Merge remote-tracking branch 'upstream/debug' into feat_regix

This commit is contained in:
Rikki
2025-03-11 05:05:07 +08:00
parent cd96644a4c
commit a54ca8ce18
35 changed files with 1746 additions and 341 deletions

View File

@@ -1,6 +1,6 @@
import os
from dataclasses import dataclass, field
from typing import Dict, Optional
from typing import Dict, List, Optional
import tomli
from loguru import logger
@@ -12,10 +12,12 @@ 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
BOT_ALIAS_NAMES: List[str] = field(default_factory=list) # 别名,可以通过这个叫它
# 消息处理相关配置
MIN_TEXT_LENGTH: int = 2 # 最小处理文本长度
@@ -82,23 +84,25 @@ class BotConfig:
PROMPT_PERSONALITY = [
"曾经是一个学习地质的女大学生,现在学习心理学和脑科学,你会刷贴吧",
"是一个女大学生,你有黑色头发,你会刷小红书",
"是一个女大学生你会刷b站对ACG文化感兴趣"
"是一个女大学生你会刷b站对ACG文化感兴趣",
]
PROMPT_SCHEDULE_GEN="一个曾经学习地质,现在学习心理学和脑科学的女大学生喜欢刷qq贴吧知乎和小红书"
PERSONALITY_1: float = 0.6 # 第一种人格概率
PERSONALITY_2: float = 0.3 # 第二种人格概率
PERSONALITY_3: float = 0.1 # 第三种人格概率
memory_ban_words: list = field(default_factory=lambda: ['表情包', '图片', '回复', '聊天记录']) # 添加新的配置项默认值
PROMPT_SCHEDULE_GEN = "一个曾经学习地质,现在学习心理学和脑科学的女大学生喜欢刷qq贴吧知乎和小红书"
PERSONALITY_1: float = 0.6 # 第一种人格概率
PERSONALITY_2: float = 0.3 # 第二种人格概率
PERSONALITY_3: float = 0.1 # 第三种人格概率
memory_ban_words: list = field(
default_factory=lambda: ["表情包", "图片", "回复", "聊天记录"]
) # 添加新的配置项默认值
@staticmethod
def get_config_dir() -> str:
"""获取配置文件目录"""
current_dir = os.path.dirname(os.path.abspath(__file__))
root_dir = os.path.abspath(os.path.join(current_dir, '..', '..', '..'))
config_dir = os.path.join(root_dir, 'config')
root_dir = os.path.abspath(os.path.join(current_dir, "..", "..", ".."))
config_dir = os.path.join(root_dir, "config")
if not os.path.exists(config_dir):
os.makedirs(config_dir)
return config_dir
@@ -109,35 +113,32 @@ class BotConfig:
Args:
value[str]: 版本表达式(字符串)
Returns:
SpecifierSet
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"
)
except InvalidSpecifier:
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
"""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
Version
"""
if 'inner' in toml:
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, 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")
logger.error("配置文件中 inner 段 不存在, 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") from e
else:
toml["inner"] = {"version": "0.0.0"}
config_version = toml["inner"]["version"]
@@ -150,7 +151,7 @@ class BotConfig:
"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")
raise InvalidVersion("配置文件中 inner段 的 version 键是错误的版本描述\n") from e
return ver
@@ -160,26 +161,26 @@ class BotConfig:
config = cls()
def personality(parent: dict):
personality_config = parent['personality']
personality = personality_config.get('prompt_personality')
personality_config = parent["personality"]
personality = personality_config.get("prompt_personality")
if len(personality) >= 2:
logger.debug(f"载入自定义人格:{personality}")
config.PROMPT_PERSONALITY = personality_config.get('prompt_personality', config.PROMPT_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.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)
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)
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"]
@@ -192,12 +193,16 @@ class BotConfig:
config.BOT_QQ = int(bot_qq)
config.BOT_NICKNAME = bot_config.get("nickname", config.BOT_NICKNAME)
if config.INNER_VERSION in SpecifierSet(">=0.0.5"):
config.BOT_ALIAS_NAMES = bot_config.get("alias_names", config.BOT_ALIAS_NAMES)
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.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):
@@ -214,7 +219,7 @@ class BotConfig:
"llm_emotion_judge",
"vlm",
"embedding",
"moderation"
"moderation",
]
for item in config_list:
@@ -223,13 +228,7 @@ class BotConfig:
# 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
}
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
@@ -248,7 +247,7 @@ class BotConfig:
cfg_target[i] = cfg_item[i]
except KeyError as e:
logger.error(f"{item} 中的必要字段不存在,请检查")
raise KeyError(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") from e
provider = cfg_item.get("provider")
if provider is None:
@@ -273,10 +272,12 @@ class BotConfig:
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.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)
if config.INNER_VERSION in SpecifierSet(">=0.0.5"):
@@ -286,7 +287,7 @@ class BotConfig:
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)
# 在版本 >= 0.0.4 时才处理新增的配置项
if config.INNER_VERSION in SpecifierSet(">=0.0.4"):
config.memory_ban_words = set(memory_config.get("memory_ban_words", []))
@@ -307,10 +308,12 @@ class BotConfig:
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)
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"]
@@ -329,61 +332,19 @@ class BotConfig:
# 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",
"necessary": False
},
"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"
}
"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", "necessary": False},
"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
@@ -395,7 +356,7 @@ class BotConfig:
with open(config_path, "rb") as f:
try:
toml_dict = tomli.load(f)
except(tomli.TOMLDecodeError) as e:
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)
@@ -410,7 +371,7 @@ class BotConfig:
# 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]:
if "notice" in include_configs[key]:
logger.warning(include_configs[key]["notice"])
include_configs[key]["func"](toml_dict)
@@ -424,7 +385,7 @@ class BotConfig:
raise InvalidVersion(f"当前程序仅支持以下版本范围: {group_specifierset}")
# 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:
elif "necessary" in include_configs[key] and include_configs[key].get("necessary") is 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