mirror of https://github.com/Mai-with-u/MaiBot.git
feat：非核心动作已从planner分离到副planner，移除action的激活机制，添加模型
SengokuCola
parent
1ddfa47e68
commit
b525e1e098
|
|
@ -29,6 +29,7 @@ from src.plugin_system.core import events_manager
|
|||
from src.plugin_system.apis import generator_api, send_api, message_api, database_api
|
||||
from src.mais4u.mai_think import mai_thinking_manager
|
||||
from src.mais4u.s4u_config import s4u_config
|
||||
from src.chat.utils.chat_message_builder import build_readable_messages_with_id, build_readable_actions, get_actions_by_timestamp_with_chat, get_raw_msg_before_timestamp_with_chat
|
||||
|
||||
if TYPE_CHECKING:
|
||||
from src.common.data_models.database_data_model import DatabaseMessages
|
||||
|
|
@ -402,8 +403,8 @@ class HeartFChatting:
|
|||
)
|
||||
]
|
||||
else:
|
||||
# 第一步：动作修改
|
||||
with Timer("动作修改", cycle_timers):
|
||||
# 第一步：动作检查
|
||||
with Timer("动作检查", cycle_timers):
|
||||
try:
|
||||
await self.action_modifier.modify_actions()
|
||||
available_actions = self.action_manager.get_using_actions()
|
||||
|
|
@ -412,10 +413,45 @@ class HeartFChatting:
|
|||
|
||||
# 执行planner
|
||||
planner_info = self.action_planner.get_necessary_info()
|
||||
|
||||
|
||||
|
||||
|
||||
|
||||
message_list_before_now = get_raw_msg_before_timestamp_with_chat(
|
||||
chat_id=self.stream_id,
|
||||
timestamp=time.time(),
|
||||
limit=int(global_config.chat.max_context_size * 0.6),
|
||||
)
|
||||
chat_content_block, message_id_list = build_readable_messages_with_id(
|
||||
messages=message_list_before_now,
|
||||
timestamp_mode="normal_no_YMD",
|
||||
read_mark=self.action_planner.last_obs_time_mark,
|
||||
truncate=True,
|
||||
show_actions=True,
|
||||
)
|
||||
|
||||
actions_before_now = get_actions_by_timestamp_with_chat(
|
||||
chat_id=self.stream_id,
|
||||
timestamp_start=time.time() - 600,
|
||||
timestamp_end=time.time(),
|
||||
limit=5,
|
||||
)
|
||||
|
||||
actions_before_now_block = build_readable_actions(
|
||||
actions=actions_before_now,
|
||||
)
|
||||
|
||||
|
||||
|
||||
|
||||
prompt_info = await self.action_planner.build_planner_prompt(
|
||||
is_group_chat=planner_info[0],
|
||||
chat_target_info=planner_info[1],
|
||||
current_available_actions=planner_info[2],
|
||||
chat_content_block=chat_content_block,
|
||||
actions_before_now_block=actions_before_now_block,
|
||||
message_id_list=message_id_list,
|
||||
)
|
||||
if not await events_manager.handle_mai_events(
|
||||
EventType.ON_PLAN, None, prompt_info[0], None, self.chat_stream.stream_id
|
||||
|
|
@ -427,6 +463,9 @@ class HeartFChatting:
|
|||
loop_start_time=self.last_read_time,
|
||||
available_actions=available_actions,
|
||||
)
|
||||
|
||||
for action in action_to_use_info:
|
||||
print(action.action_type)
|
||||
|
||||
# 3. 并行执行所有动作
|
||||
action_tasks = [
|
||||
|
|
|
|||
|
|
@ -60,7 +60,7 @@ class ActionModifier:
|
|||
|
||||
removals_s1: List[Tuple[str, str]] = []
|
||||
removals_s2: List[Tuple[str, str]] = []
|
||||
removals_s3: List[Tuple[str, str]] = []
|
||||
# removals_s3: List[Tuple[str, str]] = []
|
||||
|
||||
self.action_manager.restore_actions()
|
||||
all_actions = self.action_manager.get_using_actions()
|
||||
|
|
@ -103,26 +103,28 @@ class ActionModifier:
|
|||
self.action_manager.remove_action_from_using(action_name)
|
||||
logger.debug(f"{self.log_prefix}阶段二移除动作: {action_name}，原因: {reason}")
|
||||
|
||||
|
||||
|
||||
# === 第三阶段：激活类型判定 ===
|
||||
if chat_content is not None:
|
||||
logger.debug(f"{self.log_prefix}开始激活类型判定阶段")
|
||||
# if chat_content is not None:
|
||||
# logger.debug(f"{self.log_prefix}开始激活类型判定阶段")
|
||||
|
||||
# 获取当前使用的动作集（经过第一阶段处理）
|
||||
current_using_actions = self.action_manager.get_using_actions()
|
||||
# current_using_actions = self.action_manager.get_using_actions()
|
||||
|
||||
# 获取因激活类型判定而需要移除的动作
|
||||
removals_s3 = await self._get_deactivated_actions_by_type(
|
||||
current_using_actions,
|
||||
chat_content,
|
||||
)
|
||||
# removals_s3 = await self._get_deactivated_actions_by_type(
|
||||
# current_using_actions,
|
||||
# chat_content,
|
||||
# )
|
||||
|
||||
# 应用第三阶段的移除
|
||||
for action_name, reason in removals_s3:
|
||||
self.action_manager.remove_action_from_using(action_name)
|
||||
logger.debug(f"{self.log_prefix}阶段三移除动作: {action_name}，原因: {reason}")
|
||||
# for action_name, reason in removals_s3:
|
||||
# self.action_manager.remove_action_from_using(action_name)
|
||||
# logger.debug(f"{self.log_prefix}阶段三移除动作: {action_name}，原因: {reason}")
|
||||
|
||||
# === 统一日志记录 ===
|
||||
all_removals = removals_s1 + removals_s2 + removals_s3
|
||||
all_removals = removals_s1 + removals_s2
|
||||
removals_summary: str = ""
|
||||
if all_removals:
|
||||
removals_summary = " | ".join([f"{name}({reason})" for name, reason in all_removals])
|
||||
|
|
|
|||
|
|
@ -21,8 +21,9 @@ from src.chat.utils.chat_message_builder import (
|
|||
from src.chat.utils.utils import get_chat_type_and_target_info
|
||||
from src.chat.planner_actions.action_manager import ActionManager
|
||||
from src.chat.message_receive.chat_stream import get_chat_manager
|
||||
from src.plugin_system.base.component_types import ActionInfo, ChatMode, ComponentType
|
||||
from src.plugin_system.base.component_types import ActionInfo, ChatMode, ComponentType, ActionActivationType
|
||||
from src.plugin_system.core.component_registry import component_registry
|
||||
import random
|
||||
|
||||
logger = get_logger("planner")
|
||||
|
||||
|
|
@ -82,6 +83,36 @@ def init_prompt():
|
|||
""",
|
||||
"action_prompt",
|
||||
)
|
||||
|
||||
|
||||
Prompt(
|
||||
"""
|
||||
{time_block}
|
||||
{name_block}
|
||||
请你根据聊天内容，选择一个或多个action来参与聊天。如果没有合适的action，请选择no_action。
|
||||
|
||||
{chat_context_description}，以下是具体的聊天内容
|
||||
{chat_content_block}
|
||||
|
||||
{moderation_prompt}
|
||||
|
||||
现在请你根据聊天内容和用户的最新消息选择合适的action和触发action的消息:
|
||||
{actions_before_now_block}
|
||||
|
||||
no_action：不选择任何动作
|
||||
{{
|
||||
"action": "no_action",
|
||||
"reason":"不动作的原因"
|
||||
}}
|
||||
|
||||
{action_options_text}
|
||||
|
||||
请选择，并说明触发action的消息id和选择该action的原因。消息id格式:m+数字
|
||||
|
||||
请根据动作示例，以严格的 JSON 格式输出，且仅包含 JSON 内容：
|
||||
""",
|
||||
"sub_planner_prompt",
|
||||
)
|
||||
|
||||
|
||||
class ActionPlanner:
|
||||
|
|
@ -93,6 +124,9 @@ class ActionPlanner:
|
|||
self.planner_llm = LLMRequest(
|
||||
model_set=model_config.model_task_config.planner, request_type="planner"
|
||||
) # 用于动作规划
|
||||
self.planner_small_llm = LLMRequest(
|
||||
model_set=model_config.model_task_config.planner_small, request_type="planner_small"
|
||||
) # 用于动作规划
|
||||
|
||||
self.last_obs_time_mark = 0.0
|
||||
# 添加重试计数器
|
||||
|
|
@ -100,7 +134,7 @@ class ActionPlanner:
|
|||
self.max_plan_retries = 3
|
||||
|
||||
def find_message_by_id(
|
||||
self, message_id: str, message_id_list: List[DatabaseMessages]
|
||||
self, message_id: str, message_id_list: List[Tuple[str, DatabaseMessages]]
|
||||
) -> Optional[DatabaseMessages]:
|
||||
# sourcery skip: use-next
|
||||
"""
|
||||
|
|
@ -114,8 +148,8 @@ class ActionPlanner:
|
|||
找到的原始消息字典，如果未找到则返回None
|
||||
"""
|
||||
for item in message_id_list:
|
||||
if item.message_id == message_id:
|
||||
return item
|
||||
if item[0] == message_id:
|
||||
return item[1]
|
||||
return None
|
||||
|
||||
def get_latest_message(self, message_id_list: List[DatabaseMessages]) -> Optional[DatabaseMessages]:
|
||||
|
|
@ -129,6 +163,247 @@ class ActionPlanner:
|
|||
最新的消息字典，如果列表为空则返回None
|
||||
"""
|
||||
return message_id_list[-1] if message_id_list else None
|
||||
|
||||
def _parse_single_action(self, action_json: dict, message_id_list: List[Tuple[str, DatabaseMessages]], current_available_actions: List[Tuple[str, ActionInfo]]) -> List[ActionPlannerInfo]:
|
||||
"""解析单个action JSON并返回ActionPlannerInfo列表"""
|
||||
action_planner_infos = []
|
||||
|
||||
try:
|
||||
action = action_json.get("action", "no_action")
|
||||
reasoning = action_json.get("reason", "未提供原因")
|
||||
action_data = {}
|
||||
|
||||
# 将所有其他属性添加到action_data
|
||||
for key, value in action_json.items():
|
||||
if key not in ["action", "reasoning"]:
|
||||
action_data[key] = value
|
||||
|
||||
# 非no_action动作需要target_message_id
|
||||
target_message = None
|
||||
if action != "no_action":
|
||||
if target_message_id := action_json.get("target_message_id"):
|
||||
# 根据target_message_id查找原始消息
|
||||
target_message = self.find_message_by_id(target_message_id, message_id_list)
|
||||
if target_message is None:
|
||||
logger.warning(f"{self.log_prefix}无法找到target_message_id '{target_message_id}' 对应的消息")
|
||||
# 选择最新消息作为target_message
|
||||
target_message = self.get_latest_message(message_id_list)
|
||||
else:
|
||||
logger.warning(f"{self.log_prefix}动作'{action}'缺少target_message_id")
|
||||
|
||||
# 验证action是否可用
|
||||
available_action_names = [action_name for action_name, _ in current_available_actions]
|
||||
if action != "no_action" and action != "reply" and action not in available_action_names:
|
||||
logger.warning(
|
||||
f"{self.log_prefix}LLM 返回了当前不可用或无效的动作: '{action}' (可用: {available_action_names})，将强制使用 'no_action'"
|
||||
)
|
||||
reasoning = f"LLM 返回了当前不可用的动作 '{action}' (可用: {available_action_names})。原始理由: {reasoning}"
|
||||
action = "no_action"
|
||||
|
||||
# 创建ActionPlannerInfo对象
|
||||
# 将列表转换为字典格式
|
||||
available_actions_dict = dict(current_available_actions)
|
||||
action_planner_infos.append(ActionPlannerInfo(
|
||||
action_type=action,
|
||||
reasoning=reasoning,
|
||||
action_data=action_data,
|
||||
action_message=target_message,
|
||||
available_actions=available_actions_dict,
|
||||
))
|
||||
|
||||
except Exception as e:
|
||||
logger.error(f"{self.log_prefix}解析单个action时出错: {e}")
|
||||
# 将列表转换为字典格式
|
||||
available_actions_dict = dict(current_available_actions)
|
||||
action_planner_infos.append(ActionPlannerInfo(
|
||||
action_type="no_action",
|
||||
reasoning=f"解析单个action时出错: {e}",
|
||||
action_data={},
|
||||
action_message=None,
|
||||
available_actions=available_actions_dict,
|
||||
))
|
||||
|
||||
return action_planner_infos
|
||||
|
||||
async def sub_plan(
|
||||
self,
|
||||
action_list: List[Tuple[str, ActionInfo]],
|
||||
actions_before_now_block: str,
|
||||
chat_content_block: str,
|
||||
message_id_list: List[Tuple[str, DatabaseMessages]],
|
||||
is_group_chat: bool = False,
|
||||
chat_target_info: Optional[dict] = None,
|
||||
# current_available_actions: Dict[str, ActionInfo] = {},
|
||||
) -> List[ActionPlannerInfo]:
|
||||
# 构建副planner并执行(单个副planner)
|
||||
try:
|
||||
if actions_before_now_block:
|
||||
actions_before_now_block = f"你刚刚选择并执行过的action是：\n{actions_before_now_block}"
|
||||
else:
|
||||
actions_before_now_block = ""
|
||||
|
||||
chat_context_description = "你现在正在一个群聊中"
|
||||
chat_target_name = None
|
||||
if not is_group_chat and chat_target_info:
|
||||
chat_target_name = (
|
||||
chat_target_info.get("person_name") or chat_target_info.get("user_nickname") or "对方"
|
||||
)
|
||||
chat_context_description = f"你正在和 {chat_target_name} 私聊"
|
||||
|
||||
action_options_block = ""
|
||||
|
||||
for using_actions_name, using_actions_info in action_list:
|
||||
if using_actions_info.action_parameters:
|
||||
param_text = "\n"
|
||||
for param_name, param_description in using_actions_info.action_parameters.items():
|
||||
param_text += f' "{param_name}":"{param_description}"\n'
|
||||
param_text = param_text.rstrip("\n")
|
||||
else:
|
||||
param_text = ""
|
||||
|
||||
require_text = ""
|
||||
for require_item in using_actions_info.action_require:
|
||||
require_text += f"- {require_item}\n"
|
||||
require_text = require_text.rstrip("\n")
|
||||
|
||||
using_action_prompt = await global_prompt_manager.get_prompt_async("action_prompt")
|
||||
using_action_prompt = using_action_prompt.format(
|
||||
action_name=using_actions_name,
|
||||
action_description=using_actions_info.description,
|
||||
action_parameters=param_text,
|
||||
action_require=require_text,
|
||||
)
|
||||
|
||||
action_options_block += using_action_prompt
|
||||
|
||||
moderation_prompt_block = "请不要输出违法违规内容，不要输出色情，暴力，政治相关内容，如有敏感内容，请规避。"
|
||||
time_block = f"当前时间：{datetime.now().strftime('%Y-%m-%d %H:%M:%S')}"
|
||||
bot_name = global_config.bot.nickname
|
||||
if global_config.bot.alias_names:
|
||||
bot_nickname = f",也有人叫你{','.join(global_config.bot.alias_names)}"
|
||||
else:
|
||||
bot_nickname = ""
|
||||
name_block = f"你的名字是{bot_name}{bot_nickname}，请注意哪些是你自己的发言。"
|
||||
|
||||
planner_prompt_template = await global_prompt_manager.get_prompt_async("sub_planner_prompt")
|
||||
prompt = planner_prompt_template.format(
|
||||
time_block=time_block,
|
||||
chat_context_description=chat_context_description,
|
||||
chat_content_block=chat_content_block,
|
||||
actions_before_now_block=actions_before_now_block,
|
||||
action_options_text=action_options_block,
|
||||
moderation_prompt=moderation_prompt_block,
|
||||
name_block=name_block,
|
||||
)
|
||||
# return prompt, message_id_list
|
||||
except Exception as e:
|
||||
logger.error(f"构建 Planner 提示词时出错: {e}")
|
||||
logger.error(traceback.format_exc())
|
||||
return "构建 Planner Prompt 时出错", []
|
||||
|
||||
# --- 调用 LLM (普通文本生成) ---
|
||||
llm_content = None
|
||||
action_planner_infos = [] # 存储多个ActionPlannerInfo对象
|
||||
|
||||
try:
|
||||
llm_content, (reasoning_content, _, _) = await self.planner_llm.generate_response_async(prompt=prompt)
|
||||
|
||||
if global_config.debug.show_prompt:
|
||||
logger.info(f"{self.log_prefix}副规划器原始提示词: {prompt}")
|
||||
logger.info(f"{self.log_prefix}副规划器原始响应: {llm_content}")
|
||||
if reasoning_content:
|
||||
logger.info(f"{self.log_prefix}副规划器推理: {reasoning_content}")
|
||||
else:
|
||||
logger.debug(f"{self.log_prefix}副规划器原始提示词: {prompt}")
|
||||
logger.debug(f"{self.log_prefix}副规划器原始响应: {llm_content}")
|
||||
if reasoning_content:
|
||||
logger.debug(f"{self.log_prefix}副规划器推理: {reasoning_content}")
|
||||
|
||||
except Exception as req_e:
|
||||
logger.error(f"{self.log_prefix}副规划器LLM 请求执行失败: {req_e}")
|
||||
# 返回一个默认的no_action
|
||||
action_planner_infos.append(ActionPlannerInfo(
|
||||
action_type="no_action",
|
||||
reasoning=f"副规划器LLM 请求失败，模型出现问题: {req_e}",
|
||||
action_data={},
|
||||
action_message=None,
|
||||
available_actions=action_list,
|
||||
))
|
||||
return action_planner_infos
|
||||
|
||||
if llm_content:
|
||||
try:
|
||||
parsed_json = json.loads(repair_json(llm_content))
|
||||
|
||||
# 处理不同的JSON格式
|
||||
if isinstance(parsed_json, list):
|
||||
# 如果是列表，处理每个action
|
||||
if parsed_json:
|
||||
logger.info(f"{self.log_prefix}LLM返回了{len(parsed_json)}个action")
|
||||
for action_item in parsed_json:
|
||||
if isinstance(action_item, dict):
|
||||
action_planner_infos.extend(self._parse_single_action(
|
||||
action_item, message_id_list, action_list
|
||||
))
|
||||
else:
|
||||
logger.warning(f"{self.log_prefix}列表中的action项不是字典类型: {type(action_item)}")
|
||||
else:
|
||||
logger.warning(f"{self.log_prefix}LLM返回了空列表")
|
||||
action_planner_infos.append(ActionPlannerInfo(
|
||||
action_type="no_action",
|
||||
reasoning="LLM返回了空列表，选择no_action",
|
||||
action_data={},
|
||||
action_message=None,
|
||||
available_actions=action_list,
|
||||
))
|
||||
elif isinstance(parsed_json, dict):
|
||||
# 如果是单个字典，处理单个action
|
||||
action_planner_infos.extend(self._parse_single_action(
|
||||
parsed_json, message_id_list, action_list
|
||||
))
|
||||
else:
|
||||
logger.error(f"{self.log_prefix}解析后的JSON不是字典或列表类型: {type(parsed_json)}")
|
||||
action_planner_infos.append(ActionPlannerInfo(
|
||||
action_type="no_action",
|
||||
reasoning=f"解析后的JSON类型错误: {type(parsed_json)}",
|
||||
action_data={},
|
||||
action_message=None,
|
||||
available_actions=action_list,
|
||||
))
|
||||
|
||||
except Exception as json_e:
|
||||
logger.warning(f"{self.log_prefix}解析LLM响应JSON失败 {json_e}. LLM原始输出: '{llm_content}'")
|
||||
traceback.print_exc()
|
||||
action_planner_infos.append(ActionPlannerInfo(
|
||||
action_type="no_action",
|
||||
reasoning=f"解析LLM响应JSON失败: {json_e}. 将使用默认动作 'no_action'.",
|
||||
action_data={},
|
||||
action_message=None,
|
||||
available_actions=action_list,
|
||||
))
|
||||
else:
|
||||
# 如果没有LLM内容，返回默认的no_action
|
||||
action_planner_infos.append(ActionPlannerInfo(
|
||||
action_type="no_action",
|
||||
reasoning="副规划器没有获得LLM响应",
|
||||
action_data={},
|
||||
action_message=None,
|
||||
available_actions=action_list,
|
||||
))
|
||||
|
||||
# 如果没有解析到任何action，返回默认的no_action
|
||||
if not action_planner_infos:
|
||||
action_planner_infos.append(ActionPlannerInfo(
|
||||
action_type="no_action",
|
||||
reasoning="副规划器没有解析到任何有效action",
|
||||
action_data={},
|
||||
action_message=None,
|
||||
available_actions=action_list,
|
||||
))
|
||||
|
||||
logger.info(f"{self.log_prefix}副规划器返回了{len(action_planner_infos)}个action")
|
||||
return action_planner_infos
|
||||
|
||||
|
||||
async def plan(
|
||||
self,
|
||||
|
|
@ -147,17 +422,144 @@ class ActionPlanner:
|
|||
target_message: Optional[DatabaseMessages] = None # 初始化target_message变量
|
||||
prompt: str = ""
|
||||
message_id_list: list = []
|
||||
|
||||
message_list_before_now = get_raw_msg_before_timestamp_with_chat(
|
||||
chat_id=self.chat_id,
|
||||
timestamp=time.time(),
|
||||
limit=int(global_config.chat.max_context_size * 0.6),
|
||||
)
|
||||
chat_content_block, message_id_list = build_readable_messages_with_id(
|
||||
messages=message_list_before_now,
|
||||
timestamp_mode="normal_no_YMD",
|
||||
read_mark=self.last_obs_time_mark,
|
||||
truncate=True,
|
||||
show_actions=True,
|
||||
)
|
||||
|
||||
actions_before_now = get_actions_by_timestamp_with_chat(
|
||||
chat_id=self.chat_id,
|
||||
timestamp_start=time.time() - 600,
|
||||
timestamp_end=time.time(),
|
||||
limit=5,
|
||||
)
|
||||
|
||||
actions_before_now_block = build_readable_actions(
|
||||
actions=actions_before_now,
|
||||
)
|
||||
|
||||
|
||||
message_list_before_now_short = message_list_before_now[:5]
|
||||
|
||||
chat_content_block_short, message_id_list_short = build_readable_messages_with_id(
|
||||
messages=message_list_before_now_short,
|
||||
timestamp_mode="normal_no_YMD",
|
||||
truncate=False,
|
||||
show_actions=False,
|
||||
)
|
||||
|
||||
self.last_obs_time_mark = time.time()
|
||||
|
||||
try:
|
||||
logger.info(f"{self.log_prefix}开始构建副Planner")
|
||||
sub_planner_actions = {}
|
||||
|
||||
for action_name, action_info in available_actions.items():
|
||||
if action_info.activation_type == ActionActivationType.LLM_JUDGE or action_info.activation_type == ActionActivationType.ALWAYS:
|
||||
sub_planner_actions[action_name] = action_info
|
||||
elif action_info.activation_type == ActionActivationType.RANDOM:
|
||||
if random.random() < action_info.random_activation_probability:
|
||||
sub_planner_actions[action_name] = action_info
|
||||
elif action_info.activation_type == ActionActivationType.KEYWORD:
|
||||
if action_info.activation_keywords:
|
||||
for keyword in action_info.activation_keywords:
|
||||
if keyword in chat_content_block_short:
|
||||
sub_planner_actions[action_name] = action_info
|
||||
elif action_info.activation_type == ActionActivationType.NEVER:
|
||||
pass
|
||||
else:
|
||||
logger.warning(f"{self.log_prefix}未知的激活类型: {action_info.activation_type}，跳过处理")
|
||||
|
||||
sub_planner_actions_num = len(sub_planner_actions)
|
||||
sub_planner_size = global_config.chat.planner_size
|
||||
if global_config.chat.planner_size > int(global_config.chat.planner_size):
|
||||
if random.random() < global_config.chat.planner_size - int(global_config.chat.planner_size):
|
||||
sub_planner_size = int(global_config.chat.planner_size) + 1
|
||||
sub_planner_num = int(sub_planner_actions_num / sub_planner_size)
|
||||
if sub_planner_actions_num % sub_planner_size != 0:
|
||||
sub_planner_num += 1
|
||||
|
||||
logger.info(f"{self.log_prefix}副规划器数量: {sub_planner_num}, 副规划器大小: {sub_planner_size}")
|
||||
|
||||
# 将sub_planner_actions随机分配到sub_planner_num个List中
|
||||
sub_planner_lists = []
|
||||
if sub_planner_actions_num > 0:
|
||||
# 将actions转换为列表并随机打乱
|
||||
action_items = list(sub_planner_actions.items())
|
||||
random.shuffle(action_items)
|
||||
|
||||
# 初始化所有子列表
|
||||
for i in range(sub_planner_num):
|
||||
sub_planner_lists.append([])
|
||||
|
||||
# 分配actions到各个子列表
|
||||
for i, (action_name, action_info) in enumerate(action_items):
|
||||
# 确保每个列表至少有一个action
|
||||
if i < sub_planner_num:
|
||||
sub_planner_lists[i].append((action_name, action_info))
|
||||
else:
|
||||
# 随机选择一个列表添加action，但不超过最大大小限制
|
||||
available_lists = [j for j, lst in enumerate(sub_planner_lists)
|
||||
if len(lst) < sub_planner_size]
|
||||
if available_lists:
|
||||
target_list = random.choice(available_lists)
|
||||
sub_planner_lists[target_list].append((action_name, action_info))
|
||||
|
||||
logger.info(f"{self.log_prefix}成功将{len(sub_planner_actions)}个actions分配到{sub_planner_num}个子列表中")
|
||||
for i, lst in enumerate(sub_planner_lists):
|
||||
logger.debug(f"{self.log_prefix}子列表{i+1}: {len(lst)}个actions")
|
||||
else:
|
||||
logger.info(f"{self.log_prefix}没有可用的actions需要分配")
|
||||
|
||||
|
||||
# 先获取必要信息
|
||||
is_group_chat, chat_target_info, current_available_actions = self.get_necessary_info()
|
||||
|
||||
# 并行执行所有副规划器
|
||||
import asyncio
|
||||
|
||||
async def execute_sub_plan(action_list):
|
||||
return await self.sub_plan(
|
||||
action_list=action_list,
|
||||
actions_before_now_block=actions_before_now_block,
|
||||
chat_content_block=chat_content_block_short,
|
||||
message_id_list=message_id_list_short,
|
||||
is_group_chat=is_group_chat,
|
||||
chat_target_info=chat_target_info,
|
||||
# current_available_actions=current_available_actions,
|
||||
)
|
||||
|
||||
# 创建所有任务
|
||||
sub_plan_tasks = [execute_sub_plan(action_list) for action_list in sub_planner_lists]
|
||||
|
||||
# 并行执行所有任务
|
||||
sub_plan_results = await asyncio.gather(*sub_plan_tasks)
|
||||
|
||||
# 收集所有结果
|
||||
all_sub_planner_results = []
|
||||
for sub_result in sub_plan_results:
|
||||
all_sub_planner_results.extend(sub_result)
|
||||
|
||||
logger.info(f"{self.log_prefix}所有副规划器共返回了{len(all_sub_planner_results)}个action")
|
||||
|
||||
# --- 构建提示词 (调用修改后的 PromptBuilder 方法) ---
|
||||
prompt, message_id_list = await self.build_planner_prompt(
|
||||
is_group_chat=is_group_chat, # <-- Pass HFC state
|
||||
chat_target_info=chat_target_info, # <-- 传递获取到的聊天目标信息
|
||||
current_available_actions=current_available_actions, # <-- Pass determined actions
|
||||
current_available_actions="", # <-- Pass determined actions
|
||||
mode=mode,
|
||||
refresh_time=True,
|
||||
chat_content_block=chat_content_block,
|
||||
actions_before_now_block=actions_before_now_block,
|
||||
message_id_list=message_id_list,
|
||||
)
|
||||
|
||||
# --- 调用 LLM (普通文本生成) ---
|
||||
|
|
@ -185,60 +587,54 @@ class ActionPlanner:
|
|||
try:
|
||||
parsed_json = json.loads(repair_json(llm_content))
|
||||
|
||||
# 处理不同的JSON格式，复用_parse_single_action函数
|
||||
if isinstance(parsed_json, list):
|
||||
if parsed_json:
|
||||
# 使用最后一个action（保持原有逻辑）
|
||||
parsed_json = parsed_json[-1]
|
||||
logger.warning(f"{self.log_prefix}LLM返回了多个JSON对象，使用最后一个: {parsed_json}")
|
||||
else:
|
||||
parsed_json = {}
|
||||
|
||||
if not isinstance(parsed_json, dict):
|
||||
logger.error(f"{self.log_prefix}解析后的JSON不是字典类型: {type(parsed_json)}")
|
||||
parsed_json = {}
|
||||
|
||||
action = parsed_json.get("action", "no_action")
|
||||
reasoning = parsed_json.get("reason", "未提供原因")
|
||||
|
||||
# 将所有其他属性添加到action_data
|
||||
for key, value in parsed_json.items():
|
||||
if key not in ["action", "reasoning"]:
|
||||
action_data[key] = value
|
||||
|
||||
# 非no_action动作需要target_message_id
|
||||
if action != "no_action":
|
||||
if target_message_id := parsed_json.get("target_message_id"):
|
||||
# 根据target_message_id查找原始消息
|
||||
target_message = self.find_message_by_id(target_message_id, message_id_list)
|
||||
# 如果获取的target_message为None，输出warning并重新plan
|
||||
if target_message is None:
|
||||
self.plan_retry_count += 1
|
||||
logger.warning(
|
||||
f"{self.log_prefix}无法找到target_message_id '{target_message_id}' 对应的消息，重试次数: {self.plan_retry_count}/{self.max_plan_retries}"
|
||||
)
|
||||
# 仍有重试次数
|
||||
if self.plan_retry_count < self.max_plan_retries:
|
||||
# 递归重新plan
|
||||
return await self.plan(mode, loop_start_time, available_actions)
|
||||
logger.error(
|
||||
f"{self.log_prefix}连续{self.max_plan_retries}次plan获取target_message失败，选择最新消息作为target_message"
|
||||
)
|
||||
target_message = self.get_latest_message(message_id_list)
|
||||
self.plan_retry_count = 0 # 重置计数器
|
||||
else:
|
||||
logger.warning(f"{self.log_prefix}动作'{action}'缺少target_message_id")
|
||||
|
||||
if action != "no_action" and action != "reply" and action not in current_available_actions:
|
||||
logger.warning(
|
||||
f"{self.log_prefix}LLM 返回了当前不可用或无效的动作: '{action}' (可用: {list(current_available_actions.keys())})，将强制使用 'no_action'"
|
||||
if isinstance(parsed_json, dict):
|
||||
# 使用_parse_single_action函数解析单个action
|
||||
# 将字典转换为列表格式
|
||||
current_available_actions_list = list(current_available_actions.items())
|
||||
action_planner_infos = self._parse_single_action(
|
||||
parsed_json, message_id_list, current_available_actions_list
|
||||
)
|
||||
reasoning = f"LLM 返回了当前不可用的动作 '{action}' (可用: {list(current_available_actions.keys())})。原始理由: {reasoning}"
|
||||
|
||||
if action_planner_infos:
|
||||
# 获取第一个（也是唯一一个）action的信息
|
||||
action_info = action_planner_infos[0]
|
||||
action = action_info.action_type
|
||||
reasoning = action_info.reasoning
|
||||
action_data.update(action_info.action_data)
|
||||
target_message = action_info.action_message
|
||||
|
||||
# 处理target_message为None的情况（保持原有的重试逻辑）
|
||||
if target_message is None and action != "no_action":
|
||||
# 尝试获取最新消息作为target_message
|
||||
target_message = self.get_latest_message(message_id_list)
|
||||
if target_message is None:
|
||||
logger.warning(f"{self.log_prefix}无法获取任何消息作为target_message")
|
||||
else:
|
||||
# 如果没有解析到action，使用默认值
|
||||
action = "no_action"
|
||||
reasoning = "解析action失败"
|
||||
target_message = None
|
||||
else:
|
||||
logger.error(f"{self.log_prefix}解析后的JSON不是字典类型: {type(parsed_json)}")
|
||||
action = "no_action"
|
||||
reasoning = f"解析后的JSON类型错误: {type(parsed_json)}"
|
||||
target_message = None
|
||||
|
||||
except Exception as json_e:
|
||||
logger.warning(f"{self.log_prefix}解析LLM响应JSON失败 {json_e}. LLM原始输出: '{llm_content}'")
|
||||
traceback.print_exc()
|
||||
reasoning = f"解析LLM响应JSON失败: {json_e}. 将使用默认动作 'no_action'."
|
||||
action = "no_action"
|
||||
reasoning = f"解析LLM响应JSON失败: {json_e}. 将使用默认动作 'no_action'."
|
||||
target_message = None
|
||||
|
||||
except Exception as outer_e:
|
||||
logger.error(f"{self.log_prefix}Planner 处理过程中发生意外错误，规划失败，将执行 no_action: {outer_e}")
|
||||
|
|
@ -246,30 +642,70 @@ class ActionPlanner:
|
|||
action = "no_action"
|
||||
reasoning = f"Planner 内部处理错误: {outer_e}"
|
||||
|
||||
is_parallel = False
|
||||
is_parallel = True
|
||||
if mode == ChatMode.NORMAL and action in current_available_actions:
|
||||
is_parallel = current_available_actions[action].parallel_action
|
||||
if is_parallel:
|
||||
is_parallel = current_available_actions[action].parallel_action
|
||||
|
||||
action_data["loop_start_time"] = loop_start_time
|
||||
|
||||
actions = [
|
||||
ActionPlannerInfo(
|
||||
action_type=action,
|
||||
reasoning=reasoning,
|
||||
action_data=action_data,
|
||||
action_message=target_message,
|
||||
available_actions=available_actions,
|
||||
)
|
||||
]
|
||||
|
||||
if action != "reply" and is_parallel:
|
||||
actions.append(
|
||||
ActionPlannerInfo(
|
||||
action_type="reply",
|
||||
# 过滤掉no_action，除非所有结果都是no_action
|
||||
def filter_no_actions(action_list):
|
||||
"""过滤no_action，如果所有都是no_action则返回一个"""
|
||||
non_no_actions = [a for a in action_list if a.action_type != "no_action"]
|
||||
if non_no_actions:
|
||||
return non_no_actions
|
||||
else:
|
||||
# 如果所有都是no_action，返回第一个
|
||||
return [action_list[0]] if action_list else []
|
||||
|
||||
# 根据is_parallel决定返回值
|
||||
if is_parallel:
|
||||
# 如果为真，将主规划器的结果和副规划器的结果都返回
|
||||
main_actions = []
|
||||
|
||||
# 添加主规划器的action（如果不是no_action）
|
||||
if action != "no_action":
|
||||
main_actions.append(ActionPlannerInfo(
|
||||
action_type=action,
|
||||
reasoning=reasoning,
|
||||
action_data=action_data,
|
||||
action_message=target_message,
|
||||
available_actions=available_actions,
|
||||
)
|
||||
)
|
||||
))
|
||||
|
||||
# 先合并主副规划器的结果
|
||||
all_actions = main_actions + all_sub_planner_results
|
||||
|
||||
# 然后统一过滤no_action
|
||||
actions = filter_no_actions(all_actions)
|
||||
|
||||
# 如果所有结果都是no_action，返回一个no_action
|
||||
if not actions:
|
||||
actions = [ActionPlannerInfo(
|
||||
action_type="no_action",
|
||||
reasoning="所有规划器都选择不执行动作",
|
||||
action_data={},
|
||||
action_message=None,
|
||||
available_actions=available_actions,
|
||||
)]
|
||||
|
||||
logger.info(f"{self.log_prefix}并行模式：返回主规划器{len(main_actions)}个action + 副规划器{len(all_sub_planner_results)}个action，过滤后总计{len(actions)}个action")
|
||||
else:
|
||||
# 如果为假，只返回副规划器的结果
|
||||
actions = filter_no_actions(all_sub_planner_results)
|
||||
|
||||
# 如果所有结果都是no_action，返回一个no_action
|
||||
if not actions:
|
||||
actions = [ActionPlannerInfo(
|
||||
action_type="no_action",
|
||||
reasoning="副规划器都选择不执行动作",
|
||||
action_data={},
|
||||
action_message=None,
|
||||
available_actions=available_actions,
|
||||
)]
|
||||
|
||||
logger.info(f"{self.log_prefix}非并行模式：返回副规划器的{len(actions)}个action（已过滤no_action）")
|
||||
|
||||
return actions, target_message
|
||||
|
||||
|
|
@ -278,43 +714,19 @@ class ActionPlanner:
|
|||
is_group_chat: bool, # Now passed as argument
|
||||
chat_target_info: Optional[dict], # Now passed as argument
|
||||
current_available_actions: Dict[str, ActionInfo],
|
||||
refresh_time: bool = False,
|
||||
mode: ChatMode = ChatMode.FOCUS,
|
||||
actions_before_now_block :str = "",
|
||||
chat_content_block :str = "",
|
||||
message_id_list :List[Tuple[str, DatabaseMessages]] = None,
|
||||
) -> tuple[str, List[DatabaseMessages]]: # sourcery skip: use-join
|
||||
"""构建 Planner LLM 的提示词 (获取模板并填充数据)"""
|
||||
try:
|
||||
message_list_before_now = get_raw_msg_before_timestamp_with_chat(
|
||||
chat_id=self.chat_id,
|
||||
timestamp=time.time(),
|
||||
limit=int(global_config.chat.max_context_size * 0.6),
|
||||
)
|
||||
chat_content_block, message_id_list = build_readable_messages_with_id(
|
||||
messages=message_list_before_now,
|
||||
timestamp_mode="normal_no_YMD",
|
||||
read_mark=self.last_obs_time_mark,
|
||||
truncate=True,
|
||||
show_actions=True,
|
||||
)
|
||||
|
||||
actions_before_now = get_actions_by_timestamp_with_chat(
|
||||
chat_id=self.chat_id,
|
||||
timestamp_start=time.time() - 600,
|
||||
timestamp_end=time.time(),
|
||||
limit=5,
|
||||
)
|
||||
|
||||
actions_before_now_block = build_readable_actions(
|
||||
actions=actions_before_now,
|
||||
)
|
||||
|
||||
if actions_before_now:
|
||||
|
||||
if actions_before_now_block:
|
||||
actions_before_now_block = f"你刚刚选择并执行过的action是：\n{actions_before_now_block}"
|
||||
else:
|
||||
actions_before_now_block = ""
|
||||
|
||||
if refresh_time:
|
||||
self.last_obs_time_mark = time.time()
|
||||
|
||||
mentioned_bonus = ""
|
||||
if global_config.chat.mentioned_bot_inevitable_reply:
|
||||
mentioned_bonus = "\n- 有人提到你"
|
||||
|
|
@ -348,29 +760,32 @@ class ActionPlanner:
|
|||
|
||||
action_options_block = ""
|
||||
|
||||
for using_actions_name, using_actions_info in current_available_actions.items():
|
||||
if using_actions_info.action_parameters:
|
||||
param_text = "\n"
|
||||
for param_name, param_description in using_actions_info.action_parameters.items():
|
||||
param_text += f' "{param_name}":"{param_description}"\n'
|
||||
param_text = param_text.rstrip("\n")
|
||||
else:
|
||||
param_text = ""
|
||||
if current_available_actions:
|
||||
for using_actions_name, using_actions_info in current_available_actions.items():
|
||||
if using_actions_info.action_parameters:
|
||||
param_text = "\n"
|
||||
for param_name, param_description in using_actions_info.action_parameters.items():
|
||||
param_text += f' "{param_name}":"{param_description}"\n'
|
||||
param_text = param_text.rstrip("\n")
|
||||
else:
|
||||
param_text = ""
|
||||
|
||||
require_text = ""
|
||||
for require_item in using_actions_info.action_require:
|
||||
require_text += f"- {require_item}\n"
|
||||
require_text = require_text.rstrip("\n")
|
||||
require_text = ""
|
||||
for require_item in using_actions_info.action_require:
|
||||
require_text += f"- {require_item}\n"
|
||||
require_text = require_text.rstrip("\n")
|
||||
|
||||
using_action_prompt = await global_prompt_manager.get_prompt_async("action_prompt")
|
||||
using_action_prompt = using_action_prompt.format(
|
||||
action_name=using_actions_name,
|
||||
action_description=using_actions_info.description,
|
||||
action_parameters=param_text,
|
||||
action_require=require_text,
|
||||
)
|
||||
using_action_prompt = await global_prompt_manager.get_prompt_async("action_prompt")
|
||||
using_action_prompt = using_action_prompt.format(
|
||||
action_name=using_actions_name,
|
||||
action_description=using_actions_info.description,
|
||||
action_parameters=param_text,
|
||||
action_require=require_text,
|
||||
)
|
||||
|
||||
action_options_block += using_action_prompt
|
||||
action_options_block += using_action_prompt
|
||||
else:
|
||||
action_options_block = ""
|
||||
|
||||
moderation_prompt_block = "请不要输出违法违规内容，不要输出色情，暴力，政治相关内容，如有敏感内容，请规避。"
|
||||
|
||||
|
|
|
|||
|
|
@ -361,10 +361,10 @@ def _build_readable_messages_internal(
|
|||
# 创建时间戳到消息ID的映射，用于在消息前添加[id]标识符
|
||||
timestamp_to_id_mapping: Dict[float, str] = {}
|
||||
if message_id_list:
|
||||
for msg in message_id_list:
|
||||
for msg_id, msg in message_id_list:
|
||||
timestamp = msg.time
|
||||
if timestamp is not None:
|
||||
timestamp_to_id_mapping[timestamp] = msg.message_id
|
||||
timestamp_to_id_mapping[timestamp] = msg_id
|
||||
|
||||
def process_pic_ids(content: Optional[str]) -> str:
|
||||
"""处理内容中的图片ID，将其替换为[图片x]格式"""
|
||||
|
|
@ -477,7 +477,7 @@ def _build_readable_messages_internal(
|
|||
readable_time = translate_timestamp_to_human_readable(timestamp, mode=timestamp_mode)
|
||||
|
||||
# 查找消息id（如果有）并构建id_prefix
|
||||
message_id = timestamp_to_id_mapping.get(timestamp)
|
||||
message_id = timestamp_to_id_mapping.get(timestamp, "")
|
||||
id_prefix = f"[{message_id}]" if message_id else ""
|
||||
|
||||
if is_action:
|
||||
|
|
@ -606,7 +606,7 @@ def build_readable_messages_with_id(
|
|||
truncate: bool = False,
|
||||
show_actions: bool = False,
|
||||
show_pic: bool = True,
|
||||
) -> Tuple[str, List[DatabaseMessages]]:
|
||||
) -> Tuple[str, List[Tuple[str, DatabaseMessages]]]:
|
||||
"""
|
||||
将消息列表转换为可读的文本格式，并返回原始(时间戳, 昵称, 内容)列表。
|
||||
允许通过参数控制格式化行为。
|
||||
|
|
|
|||
|
|
@ -685,7 +685,7 @@ def assign_message_ids(messages: List[DatabaseMessages]) -> List[DatabaseMessage
|
|||
Returns:
|
||||
List[DatabaseMessages]: 分配了唯一ID的消息列表(写入message_id属性)
|
||||
"""
|
||||
result: List[DatabaseMessages] = list(messages) # 复制原始消息列表
|
||||
result: List[Tuple[str, DatabaseMessages]] = [] # 复制原始消息列表
|
||||
used_ids = set()
|
||||
len_i = len(messages)
|
||||
if len_i > 100:
|
||||
|
|
@ -695,7 +695,7 @@ def assign_message_ids(messages: List[DatabaseMessages]) -> List[DatabaseMessage
|
|||
a = 1
|
||||
b = 9
|
||||
|
||||
for i, _ in enumerate(result):
|
||||
for i, message in enumerate(messages):
|
||||
# 生成唯一的简短ID
|
||||
while True:
|
||||
# 使用索引+随机数生成简短ID
|
||||
|
|
@ -705,7 +705,7 @@ def assign_message_ids(messages: List[DatabaseMessages]) -> List[DatabaseMessage
|
|||
if message_id not in used_ids:
|
||||
used_ids.add(message_id)
|
||||
break
|
||||
result[i].message_id = message_id
|
||||
result.append((message_id, message))
|
||||
|
||||
return result
|
||||
|
||||
|
|
|
|||
|
|
@ -117,6 +117,9 @@ class ModelTaskConfig(ConfigBase):
|
|||
planner: TaskConfig
|
||||
"""规划模型配置"""
|
||||
|
||||
planner_small: TaskConfig
|
||||
"""副规划模型配置"""
|
||||
|
||||
embedding: TaskConfig
|
||||
"""嵌入模型配置"""
|
||||
|
||||
|
|
|
|||
|
|
@ -76,6 +76,9 @@ class ChatConfig(ConfigBase):
|
|||
|
||||
mentioned_bot_inevitable_reply: bool = False
|
||||
"""提及 bot 必然回复"""
|
||||
|
||||
planner_size: int = 1
|
||||
"""副规划器大小，越小，麦麦的动作执行能力越精细，但是消耗更多token，调大可以缓解429类错误"""
|
||||
|
||||
at_bot_inevitable_reply: bool = False
|
||||
"""@bot 必然回复"""
|
||||
|
|
|
|||
|
|
@ -115,9 +115,9 @@ class ActionInfo(ComponentInfo):
|
|||
action_require: List[str] = field(default_factory=list) # 动作需求说明
|
||||
associated_types: List[str] = field(default_factory=list) # 关联的消息类型
|
||||
# 激活类型相关
|
||||
focus_activation_type: ActionActivationType = ActionActivationType.ALWAYS
|
||||
normal_activation_type: ActionActivationType = ActionActivationType.ALWAYS
|
||||
activation_type: ActionActivationType = ActionActivationType.ALWAYS
|
||||
focus_activation_type: ActionActivationType = ActionActivationType.ALWAYS #已弃用
|
||||
normal_activation_type: ActionActivationType = ActionActivationType.ALWAYS #已弃用
|
||||
activation_type: ActionActivationType = ActionActivationType.ALWAYS
|
||||
random_activation_probability: float = 0.0
|
||||
llm_judge_prompt: str = ""
|
||||
activation_keywords: List[str] = field(default_factory=list) # 激活关键词列表
|
||||
|
|
|
|||
|
|
@ -1,5 +1,5 @@
|
|||
[inner]
|
||||
version = "6.6.1"
|
||||
version = "6.7.0"
|
||||
|
||||
#----以下是给开发人员阅读的，如果你只是部署了麦麦，不需要阅读----
|
||||
#如果你想要修改配置文件，请递增version的值
|
||||
|
|
@ -67,6 +67,8 @@ max_context_size = 20 # 上下文长度
|
|||
|
||||
interest_rate_mode = "fast" #激活值计算模式，可选fast或者accurate
|
||||
|
||||
planner_size = 1 # 副规划器大小，越小，麦麦的动作执行能力越精细，但是消耗更多token，调大可以缓解429类错误
|
||||
|
||||
mentioned_bot_inevitable_reply = true # 提及 bot 大概率回复
|
||||
at_bot_inevitable_reply = true # @bot 或 提及bot 大概率回复
|
||||
|
||||
|
|
|
|||
|
|
@ -1,5 +1,5 @@
|
|||
[inner]
|
||||
version = "1.3.1"
|
||||
version = "1.4.0"
|
||||
|
||||
# 配置文件版本号迭代规则同bot_config.toml
|
||||
|
||||
|
|
@ -117,11 +117,16 @@ model_list = ["siliconflow-deepseek-v3"]
|
|||
temperature = 0.2 # 模型温度，新V3建议0.1-0.3
|
||||
max_tokens = 800
|
||||
|
||||
[model_task_config.planner] #决策：负责决定麦麦该做什么的模型
|
||||
[model_task_config.planner] #决策：负责决定麦麦该什么时候回复的模型
|
||||
model_list = ["siliconflow-deepseek-v3"]
|
||||
temperature = 0.3
|
||||
max_tokens = 800
|
||||
|
||||
[model_task_config.planner_small] #副决策：负责决定麦麦该做什么的模型
|
||||
model_list = ["qwen3-14b"]
|
||||
temperature = 0.3
|
||||
max_tokens = 800
|
||||
|
||||
[model_task_config.emotion] #负责麦麦的情绪变化
|
||||
model_list = ["siliconflow-deepseek-v3"]
|
||||
temperature = 0.3
|
||||
|
|
|
|||
Loading…
Reference in New Issue