551 lines
23 KiB
Python
551 lines
23 KiB
Python
import hashlib
|
||
import asyncio
|
||
import json
|
||
import time
|
||
import random
|
||
|
||
from json_repair import repair_json
|
||
from typing import Union
|
||
|
||
from src.common.logger import get_logger
|
||
from src.common.database.database import db
|
||
from src.common.database.database_model import PersonInfo
|
||
from src.llm_models.utils_model import LLMRequest
|
||
from src.config.config import global_config, model_config
|
||
|
||
|
||
logger = get_logger("person_info")
|
||
|
||
def get_person_id(platform: str, user_id: Union[int, str]) -> str:
|
||
"""获取唯一id"""
|
||
if "-" in platform:
|
||
platform = platform.split("-")[1]
|
||
components = [platform, str(user_id)]
|
||
key = "_".join(components)
|
||
return hashlib.md5(key.encode()).hexdigest()
|
||
|
||
def get_person_id_by_person_name(person_name: str) -> str:
|
||
"""根据用户名获取用户ID"""
|
||
try:
|
||
record = PersonInfo.get_or_none(PersonInfo.person_name == person_name)
|
||
return record.person_id if record else ""
|
||
except Exception as e:
|
||
logger.error(f"根据用户名 {person_name} 获取用户ID时出错 (Peewee): {e}")
|
||
return ""
|
||
|
||
def is_person_known(person_id: str = None,user_id: str = None,platform: str = None,person_name: str = None) -> bool:
|
||
if person_id:
|
||
person = PersonInfo.get_or_none(PersonInfo.person_id == person_id)
|
||
return person.is_known if person else False
|
||
elif user_id and platform:
|
||
person_id = get_person_id(platform, user_id)
|
||
person = PersonInfo.get_or_none(PersonInfo.person_id == person_id)
|
||
return person.is_known if person else False
|
||
elif person_name:
|
||
person_id = get_person_id_by_person_name(person_name)
|
||
person = PersonInfo.get_or_none(PersonInfo.person_id == person_id)
|
||
return person.is_known if person else False
|
||
else:
|
||
return False
|
||
|
||
class Person:
|
||
@classmethod
|
||
def register_person(cls, platform: str, user_id: str, nickname: str):
|
||
"""
|
||
注册新用户的类方法
|
||
必须输入 platform、user_id 和 nickname 参数
|
||
|
||
Args:
|
||
platform: 平台名称
|
||
user_id: 用户ID
|
||
nickname: 用户昵称
|
||
|
||
Returns:
|
||
Person: 新注册的Person实例
|
||
"""
|
||
if not platform or not user_id or not nickname:
|
||
logger.error("注册用户失败：platform、user_id 和 nickname 都是必需参数")
|
||
return None
|
||
|
||
# 生成唯一的person_id
|
||
person_id = get_person_id(platform, user_id)
|
||
|
||
if is_person_known(person_id=person_id):
|
||
logger.info(f"用户 {nickname} 已存在")
|
||
return Person(person_id=person_id)
|
||
|
||
# 创建Person实例
|
||
person = cls.__new__(cls)
|
||
|
||
# 设置基本属性
|
||
person.person_id = person_id
|
||
person.platform = platform
|
||
person.user_id = user_id
|
||
person.nickname = nickname
|
||
|
||
# 初始化默认值
|
||
person.is_known = True # 注册后立即标记为已认识
|
||
person.person_name = nickname # 使用nickname作为初始person_name
|
||
person.name_reason = "用户注册时设置的昵称"
|
||
person.know_times = 1
|
||
person.know_since = time.time()
|
||
person.last_know = time.time()
|
||
person.points = []
|
||
|
||
# 初始化性格特征相关字段
|
||
person.attitude_to_me = 0
|
||
person.attitude_to_me_confidence = 1
|
||
|
||
person.neuroticism = 5
|
||
person.neuroticism_confidence = 1
|
||
|
||
person.friendly_value = 50
|
||
person.friendly_value_confidence = 1
|
||
|
||
person.rudeness = 50
|
||
person.rudeness_confidence = 1
|
||
|
||
person.conscientiousness = 50
|
||
person.conscientiousness_confidence = 1
|
||
|
||
person.likeness = 50
|
||
person.likeness_confidence = 1
|
||
|
||
# 同步到数据库
|
||
person.sync_to_database()
|
||
|
||
logger.info(f"成功注册新用户：{person_id}，平台：{platform}，昵称：{nickname}")
|
||
|
||
return person
|
||
|
||
def __init__(self, platform: str = "", user_id: str = "",person_id: str = "",person_name: str = ""):
|
||
if platform == global_config.bot.platform and user_id == global_config.bot.qq_account:
|
||
self.is_known = True
|
||
self.person_id = get_person_id(platform, user_id)
|
||
self.user_id = user_id
|
||
self.platform = platform
|
||
self.nickname = global_config.bot.nickname
|
||
self.person_name = global_config.bot.nickname
|
||
return
|
||
|
||
self.user_id = ""
|
||
self.platform = ""
|
||
|
||
if person_id:
|
||
self.person_id = person_id
|
||
elif person_name:
|
||
self.person_id = get_person_id_by_person_name(person_name)
|
||
if not self.person_id:
|
||
logger.error(f"根据用户名 {person_name} 获取用户ID时出错，不存在用户{person_name}")
|
||
return
|
||
elif platform and user_id:
|
||
self.person_id = get_person_id(platform, user_id)
|
||
self.user_id = user_id
|
||
self.platform = platform
|
||
else:
|
||
logger.error("Person 初始化失败，缺少必要参数")
|
||
raise ValueError("Person 初始化失败，缺少必要参数")
|
||
|
||
if not is_person_known(person_id=self.person_id):
|
||
self.is_known = False
|
||
logger.warning(f"用户 {platform}:{user_id}:{person_name}:{person_id} 尚未认识")
|
||
self.person_name = f"未知用户{self.person_id[:4]}"
|
||
return
|
||
|
||
self.is_known = False
|
||
|
||
# 初始化默认值
|
||
self.nickname = ""
|
||
self.person_name = None
|
||
self.name_reason = None
|
||
self.know_times = 0
|
||
self.know_since = None
|
||
self.last_know = None
|
||
self.points = []
|
||
|
||
# 初始化性格特征相关字段
|
||
self.attitude_to_me:float = 0
|
||
self.attitude_to_me_confidence:float = 1
|
||
|
||
self.neuroticism:float = 5
|
||
self.neuroticism_confidence:float = 1
|
||
|
||
self.friendly_value:float = 50
|
||
self.friendly_value_confidence:float = 1
|
||
|
||
self.rudeness:float = 50
|
||
self.rudeness_confidence:float = 1
|
||
|
||
self.conscientiousness:float = 50
|
||
self.conscientiousness_confidence:float = 1
|
||
|
||
self.likeness:float = 50
|
||
self.likeness_confidence:float = 1
|
||
|
||
# 从数据库加载数据
|
||
self.load_from_database()
|
||
|
||
def load_from_database(self):
|
||
"""从数据库加载个人信息数据"""
|
||
try:
|
||
# 查询数据库中的记录
|
||
record = PersonInfo.get_or_none(PersonInfo.person_id == self.person_id)
|
||
|
||
if record:
|
||
self.user_id = record.user_id if record.user_id else ""
|
||
self.platform = record.platform if record.platform else ""
|
||
self.is_known = record.is_known if record.is_known else False
|
||
self.nickname = record.nickname if record.nickname else ""
|
||
self.person_name = record.person_name if record.person_name else self.nickname
|
||
self.name_reason = record.name_reason if record.name_reason else None
|
||
self.know_times = record.know_times if record.know_times else 0
|
||
|
||
# 处理points字段（JSON格式的列表）
|
||
if record.points:
|
||
try:
|
||
self.points = json.loads(record.points)
|
||
except (json.JSONDecodeError, TypeError):
|
||
logger.warning(f"解析用户 {self.person_id} 的points字段失败，使用默认值")
|
||
self.points = []
|
||
else:
|
||
self.points = []
|
||
|
||
# 加载性格特征相关字段
|
||
if record.attitude_to_me and not isinstance(record.attitude_to_me, str):
|
||
self.attitude_to_me = record.attitude_to_me
|
||
|
||
if record.attitude_to_me_confidence is not None:
|
||
self.attitude_to_me_confidence = float(record.attitude_to_me_confidence)
|
||
|
||
if record.friendly_value is not None:
|
||
self.friendly_value = float(record.friendly_value)
|
||
|
||
if record.friendly_value_confidence is not None:
|
||
self.friendly_value_confidence = float(record.friendly_value_confidence)
|
||
|
||
if record.rudeness is not None:
|
||
self.rudeness = float(record.rudeness)
|
||
|
||
if record.rudeness_confidence is not None:
|
||
self.rudeness_confidence = float(record.rudeness_confidence)
|
||
|
||
if record.neuroticism and not isinstance(record.neuroticism, str):
|
||
self.neuroticism = float(record.neuroticism)
|
||
|
||
if record.neuroticism_confidence is not None:
|
||
self.neuroticism_confidence = float(record.neuroticism_confidence)
|
||
|
||
if record.conscientiousness is not None:
|
||
self.conscientiousness = float(record.conscientiousness)
|
||
|
||
if record.conscientiousness_confidence is not None:
|
||
self.conscientiousness_confidence = float(record.conscientiousness_confidence)
|
||
|
||
if record.likeness is not None:
|
||
self.likeness = float(record.likeness)
|
||
|
||
if record.likeness_confidence is not None:
|
||
self.likeness_confidence = float(record.likeness_confidence)
|
||
|
||
logger.debug(f"已从数据库加载用户 {self.person_id} 的信息")
|
||
else:
|
||
self.sync_to_database()
|
||
logger.info(f"用户 {self.person_id} 在数据库中不存在，使用默认值并创建")
|
||
|
||
except Exception as e:
|
||
logger.error(f"从数据库加载用户 {self.person_id} 信息时出错: {e}")
|
||
# 出错时保持默认值
|
||
|
||
def sync_to_database(self):
|
||
"""将所有属性同步回数据库"""
|
||
if not self.is_known:
|
||
return
|
||
try:
|
||
# 准备数据
|
||
data = {
|
||
'person_id': self.person_id,
|
||
'is_known': self.is_known,
|
||
'platform': self.platform,
|
||
'user_id': self.user_id,
|
||
'nickname': self.nickname,
|
||
'person_name': self.person_name,
|
||
'name_reason': self.name_reason,
|
||
'know_times': self.know_times,
|
||
'know_since': self.know_since,
|
||
'last_know': self.last_know,
|
||
'points': json.dumps(self.points, ensure_ascii=False) if self.points else json.dumps([], ensure_ascii=False),
|
||
'attitude_to_me': self.attitude_to_me,
|
||
'attitude_to_me_confidence': self.attitude_to_me_confidence,
|
||
'friendly_value': self.friendly_value,
|
||
'friendly_value_confidence': self.friendly_value_confidence,
|
||
'rudeness': self.rudeness,
|
||
'rudeness_confidence': self.rudeness_confidence,
|
||
'neuroticism': self.neuroticism,
|
||
'neuroticism_confidence': self.neuroticism_confidence,
|
||
'conscientiousness': self.conscientiousness,
|
||
'conscientiousness_confidence': self.conscientiousness_confidence,
|
||
'likeness': self.likeness,
|
||
'likeness_confidence': self.likeness_confidence,
|
||
}
|
||
|
||
# 检查记录是否存在
|
||
record = PersonInfo.get_or_none(PersonInfo.person_id == self.person_id)
|
||
|
||
if record:
|
||
# 更新现有记录
|
||
for field, value in data.items():
|
||
if hasattr(record, field):
|
||
setattr(record, field, value)
|
||
record.save()
|
||
logger.debug(f"已同步用户 {self.person_id} 的信息到数据库")
|
||
else:
|
||
# 创建新记录
|
||
PersonInfo.create(**data)
|
||
logger.debug(f"已创建用户 {self.person_id} 的信息到数据库")
|
||
|
||
except Exception as e:
|
||
logger.error(f"同步用户 {self.person_id} 信息到数据库时出错: {e}")
|
||
|
||
def build_relationship(self,points_num=3):
|
||
# print(self.person_name,self.nickname,self.platform,self.is_known)
|
||
|
||
|
||
if not self.is_known:
|
||
return ""
|
||
|
||
# 按时间排序forgotten_points
|
||
current_points = self.points
|
||
current_points.sort(key=lambda x: x[2])
|
||
# 按权重加权随机抽取最多3个不重复的points，point[1]的值在1-10之间，权重越高被抽到概率越大
|
||
if len(current_points) > points_num:
|
||
# point[1] 取值范围1-10，直接作为权重
|
||
weights = [max(1, min(10, int(point[1]))) for point in current_points]
|
||
# 使用加权采样不放回，保证不重复
|
||
indices = list(range(len(current_points)))
|
||
points = []
|
||
for _ in range(points_num):
|
||
if not indices:
|
||
break
|
||
sub_weights = [weights[i] for i in indices]
|
||
chosen_idx = random.choices(indices, weights=sub_weights, k=1)[0]
|
||
points.append(current_points[chosen_idx])
|
||
indices.remove(chosen_idx)
|
||
else:
|
||
points = current_points
|
||
|
||
# 构建points文本
|
||
points_text = "\n".join([f"{point[2]}：{point[0]}" for point in points])
|
||
|
||
nickname_str = ""
|
||
if self.person_name != self.nickname:
|
||
nickname_str = f"(ta在{self.platform}上的昵称是{self.nickname})"
|
||
|
||
relation_info = ""
|
||
|
||
attitude_info = ""
|
||
if self.attitude_to_me:
|
||
if self.attitude_to_me > 8:
|
||
attitude_info = f"{self.person_name}对你的态度十分好,"
|
||
elif self.attitude_to_me > 5:
|
||
attitude_info = f"{self.person_name}对你的态度较好,"
|
||
|
||
|
||
if self.attitude_to_me < -8:
|
||
attitude_info = f"{self.person_name}对你的态度十分恶劣,"
|
||
elif self.attitude_to_me < -4:
|
||
attitude_info = f"{self.person_name}对你的态度不好,"
|
||
elif self.attitude_to_me < 0:
|
||
attitude_info = f"{self.person_name}对你的态度一般,"
|
||
|
||
neuroticism_info = ""
|
||
if self.neuroticism:
|
||
if self.neuroticism > 8:
|
||
neuroticism_info = f"{self.person_name}的情绪十分活跃，容易情绪化,"
|
||
elif self.neuroticism > 6:
|
||
neuroticism_info = f"{self.person_name}的情绪比较活跃,"
|
||
elif self.neuroticism > 4:
|
||
neuroticism_info = ""
|
||
elif self.neuroticism > 2:
|
||
neuroticism_info = f"{self.person_name}的情绪比较稳定,"
|
||
else:
|
||
neuroticism_info = f"{self.person_name}的情绪非常稳定,毫无波动"
|
||
|
||
points_info = ""
|
||
if points_text:
|
||
points_info = f"你还记得ta最近做的事：{points_text}"
|
||
|
||
if not (nickname_str or attitude_info or neuroticism_info or points_info):
|
||
return ""
|
||
relation_info = f"{self.person_name}:{nickname_str}{attitude_info}{neuroticism_info}{points_info}"
|
||
|
||
return relation_info
|
||
|
||
|
||
class PersonInfoManager:
|
||
def __init__(self):
|
||
|
||
self.person_name_list = {}
|
||
self.qv_name_llm = LLMRequest(model_set=model_config.model_task_config.utils, request_type="relation.qv_name")
|
||
try:
|
||
db.connect(reuse_if_open=True)
|
||
# 设置连接池参数
|
||
if hasattr(db, "execute_sql"):
|
||
# 设置SQLite优化参数
|
||
db.execute_sql("PRAGMA cache_size = -64000") # 64MB缓存
|
||
db.execute_sql("PRAGMA temp_store = memory") # 临时存储在内存中
|
||
db.execute_sql("PRAGMA mmap_size = 268435456") # 256MB内存映射
|
||
db.create_tables([PersonInfo], safe=True)
|
||
except Exception as e:
|
||
logger.error(f"数据库连接或 PersonInfo 表创建失败: {e}")
|
||
|
||
# 初始化时读取所有person_name
|
||
try:
|
||
for record in PersonInfo.select(PersonInfo.person_id, PersonInfo.person_name).where(
|
||
PersonInfo.person_name.is_null(False)
|
||
):
|
||
if record.person_name:
|
||
self.person_name_list[record.person_id] = record.person_name
|
||
logger.debug(f"已加载 {len(self.person_name_list)} 个用户名称 (Peewee)")
|
||
except Exception as e:
|
||
logger.error(f"从 Peewee 加载 person_name_list 失败: {e}")
|
||
|
||
|
||
|
||
@staticmethod
|
||
def _extract_json_from_text(text: str) -> dict:
|
||
"""从文本中提取JSON数据的高容错方法"""
|
||
try:
|
||
fixed_json = repair_json(text)
|
||
if isinstance(fixed_json, str):
|
||
parsed_json = json.loads(fixed_json)
|
||
else:
|
||
parsed_json = fixed_json
|
||
|
||
if isinstance(parsed_json, list) and parsed_json:
|
||
parsed_json = parsed_json[0]
|
||
|
||
if isinstance(parsed_json, dict):
|
||
return parsed_json
|
||
|
||
except Exception as e:
|
||
logger.warning(f"JSON提取失败: {e}")
|
||
|
||
logger.warning(f"无法从文本中提取有效的JSON字典: {text}")
|
||
logger.info(f"文本: {text}")
|
||
return {"nickname": "", "reason": ""}
|
||
|
||
async def _generate_unique_person_name(self, base_name: str) -> str:
|
||
"""生成唯一的 person_name，如果存在重复则添加数字后缀"""
|
||
# 处理空昵称的情况
|
||
if not base_name or base_name.isspace():
|
||
base_name = "空格"
|
||
|
||
# 检查基础名称是否已存在
|
||
if base_name not in self.person_name_list.values():
|
||
return base_name
|
||
|
||
# 如果存在，添加数字后缀
|
||
counter = 1
|
||
while True:
|
||
new_name = f"{base_name}[{counter}]"
|
||
if new_name not in self.person_name_list.values():
|
||
return new_name
|
||
counter += 1
|
||
|
||
async def qv_person_name(
|
||
self, person_id: str, user_nickname: str, user_cardname: str, user_avatar: str, request: str = ""
|
||
):
|
||
"""给某个用户取名"""
|
||
if not person_id:
|
||
logger.debug("取名失败：person_id不能为空")
|
||
return None
|
||
|
||
person = Person(person_id=person_id)
|
||
old_name = person.person_name
|
||
old_reason = person.name_reason
|
||
|
||
max_retries = 8
|
||
current_try = 0
|
||
existing_names_str = ""
|
||
current_name_set = set(self.person_name_list.values())
|
||
|
||
while current_try < max_retries:
|
||
# prompt_personality =get_individuality().get_prompt(x_person=2, level=1)
|
||
bot_name = global_config.bot.nickname
|
||
|
||
qv_name_prompt = f"你是{bot_name}，一个伪装成人类的AI，你不能让别人发现这一点，"
|
||
qv_name_prompt += f"现在你想给一个用户取一个昵称，用户的qq昵称是{user_nickname}，"
|
||
qv_name_prompt += f"用户的qq群昵称名是{user_cardname}，"
|
||
if user_avatar:
|
||
qv_name_prompt += f"用户的qq头像是{user_avatar}，"
|
||
if old_name:
|
||
qv_name_prompt += f"你之前叫他{old_name}，是因为{old_reason}，"
|
||
|
||
qv_name_prompt += f"\n其他取名的要求是：{request}，不要太浮夸，简短，"
|
||
qv_name_prompt += "\n请根据以上用户信息，想想你叫他什么比较好，不要太浮夸，请最好使用用户的qq昵称或群昵称原文，可以稍作修改，优先使用原文。优先使用用户的qq昵称或者群昵称原文。"
|
||
|
||
if existing_names_str:
|
||
qv_name_prompt += f"\n请注意，以下名称已被你尝试过或已知存在，请避免：{existing_names_str}。\n"
|
||
|
||
if len(current_name_set) < 50 and current_name_set:
|
||
qv_name_prompt += f"已知的其他昵称有: {', '.join(list(current_name_set)[:10])}等。\n"
|
||
|
||
qv_name_prompt += "请用json给出你的想法，并给出理由，示例如下："
|
||
qv_name_prompt += """{
|
||
"nickname": "昵称",
|
||
"reason": "理由"
|
||
}"""
|
||
response, _ = await self.qv_name_llm.generate_response_async(qv_name_prompt)
|
||
# logger.info(f"取名提示词：{qv_name_prompt}\n取名回复：{response}")
|
||
result = self._extract_json_from_text(response)
|
||
|
||
if not result or not result.get("nickname"):
|
||
logger.error("生成的昵称为空或结果格式不正确，重试中...")
|
||
current_try += 1
|
||
continue
|
||
|
||
generated_nickname = result["nickname"]
|
||
|
||
is_duplicate = False
|
||
if generated_nickname in current_name_set:
|
||
is_duplicate = True
|
||
logger.info(f"尝试给用户{user_nickname} {person_id} 取名，但是 {generated_nickname} 已存在，重试中...")
|
||
else:
|
||
|
||
def _db_check_name_exists_sync(name_to_check):
|
||
return PersonInfo.select().where(PersonInfo.person_name == name_to_check).exists()
|
||
|
||
if await asyncio.to_thread(_db_check_name_exists_sync, generated_nickname):
|
||
is_duplicate = True
|
||
current_name_set.add(generated_nickname)
|
||
|
||
if not is_duplicate:
|
||
person.person_name = generated_nickname
|
||
person.name_reason = result.get("reason", "未提供理由")
|
||
person.sync_to_database()
|
||
|
||
logger.info(
|
||
f"成功给用户{user_nickname} {person_id} 取名 {generated_nickname}，理由：{result.get('reason', '未提供理由')}"
|
||
)
|
||
|
||
self.person_name_list[person_id] = generated_nickname
|
||
return result
|
||
else:
|
||
if existing_names_str:
|
||
existing_names_str += "、"
|
||
existing_names_str += generated_nickname
|
||
logger.debug(f"生成的昵称 {generated_nickname} 已存在，重试中...")
|
||
current_try += 1
|
||
|
||
# 如果多次尝试后仍未成功，使用唯一的 user_nickname 作为默认值
|
||
unique_nickname = await self._generate_unique_person_name(user_nickname)
|
||
logger.warning(f"在{max_retries}次尝试后未能生成唯一昵称，使用默认昵称 {unique_nickname}")
|
||
person.person_name = unique_nickname
|
||
person.name_reason = "使用用户原始昵称作为默认值"
|
||
person.sync_to_database()
|
||
self.person_name_list[person_id] = unique_nickname
|
||
return {"nickname": unique_nickname, "reason": "使用用户原始昵称作为默认值"}
|
||
|
||
|
||
person_info_manager = PersonInfoManager()
|