128 lines
5.0 KiB
Python
128 lines
5.0 KiB
Python
from typing import Tuple
|
||
from src.common.logger import get_module_logger
|
||
from ..models.utils_model import LLM_request
|
||
from ..config.config import global_config
|
||
from .chat_observer import ChatObserver
|
||
from .pfc_utils import get_items_from_json
|
||
from src.individuality.individuality import Individuality
|
||
from .observation_info import ObservationInfo
|
||
from .conversation_info import ConversationInfo
|
||
|
||
logger = get_module_logger("action_planner")
|
||
|
||
|
||
class ActionPlannerInfo:
|
||
def __init__(self):
|
||
self.done_action = []
|
||
self.goal_list = []
|
||
self.knowledge_list = []
|
||
self.memory_list = []
|
||
|
||
|
||
class ActionPlanner:
|
||
"""行动规划器"""
|
||
|
||
def __init__(self, stream_id: str):
|
||
self.llm = LLM_request(
|
||
model=global_config.llm_normal, temperature=0.7, max_tokens=1000, request_type="action_planning"
|
||
)
|
||
self.personality_info = Individuality.get_instance().get_prompt(type="personality", x_person=2, level=2)
|
||
self.name = global_config.BOT_NICKNAME
|
||
self.chat_observer = ChatObserver.get_instance(stream_id)
|
||
|
||
async def plan(self, observation_info: ObservationInfo, conversation_info: ConversationInfo) -> Tuple[str, str]:
|
||
"""规划下一步行动
|
||
|
||
Args:
|
||
observation_info: 决策信息
|
||
conversation_info: 对话信息
|
||
|
||
Returns:
|
||
Tuple[str, str]: (行动类型, 行动原因)
|
||
"""
|
||
# 构建提示词
|
||
logger.debug(f"开始规划行动：当前目标: {conversation_info.goal_list}")
|
||
|
||
# 构建对话目标
|
||
if conversation_info.goal_list:
|
||
goal, reasoning = conversation_info.goal_list[-1]
|
||
else:
|
||
goal = "目前没有明确对话目标"
|
||
reasoning = "目前没有明确对话目标，最好思考一个对话目标"
|
||
|
||
# 获取聊天历史记录
|
||
chat_history_list = observation_info.chat_history
|
||
chat_history_text = ""
|
||
for msg in chat_history_list:
|
||
chat_history_text += f"{msg}\n"
|
||
|
||
if observation_info.new_messages_count > 0:
|
||
new_messages_list = observation_info.unprocessed_messages
|
||
|
||
chat_history_text += f"有{observation_info.new_messages_count}条新消息：\n"
|
||
for msg in new_messages_list:
|
||
chat_history_text += f"{msg}\n"
|
||
|
||
observation_info.clear_unprocessed_messages()
|
||
|
||
personality_text = f"你的名字是{self.name}，{self.personality_info}"
|
||
|
||
# 构建action历史文本
|
||
action_history_list = conversation_info.done_action
|
||
action_history_text = "你之前做的事情是："
|
||
for action in action_history_list:
|
||
action_history_text += f"{action}\n"
|
||
|
||
prompt = f"""{personality_text}。现在你在参与一场QQ聊天，请分析以下内容，根据信息决定下一步行动：
|
||
|
||
当前对话目标：{goal}
|
||
产生该对话目标的原因：{reasoning}
|
||
|
||
{action_history_text}
|
||
|
||
最近的对话记录：
|
||
{chat_history_text}
|
||
|
||
请你接下去想想要你要做什么，可以发言，可以等待，可以倾听，可以调取知识。注意不同行动类型的要求，不要重复发言：
|
||
行动类型：
|
||
fetch_knowledge: 需要调取知识，当需要专业知识或特定信息时选择
|
||
wait: 当你做出了发言,对方尚未回复时等待对方的回复
|
||
listening: 倾听对方发言，当你认为对方发言尚未结束时采用
|
||
direct_reply: 不符合上述情况，回复对方，注意不要过多或者重复发言
|
||
rethink_goal: 重新思考对话目标，当发现对话目标不合适时选择，会重新思考对话目标
|
||
|
||
请以JSON格式输出，包含以下字段：
|
||
1. action: 行动类型，注意你之前的行为
|
||
2. reason: 选择该行动的原因，注意你之前的行为（简要解释）
|
||
|
||
注意：请严格按照JSON格式输出，不要包含任何其他内容。"""
|
||
|
||
logger.debug(f"发送到LLM的提示词: {prompt}")
|
||
try:
|
||
content, _ = await self.llm.generate_response_async(prompt)
|
||
logger.debug(f"LLM原始返回内容: {content}")
|
||
|
||
# 使用简化函数提取JSON内容
|
||
success, result = get_items_from_json(
|
||
content, "action", "reason", default_values={"action": "direct_reply", "reason": "没有明确原因"}
|
||
)
|
||
|
||
if not success:
|
||
return "direct_reply", "JSON解析失败，选择直接回复"
|
||
|
||
action = result["action"]
|
||
reason = result["reason"]
|
||
|
||
# 验证action类型
|
||
if action not in ["direct_reply", "fetch_knowledge", "wait", "listening", "rethink_goal"]:
|
||
logger.warning(f"未知的行动类型: {action}，默认使用listening")
|
||
action = "listening"
|
||
|
||
logger.info(f"规划的行动: {action}")
|
||
logger.info(f"行动原因: {reason}")
|
||
return action, reason
|
||
|
||
except Exception as e:
|
||
logger.error(f"规划行动时出错: {str(e)}")
|
||
return "direct_reply", "发生错误，选择直接回复"
|