||
# })
|
||||
# print("Stream ID:", stream_id)
|
||||
|
||||
# 如果结束消息的时间等于开始时间，返回0
|
||||
if end_message["time"] == start_message["time"]:
|
||||
return 0, 0
|
||||
|
||||
# 获取并打印这个时间范围内的所有消息
|
||||
# print("\n=== 时间范围内的所有消息 ===")
|
||||
all_messages = list(db.messages.find(
|
||||
{
|
||||
"chat_id": stream_id,
|
||||
"time": {
|
||||
"$gte": start_message["time"],
|
||||
"$lte": end_message["time"]
|
||||
}
|
||||
},
|
||||
sort=[("time", 1), ("_id", 1)] # 按时间正序，_id正序
|
||||
))
|
||||
|
||||
count = 0
|
||||
total_length = 0
|
||||
for msg in all_messages:
|
||||
count += 1
|
||||
text_length = len(msg.get("processed_plain_text", ""))
|
||||
total_length += text_length
|
||||
# print(f"\n消息 {count}:")
|
||||
# print({
|
||||
# "message_id": msg.get("message_id"),
|
||||
# "time": msg.get("time"),
|
||||
# "text": msg.get("processed_plain_text", ""),
|
||||
# "text_length": text_length,
|
||||
# "_id": str(msg.get("_id"))
|
||||
# })
|
||||
|
||||
# 如果时间不同，需要把end_message本身也计入
|
||||
return count - 1, total_length
|
||||
|
||||
except Exception as e:
|
||||
logger.error(f"计算消息数量时出错: {str(e)}")
|
||||
return 0, 0
|
||||
|
||||
@@ -112,8 +112,13 @@ class ImageManager:
|
||||
return f"[表情包：{cached_description}]"
|
||||
|
||||
# 调用AI获取描述
|
||||
prompt = "这是一个表情包，使用中文简洁的描述一下表情包的内容和表情包所表达的情感"
|
||||
description, _ = await self._llm.generate_response_for_image(prompt, image_base64, image_format)
|
||||
if image_format == "gif" or image_format == "GIF":
|
||||
image_base64 = self.transform_gif(image_base64)
|
||||
prompt = "这是一个动态图表情包，每一张图代表了动态图的某一帧，黑色背景代表透明，使用中文简洁的描述一下表情包的内容和表达的情感，简短一些"
|
||||
description, _ = await self._llm.generate_response_for_image(prompt, image_base64, "jpg")
|
||||
else:
|
||||
prompt = "这是一个表情包，使用中文简洁的描述一下表情包的内容和表情包所表达的情感"
|
||||
description, _ = await self._llm.generate_response_for_image(prompt, image_base64, image_format)
|
||||
|
||||
cached_description = self._get_description_from_db(image_hash, "emoji")
|
||||
if cached_description:
|
||||
@@ -221,6 +226,72 @@ class ImageManager:
|
||||
logger.error(f"获取图片描述失败: {str(e)}")
|
||||
return "[图片]"
|
||||
|
||||
def transform_gif(self, gif_base64: str) -> str:
|
||||
"""将GIF转换为水平拼接的静态图像
|
||||
|
||||
Args:
|
||||
gif_base64: GIF的base64编码字符串
|
||||
|
||||
Returns:
|
||||
str: 拼接后的JPG图像的base64编码字符串
|
||||
"""
|
||||
try:
|
||||
# 解码base64
|
||||
gif_data = base64.b64decode(gif_base64)
|
||||
gif = Image.open(io.BytesIO(gif_data))
|
||||
|
||||
# 收集所有帧
|
||||
frames = []
|
||||
try:
|
||||
while True:
|
||||
gif.seek(len(frames))
|
||||
frame = gif.convert('RGB')
|
||||
frames.append(frame.copy())
|
||||
except EOFError:
|
||||
pass
|
||||
|
||||
if not frames:
|
||||
raise ValueError("No frames found in GIF")
|
||||
|
||||
# 计算需要抽取的帧的索引
|
||||
total_frames = len(frames)
|
||||
if total_frames <= 15:
|
||||
selected_frames = frames
|
||||
else:
|
||||
# 均匀抽取10帧
|
||||
indices = [int(i * (total_frames - 1) / 14) for i in range(15)]
|
||||
selected_frames = [frames[i] for i in indices]
|
||||
|
||||
# 获取单帧的尺寸
|
||||
frame_width, frame_height = selected_frames[0].size
|
||||
|
||||
# 计算目标尺寸，保持宽高比
|
||||
target_height = 200 # 固定高度
|
||||
target_width = int((target_height / frame_height) * frame_width)
|
||||
|
||||
# 调整所有帧的大小
|
||||
resized_frames = [frame.resize((target_width, target_height), Image.Resampling.LANCZOS)
|
||||
for frame in selected_frames]
|
||||
|
||||
# 创建拼接图像
|
||||
total_width = target_width * len(resized_frames)
|
||||
combined_image = Image.new('RGB', (total_width, target_height))
|
||||
|
||||
# 水平拼接图像
|
||||
for idx, frame in enumerate(resized_frames):
|
||||
combined_image.paste(frame, (idx * target_width, 0))
|
||||
|
||||
# 转换为base64
|
||||
buffer = io.BytesIO()
|
||||
combined_image.save(buffer, format='JPEG', quality=85)
|
||||
result_base64 = base64.b64encode(buffer.getvalue()).decode('utf-8')
|
||||
|
||||
return result_base64
|
||||
|
||||
except Exception as e:
|
||||
logger.error(f"GIF转换失败: {str(e)}")
|
||||
return None
|
||||
|
||||
|
||||
# 创建全局单例
|
||||
image_manager = ImageManager()
|
||||
|
||||
@@ -177,14 +177,17 @@ class ThinkFlowChat:
|
||||
heartflow.create_subheartflow(chat.stream_id)
|
||||
|
||||
await message.process()
|
||||
logger.debug(f"消息处理成功{message.processed_plain_text}")
|
||||
|
||||
# 过滤词/正则表达式过滤
|
||||
if self._check_ban_words(message.processed_plain_text, chat, userinfo) or self._check_ban_regex(
|
||||
message.raw_message, chat, userinfo
|
||||
):
|
||||
return
|
||||
logger.debug(f"过滤词/正则表达式过滤成功{message.processed_plain_text}")
|
||||
|
||||
await self.storage.store_message(message, chat)
|
||||
logger.debug(f"存储成功{message.processed_plain_text}")
|
||||
|
||||
# 记忆激活
|
||||
timer1 = time.time()
|
||||
|
||||
@@ -1,6 +1,7 @@
|
||||
import os
|
||||
from dataclasses import dataclass, field
|
||||
from typing import Dict, List, Optional
|
||||
from dateutil import tz
|
||||
|
||||
import tomli
|
||||
import tomlkit
|
||||
@@ -24,8 +25,8 @@ config_config = LogConfig(
|
||||
logger = get_module_logger("config", config=config_config)
|
||||
|
||||
#考虑到，实际上配置文件中的mai_version是不会自动更新的,所以采用硬编码
|
||||
mai_version_main = "test-0.6.0"
|
||||
mai_version_fix = "snapshot-7"
|
||||
mai_version_main = "0.6.0"
|
||||
mai_version_fix = ""
|
||||
mai_version = f"{mai_version_main}-{mai_version_fix}"
|
||||
|
||||
def update_config():
|
||||
@@ -151,6 +152,7 @@ class BotConfig:
|
||||
PROMPT_SCHEDULE_GEN = "无日程"
|
||||
SCHEDULE_DOING_UPDATE_INTERVAL: int = 300 # 日程表更新间隔 单位秒
|
||||
SCHEDULE_TEMPERATURE: float = 0.5 # 日程表温度，建议0.5-1.0
|
||||
TIME_ZONE: str = "Asia/Shanghai" # 时区
|
||||
|
||||
# message
|
||||
MAX_CONTEXT_SIZE: int = 15 # 上下文最大消息数
|
||||
@@ -182,6 +184,8 @@ class BotConfig:
|
||||
# MODEL_R1_DISTILL_PROBABILITY: float = 0.1 # R1蒸馏模型概率
|
||||
|
||||
# emoji
|
||||
max_emoji_num: int = 200 # 表情包最大数量
|
||||
max_reach_deletion: bool = True # 开启则在达到最大数量时删除表情包，关闭则不会继续收集表情包
|
||||
EMOJI_CHECK_INTERVAL: int = 120 # 表情包检查间隔（分钟）
|
||||
EMOJI_REGISTER_INTERVAL: int = 10 # 表情包注册间隔（分钟）
|
||||
EMOJI_SAVE: bool = True # 偷表情包
|
||||
@@ -353,6 +357,11 @@ class BotConfig:
|
||||
)
|
||||
if config.INNER_VERSION in SpecifierSet(">=1.0.2"):
|
||||
config.SCHEDULE_TEMPERATURE = schedule_config.get("schedule_temperature", config.SCHEDULE_TEMPERATURE)
|
||||
time_zone = schedule_config.get("time_zone", config.TIME_ZONE)
|
||||
if tz.gettz(time_zone) is None:
|
||||
logger.error(f"无效的时区: {time_zone}，使用默认值: {config.TIME_ZONE}")
|
||||
else:
|
||||
config.TIME_ZONE = time_zone
|
||||
|
||||
def emoji(parent: dict):
|
||||
emoji_config = parent["emoji"]
|
||||
@@ -361,6 +370,9 @@ class BotConfig:
|
||||
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 config.INNER_VERSION in SpecifierSet(">=1.1.1"):
|
||||
config.max_emoji_num = emoji_config.get("max_emoji_num", config.max_emoji_num)
|
||||
config.max_reach_deletion = emoji_config.get("max_reach_deletion", config.max_reach_deletion)
|
||||
|
||||
def bot(parent: dict):
|
||||
# 机器人基础配置
|
||||
|
||||
@@ -14,7 +14,6 @@ from src.common.logger import get_module_logger, LogConfig, MEMORY_STYLE_CONFIG
|
||||
from src.plugins.memory_system.sample_distribution import MemoryBuildScheduler # 分布生成器
|
||||
from .memory_config import MemoryConfig
|
||||
|
||||
|
||||
def get_closest_chat_from_db(length: int, timestamp: str):
|
||||
# print(f"获取最接近指定时间戳的聊天记录，长度: {length}, 时间戳: {timestamp}")
|
||||
# print(f"当前时间: {timestamp},转换后时间: {time.strftime('%Y-%m-%d %H:%M:%S', time.localtime(timestamp))}")
|
||||
|
||||
@@ -1,6 +1,7 @@
|
||||
from fastapi import FastAPI, HTTPException
|
||||
from typing import Dict, Any, Callable, List
|
||||
from fastapi import FastAPI, HTTPException, WebSocket, WebSocketDisconnect
|
||||
from typing import Dict, Any, Callable, List, Set
|
||||
from src.common.logger import get_module_logger
|
||||
from src.plugins.message.message_base import MessageBase
|
||||
import aiohttp
|
||||
import asyncio
|
||||
import uvicorn
|
||||
@@ -10,6 +11,212 @@ import traceback
|
||||
logger = get_module_logger("api")
|
||||
|
||||
|
||||
class BaseMessageHandler:
|
||||
"""消息处理基类"""
|
||||
|
||||
def __init__(self):
|
||||
self.message_handlers: List[Callable] = []
|
||||
self.background_tasks = set()
|
||||
|
||||
def register_message_handler(self, handler: Callable):
|
||||
"""注册消息处理函数"""
|
||||
self.message_handlers.append(handler)
|
||||
|
||||
async def process_message(self, message: Dict[str, Any]):
|
||||
"""处理单条消息"""
|
||||
tasks = []
|
||||
for handler in self.message_handlers:
|
||||
try:
|
||||
tasks.append(handler(message))
|
||||
except Exception as e:
|
||||
raise RuntimeError(str(e)) from e
|
||||
if tasks:
|
||||
await asyncio.gather(*tasks, return_exceptions=True)
|
||||
|
||||
async def _handle_message(self, message: Dict[str, Any]):
|
||||
"""后台处理单个消息"""
|
||||
try:
|
||||
await self.process_message(message)
|
||||
except Exception as e:
|
||||
raise RuntimeError(str(e)) from e
|
||||
|
||||
|
||||
class MessageServer(BaseMessageHandler):
|
||||
"""WebSocket服务端"""
|
||||
|
||||
_class_handlers: List[Callable] = [] # 类级别的消息处理器
|
||||
|
||||
def __init__(self, host: str = "0.0.0.0", port: int = 18000, enable_token=False):
|
||||
super().__init__()
|
||||
# 将类级别的处理器添加到实例处理器中
|
||||
self.message_handlers.extend(self._class_handlers)
|
||||
self.app = FastAPI()
|
||||
self.host = host
|
||||
self.port = port
|
||||
self.active_websockets: Set[WebSocket] = set()
|
||||
self.platform_websockets: Dict[str, WebSocket] = {} # 平台到websocket的映射
|
||||
self.valid_tokens: Set[str] = set()
|
||||
self.enable_token = enable_token
|
||||
self._setup_routes()
|
||||
self._running = False
|
||||
|
||||
@classmethod
|
||||
def register_class_handler(cls, handler: Callable):
|
||||
"""注册类级别的消息处理器"""
|
||||
if handler not in cls._class_handlers:
|
||||
cls._class_handlers.append(handler)
|
||||
|
||||
def register_message_handler(self, handler: Callable):
|
||||
"""注册实例级别的消息处理器"""
|
||||
if handler not in self.message_handlers:
|
||||
self.message_handlers.append(handler)
|
||||
|
||||
async def verify_token(self, token: str) -> bool:
|
||||
if not self.enable_token:
|
||||
return True
|
||||
return token in self.valid_tokens
|
||||
|
||||
def add_valid_token(self, token: str):
|
||||
self.valid_tokens.add(token)
|
||||
|
||||
def remove_valid_token(self, token: str):
|
||||
self.valid_tokens.discard(token)
|
||||
|
||||
def _setup_routes(self):
|
||||
@self.app.post("/api/message")
|
||||
async def handle_message(message: Dict[str, Any]):
|
||||
try:
|
||||
# 创建后台任务处理消息
|
||||
asyncio.create_task(self._handle_message(message))
|
||||
return {"status": "success"}
|
||||
except Exception as e:
|
||||
raise HTTPException(status_code=500, detail=str(e)) from e
|
||||
|
||||
@self.app.websocket("/ws")
|
||||
async def websocket_endpoint(websocket: WebSocket):
|
||||
headers = dict(websocket.headers)
|
||||
token = headers.get("authorization")
|
||||
platform = headers.get("platform", "default") # 获取platform标识
|
||||
if self.enable_token:
|
||||
if not token or not await self.verify_token(token):
|
||||
await websocket.close(code=1008, reason="Invalid or missing token")
|
||||
return
|
||||
|
||||
await websocket.accept()
|
||||
self.active_websockets.add(websocket)
|
||||
|
||||
# 添加到platform映射
|
||||
if platform not in self.platform_websockets:
|
||||
self.platform_websockets[platform] = websocket
|
||||
|
||||
try:
|
||||
while True:
|
||||
message = await websocket.receive_json()
|
||||
# print(f"Received message: {message}")
|
||||
asyncio.create_task(self._handle_message(message))
|
||||
except WebSocketDisconnect:
|
||||
self._remove_websocket(websocket, platform)
|
||||
except Exception as e:
|
||||
self._remove_websocket(websocket, platform)
|
||||
raise RuntimeError(str(e)) from e
|
||||
finally:
|
||||
self._remove_websocket(websocket, platform)
|
||||
|
||||
def _remove_websocket(self, websocket: WebSocket, platform: str):
|
||||
"""从所有集合中移除websocket"""
|
||||
if websocket in self.active_websockets:
|
||||
self.active_websockets.remove(websocket)
|
||||
if platform in self.platform_websockets:
|
||||
if self.platform_websockets[platform] == websocket:
|
||||
del self.platform_websockets[platform]
|
||||
|
||||
async def broadcast_message(self, message: Dict[str, Any]):
|
||||
disconnected = set()
|
||||
for websocket in self.active_websockets:
|
||||
try:
|
||||
await websocket.send_json(message)
|
||||
except Exception:
|
||||
disconnected.add(websocket)
|
||||
for websocket in disconnected:
|
||||
self.active_websockets.remove(websocket)
|
||||
|
||||
async def broadcast_to_platform(self, platform: str, message: Dict[str, Any]):
|
||||
"""向指定平台的所有WebSocket客户端广播消息"""
|
||||
if platform not in self.platform_websockets:
|
||||
raise ValueError(f"平台：{platform} 未连接")
|
||||
|
||||
disconnected = set()
|
||||
try:
|
||||
await self.platform_websockets[platform].send_json(message)
|
||||
except Exception:
|
||||
disconnected.add(self.platform_websockets[platform])
|
||||
|
||||
# 清理断开的连接
|
||||
for websocket in disconnected:
|
||||
self._remove_websocket(websocket, platform)
|
||||
|
||||
async def send_message(self, message: MessageBase):
|
||||
await self.broadcast_to_platform(message.message_info.platform, message.to_dict())
|
||||
|
||||
def run_sync(self):
|
||||
"""同步方式运行服务器"""
|
||||
uvicorn.run(self.app, host=self.host, port=self.port)
|
||||
|
||||
async def run(self):
|
||||
"""异步方式运行服务器"""
|
||||
config = uvicorn.Config(self.app, host=self.host, port=self.port, loop="asyncio")
|
||||
self.server = uvicorn.Server(config)
|
||||
try:
|
||||
await self.server.serve()
|
||||
except KeyboardInterrupt as e:
|
||||
await self.stop()
|
||||
raise KeyboardInterrupt from e
|
||||
|
||||
async def start_server(self):
|
||||
"""启动服务器的异步方法"""
|
||||
if not self._running:
|
||||
self._running = True
|
||||
await self.run()
|
||||
|
||||
async def stop(self):
|
||||
"""停止服务器"""
|
||||
# 清理platform映射
|
||||
self.platform_websockets.clear()
|
||||
|
||||
# 取消所有后台任务
|
||||
for task in self.background_tasks:
|
||||
task.cancel()
|
||||
# 等待所有任务完成
|
||||
await asyncio.gather(*self.background_tasks, return_exceptions=True)
|
||||
self.background_tasks.clear()
|
||||
|
||||
# 关闭所有WebSocket连接
|
||||
for websocket in self.active_websockets:
|
||||
await websocket.close()
|
||||
self.active_websockets.clear()
|
||||
|
||||
if hasattr(self, "server"):
|
||||
self._running = False
|
||||
# 正确关闭 uvicorn 服务器
|
||||
self.server.should_exit = True
|
||||
await self.server.shutdown()
|
||||
# 等待服务器完全停止
|
||||
if hasattr(self.server, "started") and self.server.started:
|
||||
await self.server.main_loop()
|
||||
# 清理处理程序
|
||||
self.message_handlers.clear()
|
||||
|
||||
async def send_message_REST(self, url: str, data: Dict[str, Any]) -> Dict[str, Any]:
|
||||
"""发送消息到指定端点"""
|
||||
async with aiohttp.ClientSession() as session:
|
||||
try:
|
||||
async with session.post(url, json=data, headers={"Content-Type": "application/json"}) as response:
|
||||
return await response.json()
|
||||
except Exception:
|
||||
# logger.error(f"发送消息失败: {str(e)}")
|
||||
pass
|
||||
|
||||
|
||||
class BaseMessageAPI:
|
||||
def __init__(self, host: str = "0.0.0.0", port: int = 18000):
|
||||
self.app = FastAPI()
|
||||
@@ -111,4 +318,4 @@ class BaseMessageAPI:
|
||||
loop.close()
|
||||
|
||||
|
||||
global_api = BaseMessageAPI(host=os.environ["HOST"], port=int(os.environ["PORT"]))
|
||||
global_api = MessageServer(host=os.environ["HOST"], port=int(os.environ["PORT"]))
|
||||
|
||||
@@ -166,7 +166,7 @@ class BaseMessageInfo:
|
||||
|
||||
platform: Optional[str] = None
|
||||
message_id: Union[str, int, None] = None
|
||||
time: Optional[int] = None
|
||||
time: Optional[float] = None
|
||||
group_info: Optional[GroupInfo] = None
|
||||
user_info: Optional[UserInfo] = None
|
||||
format_info: Optional[FormatInfo] = None
|
||||
|
||||
@@ -154,7 +154,7 @@ class LLM_request:
|
||||
# 合并重试策略
|
||||
default_retry = {
|
||||
"max_retries": 3,
|
||||
"base_wait": 15,
|
||||
"base_wait": 10,
|
||||
"retry_codes": [429, 413, 500, 503],
|
||||
"abort_codes": [400, 401, 402, 403],
|
||||
}
|
||||
@@ -179,9 +179,6 @@ class LLM_request:
|
||||
# logger.debug(f"{logger_msg}发送请求到URL: {api_url}")
|
||||
# logger.info(f"使用模型: {self.model_name}")
|
||||
|
||||
# 流式输出标志
|
||||
if stream_mode:
|
||||
payload["stream"] = stream_mode
|
||||
|
||||
# 构建请求体
|
||||
if image_base64:
|
||||
@@ -189,6 +186,11 @@ class LLM_request:
|
||||
elif payload is None:
|
||||
payload = await self._build_payload(prompt)
|
||||
|
||||
# 流式输出标志
|
||||
# 先构建payload，再添加流式输出标志
|
||||
if stream_mode:
|
||||
payload["stream"] = stream_mode
|
||||
|
||||
for retry in range(policy["max_retries"]):
|
||||
try:
|
||||
# 使用上下文管理器处理会话
|
||||
@@ -203,21 +205,21 @@ class LLM_request:
|
||||
# 处理需要重试的状态码
|
||||
if response.status in policy["retry_codes"]:
|
||||
wait_time = policy["base_wait"] * (2**retry)
|
||||
logger.warning(f"错误码: {response.status}, 等待 {wait_time}秒后重试")
|
||||
logger.warning(f"模型 {self.model_name} 错误码: {response.status}, 等待 {wait_time}秒后重试")
|
||||
if response.status == 413:
|
||||
logger.warning("请求体过大，尝试压缩...")
|
||||
image_base64 = compress_base64_image_by_scale(image_base64)
|
||||
payload = await self._build_payload(prompt, image_base64, image_format)
|
||||
elif response.status in [500, 503]:
|
||||
logger.error(f"错误码: {response.status} - {error_code_mapping.get(response.status)}")
|
||||
logger.error(f"模型 {self.model_name} 错误码: {response.status} - {error_code_mapping.get(response.status)}")
|
||||
raise RuntimeError("服务器负载过高，模型恢复失败QAQ")
|
||||
else:
|
||||
logger.warning(f"请求限制(429)，等待{wait_time}秒后重试...")
|
||||
logger.warning(f"模型 {self.model_name} 请求限制(429)，等待{wait_time}秒后重试...")
|
||||
|
||||
await asyncio.sleep(wait_time)
|
||||
continue
|
||||
elif response.status in policy["abort_codes"]:
|
||||
logger.error(f"错误码: {response.status} - {error_code_mapping.get(response.status)}")
|
||||
logger.error(f"模型 {self.model_name} 错误码: {response.status} - {error_code_mapping.get(response.status)}")
|
||||
# 尝试获取并记录服务器返回的详细错误信息
|
||||
try:
|
||||
error_json = await response.json()
|
||||
@@ -319,9 +321,9 @@ class LLM_request:
|
||||
flag_delta_content_finished = True
|
||||
|
||||
except Exception as e:
|
||||
logger.exception(f"解析流式输出错误: {str(e)}")
|
||||
logger.exception(f"模型 {self.model_name} 解析流式输出错误: {str(e)}")
|
||||
except GeneratorExit:
|
||||
logger.warning("流式输出被中断，正在清理资源...")
|
||||
logger.warning("模型 {self.model_name} 流式输出被中断，正在清理资源...")
|
||||
# 确保资源被正确清理
|
||||
await response.release()
|
||||
# 返回已经累积的内容
|
||||
@@ -335,7 +337,7 @@ class LLM_request:
|
||||
else self._default_response_handler(result, user_id, request_type, endpoint)
|
||||
)
|
||||
except Exception as e:
|
||||
logger.error(f"处理流式输出时发生错误: {str(e)}")
|
||||
logger.error(f"模型 {self.model_name} 处理流式输出时发生错误: {str(e)}")
|
||||
# 确保在发生错误时也能正确清理资源
|
||||
try:
|
||||
await response.release()
|
||||
@@ -378,21 +380,21 @@ class LLM_request:
|
||||
except (aiohttp.ClientError, asyncio.TimeoutError) as e:
|
||||
if retry < policy["max_retries"] - 1:
|
||||
wait_time = policy["base_wait"] * (2**retry)
|
||||
logger.error(f"网络错误，等待{wait_time}秒后重试... 错误: {str(e)}")
|
||||
logger.error(f"模型 {self.model_name} 网络错误，等待{wait_time}秒后重试... 错误: {str(e)}")
|
||||
await asyncio.sleep(wait_time)
|
||||
continue
|
||||
else:
|
||||
logger.critical(f"网络错误达到最大重试次数: {str(e)}")
|
||||
logger.critical(f"模型 {self.model_name} 网络错误达到最大重试次数: {str(e)}")
|
||||
raise RuntimeError(f"网络请求失败: {str(e)}") from e
|
||||
except Exception as e:
|
||||
logger.critical(f"未预期的错误: {str(e)}")
|
||||
logger.critical(f"模型 {self.model_name} 未预期的错误: {str(e)}")
|
||||
raise RuntimeError(f"请求过程中发生错误: {str(e)}") from e
|
||||
|
||||
except aiohttp.ClientResponseError as e:
|
||||
# 处理aiohttp抛出的响应错误
|
||||
if retry < policy["max_retries"] - 1:
|
||||
wait_time = policy["base_wait"] * (2**retry)
|
||||
logger.error(f"HTTP响应错误，等待{wait_time}秒后重试... 状态码: {e.status}, 错误: {e.message}")
|
||||
logger.error(f"模型 {self.model_name} HTTP响应错误，等待{wait_time}秒后重试... 状态码: {e.status}, 错误: {e.message}")
|
||||
try:
|
||||
if hasattr(e, "response") and e.response and hasattr(e.response, "text"):
|
||||
error_text = await e.response.text()
|
||||
@@ -403,27 +405,27 @@ class LLM_request:
|
||||
if "error" in error_item and isinstance(error_item["error"], dict):
|
||||
error_obj = error_item["error"]
|
||||
logger.error(
|
||||
f"服务器错误详情: 代码={error_obj.get('code')}, "
|
||||
f"模型 {self.model_name} 服务器错误详情: 代码={error_obj.get('code')}, "
|
||||
f"状态={error_obj.get('status')}, "
|
||||
f"消息={error_obj.get('message')}"
|
||||
)
|
||||
elif isinstance(error_json, dict) and "error" in error_json:
|
||||
error_obj = error_json.get("error", {})
|
||||
logger.error(
|
||||
f"服务器错误详情: 代码={error_obj.get('code')}, "
|
||||
f"模型 {self.model_name} 服务器错误详情: 代码={error_obj.get('code')}, "
|
||||
f"状态={error_obj.get('status')}, "
|
||||
f"消息={error_obj.get('message')}"
|
||||
)
|
||||
else:
|
||||
logger.error(f"服务器错误响应: {error_json}")
|
||||
logger.error(f"模型 {self.model_name} 服务器错误响应: {error_json}")
|
||||
except (json.JSONDecodeError, TypeError) as json_err:
|
||||
logger.warning(f"响应不是有效的JSON: {str(json_err)}, 原始内容: {error_text[:200]}")
|
||||
logger.warning(f"模型 {self.model_name} 响应不是有效的JSON: {str(json_err)}, 原始内容: {error_text[:200]}")
|
||||
except (AttributeError, TypeError, ValueError) as parse_err:
|
||||
logger.warning(f"无法解析响应错误内容: {str(parse_err)}")
|
||||
logger.warning(f"模型 {self.model_name} 无法解析响应错误内容: {str(parse_err)}")
|
||||
|
||||
await asyncio.sleep(wait_time)
|
||||
else:
|
||||
logger.critical(f"HTTP响应错误达到最大重试次数: 状态码: {e.status}, 错误: {e.message}")
|
||||
logger.critical(f"模型 {self.model_name} HTTP响应错误达到最大重试次数: 状态码: {e.status}, 错误: {e.message}")
|
||||
# 安全地检查和记录请求详情
|
||||
if (
|
||||
image_base64
|
||||
@@ -440,14 +442,14 @@ class LLM_request:
|
||||
f"{image_base64[:10]}...{image_base64[-10:]}"
|
||||
)
|
||||
logger.critical(f"请求头: {await self._build_headers(no_key=True)} 请求体: {payload}")
|
||||
raise RuntimeError(f"API请求失败: 状态码 {e.status}, {e.message}") from e
|
||||
raise RuntimeError(f"模型 {self.model_name} API请求失败: 状态码 {e.status}, {e.message}") from e
|
||||
except Exception as e:
|
||||
if retry < policy["max_retries"] - 1:
|
||||
wait_time = policy["base_wait"] * (2**retry)
|
||||
logger.error(f"请求失败，等待{wait_time}秒后重试... 错误: {str(e)}")
|
||||
logger.error(f"模型 {self.model_name} 请求失败，等待{wait_time}秒后重试... 错误: {str(e)}")
|
||||
await asyncio.sleep(wait_time)
|
||||
else:
|
||||
logger.critical(f"请求失败: {str(e)}")
|
||||
logger.critical(f"模型 {self.model_name} 请求失败: {str(e)}")
|
||||
# 安全地检查和记录请求详情
|
||||
if (
|
||||
image_base64
|
||||
@@ -464,10 +466,10 @@ class LLM_request:
|
||||
f"{image_base64[:10]}...{image_base64[-10:]}"
|
||||
)
|
||||
logger.critical(f"请求头: {await self._build_headers(no_key=True)} 请求体: {payload}")
|
||||
raise RuntimeError(f"API请求失败: {str(e)}") from e
|
||||
raise RuntimeError(f"模型 {self.model_name} API请求失败: {str(e)}") from e
|
||||
|
||||
logger.error("达到最大重试次数，请求仍然失败")
|
||||
raise RuntimeError("达到最大重试次数，API请求仍然失败")
|
||||
logger.error(f"模型 {self.model_name} 达到最大重试次数，请求仍然失败")
|
||||
raise RuntimeError(f"模型 {self.model_name} 达到最大重试次数，API请求仍然失败")
|
||||
|
||||
async def _transform_parameters(self, params: dict) -> dict:
|
||||
"""
|
||||
|
||||
@@ -3,6 +3,7 @@ import os
|
||||
import sys
|
||||
from typing import Dict
|
||||
import asyncio
|
||||
from dateutil import tz
|
||||
|
||||
# 添加项目根目录到 Python 路径
|
||||
root_path = os.path.abspath(os.path.join(os.path.dirname(__file__), "../../.."))
|
||||
@@ -13,6 +14,8 @@ from src.common.logger import get_module_logger, SCHEDULE_STYLE_CONFIG, LogConfi
|
||||
from src.plugins.models.utils_model import LLM_request # noqa: E402
|
||||
from src.plugins.config.config import global_config # noqa: E402
|
||||
|
||||
TIME_ZONE = tz.gettz(global_config.TIME_ZONE) # 设置时区
|
||||
|
||||
|
||||
schedule_config = LogConfig(
|
||||
# 使用海马体专用样式
|
||||
@@ -44,7 +47,7 @@ class ScheduleGenerator:
|
||||
self.personality = ""
|
||||
self.behavior = ""
|
||||
|
||||
self.start_time = datetime.datetime.now()
|
||||
self.start_time = datetime.datetime.now(TIME_ZONE)
|
||||
|
||||
self.schedule_doing_update_interval = 300 # 最好大于60
|
||||
|
||||
@@ -74,7 +77,7 @@ class ScheduleGenerator:
|
||||
while True:
|
||||
# print(self.get_current_num_task(1, True))
|
||||
|
||||
current_time = datetime.datetime.now()
|
||||
current_time = datetime.datetime.now(TIME_ZONE)
|
||||
|
||||
# 检查是否需要重新生成日程（日期变化）
|
||||
if current_time.date() != self.start_time.date():
|
||||
@@ -100,7 +103,7 @@ class ScheduleGenerator:
|
||||
Returns:
|
||||
tuple: (today_schedule_text, today_schedule) 今天的日程文本和解析后的日程字典
|
||||
"""
|
||||
today = datetime.datetime.now()
|
||||
today = datetime.datetime.now(TIME_ZONE)
|
||||
yesterday = today - datetime.timedelta(days=1)
|
||||
|
||||
# 先检查昨天的日程
|
||||
@@ -156,7 +159,7 @@ class ScheduleGenerator:
|
||||
"""打印完整的日程安排"""
|
||||
if not self.today_schedule_text:
|
||||
logger.warning("今日日程有误，将在下次运行时重新生成")
|
||||
db.schedule.delete_one({"date": datetime.datetime.now().strftime("%Y-%m-%d")})
|
||||
db.schedule.delete_one({"date": datetime.datetime.now(TIME_ZONE).strftime("%Y-%m-%d")})
|
||||
else:
|
||||
logger.info("=== 今日日程安排 ===")
|
||||
logger.info(self.today_schedule_text)
|
||||
@@ -165,7 +168,7 @@ class ScheduleGenerator:
|
||||
|
||||
async def update_today_done_list(self):
|
||||
# 更新数据库中的 today_done_list
|
||||
today_str = datetime.datetime.now().strftime("%Y-%m-%d")
|
||||
today_str = datetime.datetime.now(TIME_ZONE).strftime("%Y-%m-%d")
|
||||
existing_schedule = db.schedule.find_one({"date": today_str})
|
||||
|
||||
if existing_schedule:
|
||||
@@ -177,7 +180,7 @@ class ScheduleGenerator:
|
||||
|
||||
async def move_doing(self, mind_thinking: str = ""):
|
||||
try:
|
||||
current_time = datetime.datetime.now()
|
||||
current_time = datetime.datetime.now(TIME_ZONE)
|
||||
if mind_thinking:
|
||||
doing_prompt = self.construct_doing_prompt(current_time, mind_thinking)
|
||||
else:
|
||||
@@ -246,7 +249,7 @@ class ScheduleGenerator:
|
||||
|
||||
def save_today_schedule_to_db(self):
|
||||
"""保存日程到数据库，同时初始化 today_done_list"""
|
||||
date_str = datetime.datetime.now().strftime("%Y-%m-%d")
|
||||
date_str = datetime.datetime.now(TIME_ZONE).strftime("%Y-%m-%d")
|
||||
schedule_data = {
|
||||
"date": date_str,
|
||||
"schedule": self.today_schedule_text,
|
||||
|
||||
@@ -139,10 +139,10 @@ class LLMStatistics:
|
||||
user_info = doc.get("user_info", {})
|
||||
group_info = chat_info.get("group_info") if chat_info else {}
|
||||
# print(f"group_info: {group_info}")
|
||||
group_name = "unknown"
|
||||
group_name = None
|
||||
if group_info:
|
||||
group_name = group_info["group_name"]
|
||||
if user_info and group_name == "unknown":
|
||||
group_name = group_info.get("group_name", f"群{group_info.get('group_id')}")
|
||||
if user_info and not group_name:
|
||||
group_name = user_info["user_nickname"]
|
||||
# print(f"group_name: {group_name}")
|
||||
stats["messages_by_user"][user_id] += 1
|
||||
@@ -314,7 +314,7 @@ class LLMStatistics:
|
||||
def _console_output_loop(self):
|
||||
"""控制台输出循环，每5分钟输出一次最近1小时的统计"""
|
||||
while self.running:
|
||||
# 等待5分钟
|
||||
# 等待5分钟
|
||||
for _ in range(300): # 5分钟 = 300秒
|
||||
if not self.running:
|
||||
break
|
||||
@@ -325,14 +325,14 @@ class LLMStatistics:
|
||||
hour_stats = self._collect_statistics_for_period(now - timedelta(hours=1))
|
||||
|
||||
# 使用logger输出
|
||||
stats_output = self._format_stats_section_lite(hour_stats, "最近1小时统计：详细信息见根目录文件：llm_statistics.txt")
|
||||
stats_output = self._format_stats_section_lite(
|
||||
hour_stats, "最近1小时统计：详细信息见根目录文件：llm_statistics.txt"
|
||||
)
|
||||
logger.info("\n" + stats_output + "\n" + "=" * 50)
|
||||
|
||||
except Exception:
|
||||
logger.exception("控制台统计数据输出失败")
|
||||
|
||||
|
||||
|
||||
def _stats_loop(self):
|
||||
"""统计循环，每5分钟运行一次"""
|
||||
while self.running:
|
||||
|
||||
@@ -1,5 +1,5 @@
|
||||
[inner]
|
||||
version = "1.1.0"
|
||||
version = "1.1.3"
|
||||
|
||||
|
||||
#以下是给开发人员阅读的，一般用户不需要阅读
|
||||
@@ -48,6 +48,7 @@ enable_schedule_gen = true # 是否启用日程表(尚未完成)
|
||||
prompt_schedule_gen = "用几句话描述描述性格特点或行动规律，这个特征会用来生成日程表"
|
||||
schedule_doing_update_interval = 900 # 日程表更新间隔 单位秒
|
||||
schedule_temperature = 0.3 # 日程表温度，建议0.3-0.6
|
||||
time_zone = "Asia/Shanghai" # 给你的机器人设置时区，可以解决运行电脑时区和国内时区不同的情况，或者模拟国外留学生日程
|
||||
|
||||
[platforms] # 必填项目，填写每个平台适配器提供的链接
|
||||
nonebot-qq="http://127.0.0.1:18002/api/message"
|
||||
@@ -93,8 +94,9 @@ emoji_response_penalty = 0.1 # 表情包回复惩罚系数，设为0为不回复
|
||||
|
||||
|
||||
[emoji]
|
||||
check_interval = 15 # 检查破损表情包的时间间隔(分钟)
|
||||
register_interval = 60 # 注册表情包的时间间隔(分钟)
|
||||
max_emoji_num = 120 # 表情包最大数量
|
||||
max_reach_deletion = true # 开启则在达到最大数量时删除表情包，关闭则达到最大数量时不删除，只是不会继续收集表情包
|
||||
check_interval = 30 # 检查表情包（注册，破损，删除）的时间间隔(分钟)
|
||||
auto_save = true # 是否保存表情包和图片
|
||||
enable_check = false # 是否启用表情包过滤
|
||||
check_prompt = "符合公序良俗" # 表情包过滤要求
|
||||
|
||||
Reference in New Issue
Block a user