fix：优化prompt和log
This commit is contained in:
SengokuCola
@@ -1,231 +0,0 @@
|
||||
from src.chat.heart_flow.observation.chatting_observation import ChattingObservation
|
||||
from src.chat.heart_flow.observation.observation import Observation
|
||||
from src.llm_models.utils_model import LLMRequest
|
||||
from src.config.config import global_config
|
||||
import time
|
||||
import traceback
|
||||
from src.common.logger import get_logger
|
||||
from src.individuality.individuality import get_individuality
|
||||
from src.chat.utils.prompt_builder import Prompt, global_prompt_manager
|
||||
from src.chat.utils.json_utils import safe_json_dumps
|
||||
from src.chat.message_receive.chat_stream import get_chat_manager
|
||||
from src.person_info.relationship_manager import get_relationship_manager
|
||||
from .base_processor import BaseProcessor
|
||||
from src.chat.focus_chat.info.mind_info import MindInfo
|
||||
from typing import List
|
||||
from src.chat.heart_flow.observation.hfcloop_observation import HFCloopObservation
|
||||
from src.chat.heart_flow.observation.actions_observation import ActionObservation
|
||||
from src.chat.focus_chat.info.info_base import InfoBase
|
||||
|
||||
logger = get_logger("processor")
|
||||
|
||||
|
||||
def init_prompt():
|
||||
group_prompt = """
|
||||
{extra_info}{relation_prompt}
|
||||
{cycle_info_block}
|
||||
现在是{time_now}，你正在上网，和qq群里的网友们聊天，以下是正在进行的聊天内容：
|
||||
{chat_observe_info}
|
||||
|
||||
{action_observe_info}
|
||||
|
||||
以下是你之前对聊天的观察和规划，你的名字是{bot_name}：
|
||||
{last_mind}
|
||||
|
||||
现在请你继续输出观察和规划，输出要求：
|
||||
1. 先关注未读新消息的内容和近期回复历史
|
||||
2. 根据新信息，修改和删除之前的观察和规划
|
||||
3. 注意群聊的时间线索，话题由谁发起，进展状况如何。
|
||||
4. 语言简洁自然，不要分点，不要浮夸，不要修辞，仅输出内容就好"""
|
||||
Prompt(group_prompt, "sub_heartflow_prompt_before")
|
||||
|
||||
private_prompt = """
|
||||
你的名字是{bot_name}
|
||||
{extra_info}{relation_prompt}
|
||||
{cycle_info_block}
|
||||
现在是{time_now}，你正在上网，和qq群里的网友们聊天，以下是正在进行的聊天内容：
|
||||
{chat_observe_info}
|
||||
{action_observe_info}
|
||||
以下是你之前对聊天的观察和规划，你的名字是{bot_name}：
|
||||
{last_mind}
|
||||
|
||||
现在请你继续输出观察和规划，输出要求：
|
||||
1. 先关注未读新消息的内容和近期回复历史
|
||||
2. 根据新信息，修改和删除之前的观察和规划
|
||||
3. 根据聊天内容继续输出观察和规划
|
||||
4. 注意群聊的时间线索，话题由谁发起，进展状况如何，思考聊天的时间线。
|
||||
6. 语言简洁自然，不要分点，不要浮夸，不要修辞，仅输出思考内容就好"""
|
||||
Prompt(private_prompt, "sub_heartflow_prompt_private_before")
|
||||
|
||||
|
||||
class MindProcessor(BaseProcessor):
|
||||
log_prefix = "聊天思考"
|
||||
|
||||
def __init__(self, subheartflow_id: str):
|
||||
super().__init__()
|
||||
|
||||
self.subheartflow_id = subheartflow_id
|
||||
|
||||
self.llm_model = LLMRequest(
|
||||
model=global_config.model.planner,
|
||||
request_type="focus.processor.chat_mind",
|
||||
)
|
||||
|
||||
self.current_mind = ""
|
||||
self.past_mind = []
|
||||
self.structured_info = []
|
||||
self.structured_info_str = ""
|
||||
|
||||
name = get_chat_manager().get_stream_name(self.subheartflow_id)
|
||||
self.log_prefix = f"[{name}] "
|
||||
self._update_structured_info_str()
|
||||
|
||||
def _update_structured_info_str(self):
|
||||
"""根据 structured_info 更新 structured_info_str"""
|
||||
if not self.structured_info:
|
||||
self.structured_info_str = ""
|
||||
return
|
||||
|
||||
lines = ["【信息】"]
|
||||
for item in self.structured_info:
|
||||
# 简化展示，突出内容和类型，包含TTL供调试
|
||||
type_str = item.get("type", "未知类型")
|
||||
content_str = item.get("content", "")
|
||||
|
||||
if type_str == "info":
|
||||
lines.append(f"刚刚: {content_str}")
|
||||
elif type_str == "memory":
|
||||
lines.append(f"{content_str}")
|
||||
elif type_str == "comparison_result":
|
||||
lines.append(f"数字大小比较结果: {content_str}")
|
||||
elif type_str == "time_info":
|
||||
lines.append(f"{content_str}")
|
||||
elif type_str == "lpmm_knowledge":
|
||||
lines.append(f"你知道：{content_str}")
|
||||
else:
|
||||
lines.append(f"{type_str}的信息: {content_str}")
|
||||
|
||||
self.structured_info_str = "\n".join(lines)
|
||||
logger.debug(f"{self.log_prefix} 更新 structured_info_str: \n{self.structured_info_str}")
|
||||
|
||||
async def process_info(
|
||||
self,
|
||||
observations: List[Observation] = None,
|
||||
) -> List[InfoBase]:
|
||||
"""处理信息对象
|
||||
|
||||
Args:
|
||||
*infos: 可变数量的InfoBase类型的信息对象
|
||||
|
||||
Returns:
|
||||
List[InfoBase]: 处理后的结构化信息列表
|
||||
"""
|
||||
current_mind = await self.do_thinking_before_reply(observations)
|
||||
|
||||
mind_info = MindInfo()
|
||||
mind_info.set_current_mind(current_mind)
|
||||
|
||||
return [mind_info]
|
||||
|
||||
async def do_thinking_before_reply(self, observations: List[Observation] = None):
|
||||
"""
|
||||
在回复前进行思考，生成内心想法并收集工具调用结果
|
||||
|
||||
参数:
|
||||
observations: 观察信息
|
||||
|
||||
返回:
|
||||
如果return_prompt为False:
|
||||
tuple: (current_mind, past_mind) 当前想法和过去的想法列表
|
||||
如果return_prompt为True:
|
||||
tuple: (current_mind, past_mind, prompt) 当前想法、过去的想法列表和使用的prompt
|
||||
"""
|
||||
|
||||
# ---------- 0. 更新和清理 structured_info ----------
|
||||
if self.structured_info:
|
||||
# updated_info = []
|
||||
# for item in self.structured_info:
|
||||
# item["ttl"] -= 1
|
||||
# if item["ttl"] > 0:
|
||||
# updated_info.append(item)
|
||||
# else:
|
||||
# logger.debug(f"{self.log_prefix} 移除过期的 structured_info 项: {item['id']}")
|
||||
# self.structured_info = updated_info
|
||||
self._update_structured_info_str()
|
||||
logger.debug(
|
||||
f"{self.log_prefix} 当前完整的 structured_info: {safe_json_dumps(self.structured_info, ensure_ascii=False)}"
|
||||
)
|
||||
# ---------- 1. 准备基础数据 ----------
|
||||
# 获取现有想法和情绪状态
|
||||
previous_mind = self.current_mind if self.current_mind else ""
|
||||
|
||||
if observations is None:
|
||||
observations = []
|
||||
for observation in observations:
|
||||
if isinstance(observation, ChattingObservation):
|
||||
# 获取聊天元信息
|
||||
is_group_chat = observation.is_group_chat
|
||||
chat_target_info = observation.chat_target_info
|
||||
chat_target_name = "对方" # 私聊默认名称
|
||||
if not is_group_chat and chat_target_info:
|
||||
# 优先使用person_name，其次user_nickname，最后回退到默认值
|
||||
chat_target_name = (
|
||||
chat_target_info.get("person_name") or chat_target_info.get("user_nickname") or chat_target_name
|
||||
)
|
||||
# 获取聊天内容
|
||||
chat_observe_info = observation.get_observe_info()
|
||||
person_list = observation.person_list
|
||||
if isinstance(observation, HFCloopObservation):
|
||||
hfcloop_observe_info = observation.get_observe_info()
|
||||
if isinstance(observation, ActionObservation):
|
||||
action_observe_info = observation.get_observe_info()
|
||||
|
||||
# ---------- 3. 准备个性化数据 ----------
|
||||
# 获取个性化信息
|
||||
|
||||
relation_prompt = ""
|
||||
if global_config.relationship.enable_relationship:
|
||||
for person in person_list:
|
||||
relationship_manager = get_relationship_manager()
|
||||
relation_prompt += await relationship_manager.build_relationship_info(person, is_id=True)
|
||||
|
||||
template_name = "sub_heartflow_prompt_before" if is_group_chat else "sub_heartflow_prompt_private_before"
|
||||
logger.debug(f"{self.log_prefix} 使用{'群聊' if is_group_chat else '私聊'}思考模板")
|
||||
|
||||
prompt = (await global_prompt_manager.get_prompt_async(template_name)).format(
|
||||
bot_name=get_individuality().name,
|
||||
extra_info=self.structured_info_str,
|
||||
relation_prompt=relation_prompt,
|
||||
time_now=time.strftime("%Y-%m-%d %H:%M:%S", time.localtime()),
|
||||
chat_observe_info=chat_observe_info,
|
||||
last_mind=previous_mind,
|
||||
cycle_info_block=hfcloop_observe_info,
|
||||
action_observe_info=action_observe_info,
|
||||
chat_target_name=chat_target_name,
|
||||
)
|
||||
|
||||
content = "(不知道该想些什么...)"
|
||||
try:
|
||||
content, _ = await self.llm_model.generate_response_async(prompt=prompt)
|
||||
if not content:
|
||||
logger.warning(f"{self.log_prefix} LLM返回空结果，思考失败。")
|
||||
except Exception as e:
|
||||
# 处理总体异常
|
||||
logger.error(f"{self.log_prefix} 执行LLM请求或处理响应时出错: {e}")
|
||||
logger.error(traceback.format_exc())
|
||||
content = "注意：思考过程中出现错误，应该是LLM大模型有问题！！你需要告诉别人，检查大模型配置"
|
||||
|
||||
# 记录初步思考结果
|
||||
logger.debug(f"{self.log_prefix} 思考prompt: \n{prompt}\n")
|
||||
logger.info(f"{self.log_prefix} 聊天规划: {content}")
|
||||
self.update_current_mind(content)
|
||||
|
||||
return content
|
||||
|
||||
def update_current_mind(self, response):
|
||||
if self.current_mind: # 只有当 current_mind 非空时才添加到 past_mind
|
||||
self.past_mind.append(self.current_mind)
|
||||
self.current_mind = response
|
||||
|
||||
|
||||
init_prompt()
|
||||
@@ -644,7 +644,7 @@ class PersonImpressionpProcessor(BaseProcessor):
|
||||
)
|
||||
|
||||
try:
|
||||
logger.info(f"{self.log_prefix} 人物信息prompt: \n{prompt}\n")
|
||||
logger.debug(f"{self.log_prefix} 人物信息prompt: \n{prompt}\n")
|
||||
content, _ = await self.llm_model.generate_response_async(prompt=prompt)
|
||||
if content:
|
||||
# print(f"content: {content}")
|
||||
@@ -704,26 +704,47 @@ class PersonImpressionpProcessor(BaseProcessor):
|
||||
persons_infos_str = ""
|
||||
# 处理已获取到的信息
|
||||
if self.info_fetched_cache:
|
||||
persons_with_known_info = [] # 有已知信息的人员
|
||||
persons_with_unknown_info = [] # 有未知信息的人员
|
||||
|
||||
for person_id in self.info_fetched_cache:
|
||||
person_infos_str = ""
|
||||
unknown_info_types = [] # 收集所有unknow的信息类型
|
||||
person_known_infos = []
|
||||
person_unknown_infos = []
|
||||
person_name = ""
|
||||
|
||||
for info_type in self.info_fetched_cache[person_id]:
|
||||
person_name = self.info_fetched_cache[person_id][info_type]["person_name"]
|
||||
if not self.info_fetched_cache[person_id][info_type]["unknow"]:
|
||||
info_content = self.info_fetched_cache[person_id][info_type]["info"]
|
||||
person_infos_str += f"[{info_type}]：{info_content}；"
|
||||
person_known_infos.append(f"[{info_type}]：{info_content}")
|
||||
else:
|
||||
unknown_info_types.append(info_type)
|
||||
person_unknown_infos.append(info_type)
|
||||
|
||||
# 如果有unknow的信息类型，合并输出
|
||||
if unknown_info_types:
|
||||
unknown_types_str = "、".join(unknown_info_types)
|
||||
person_infos_str += f"你不了解{person_name}有关[{unknown_types_str}]的信息，不要胡乱回答，你可以直接说你不知道，或者你忘记了；"
|
||||
# 如果有已知信息，添加到已知信息列表
|
||||
if person_known_infos:
|
||||
known_info_str = "；".join(person_known_infos) + "；"
|
||||
persons_with_known_info.append((person_name, known_info_str))
|
||||
|
||||
if person_infos_str:
|
||||
persons_infos_str += f"你对 {person_name} 的了解：{person_infos_str}\n"
|
||||
# 如果有未知信息，添加到未知信息列表
|
||||
if person_unknown_infos:
|
||||
persons_with_unknown_info.append((person_name, person_unknown_infos))
|
||||
|
||||
# 先输出有已知信息的人员
|
||||
for person_name, known_info_str in persons_with_known_info:
|
||||
persons_infos_str += f"你对 {person_name} 的了解：{known_info_str}\n"
|
||||
|
||||
# 统一处理未知信息，避免重复的警告文本
|
||||
if persons_with_unknown_info:
|
||||
unknown_persons_details = []
|
||||
for person_name, unknown_types in persons_with_unknown_info:
|
||||
unknown_types_str = "、".join(unknown_types)
|
||||
unknown_persons_details.append(f"{person_name}的[{unknown_types_str}]")
|
||||
|
||||
if len(unknown_persons_details) == 1:
|
||||
persons_infos_str += f"你不了解{unknown_persons_details[0]}信息，不要胡乱回答，可以直接说不知道或忘记了；\n"
|
||||
else:
|
||||
unknown_all_str = "、".join(unknown_persons_details)
|
||||
persons_infos_str += f"你不了解{unknown_all_str}等信息，不要胡乱回答，可以直接说不知道或忘记了；\n"
|
||||
|
||||
return persons_infos_str
|
||||
|
||||
|
||||
Reference in New Issue
Block a user