From 1eeabe76ba0fd190820179fca4d53182761b40c7 Mon Sep 17 00:00:00 2001 From: UnCLAS-Prommer Date: Fri, 22 Aug 2025 23:40:24 +0800 Subject: [PATCH 1/2] =?UTF-8?q?=E6=9B=B4=E6=94=B9generator=E7=9A=84?= =?UTF-8?q?=E8=BF=94=E5=9B=9E=E5=80=BC=E4=B8=BA=E4=B8=80=E4=B8=AA=E6=95=B0?= =?UTF-8?q?=E6=8D=AE=E6=A8=A1=E5=9E=8B=E7=A8=B3=E5=AE=9Aapi?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- src/chat/heart_flow/heartFC_chat.py | 8 +-- src/chat/planner_actions/action_modifier.py | 8 +-- src/chat/replyer/default_generator.py | 63 +++++++++++---------- src/common/data_models/llm_data_model.py | 16 ++++++ src/plugin_system/apis/generator_api.py | 53 ++++++----------- src/plugin_system/core/events_manager.py | 31 +++++----- 6 files changed, 90 insertions(+), 89 deletions(-) create mode 100644 src/common/data_models/llm_data_model.py diff --git a/src/chat/heart_flow/heartFC_chat.py b/src/chat/heart_flow/heartFC_chat.py index f4200028..eed12623 100644 --- a/src/chat/heart_flow/heartFC_chat.py +++ b/src/chat/heart_flow/heartFC_chat.py @@ -679,7 +679,7 @@ class HeartFChatting: } else: try: - success, response_set, prompt, selected_expressions = await generator_api.generate_reply( + success, llm_response = await generator_api.generate_reply( chat_stream=self.chat_stream, reply_message=action_planner_info.action_message, available_actions=available_actions, @@ -688,10 +688,9 @@ class HeartFChatting: enable_tool=global_config.tool.enable_tool, request_type="replyer", from_plugin=False, - return_expressions=True, ) - if not success or not response_set: + if not success or not llm_response or not llm_response.reply_set: if action_planner_info.action_message: logger.info(f"对 {action_planner_info.action_message.processed_plain_text} 的回复生成失败") else: @@ -701,7 +700,8 @@ class HeartFChatting: except asyncio.CancelledError: logger.debug(f"{self.log_prefix} 并行执行：回复生成任务已被取消") return {"action_type": "reply", "success": False, "reply_text": "", "loop_info": None} - + response_set = llm_response.reply_set + selected_expressions = llm_response.selected_expressions loop_info, reply_text, _ = await self._send_and_store_reply( response_set=response_set, action_message=action_planner_info.action_message, # type: ignore diff --git a/src/chat/planner_actions/action_modifier.py b/src/chat/planner_actions/action_modifier.py index 03c72ffc..d7f46e87 100644 --- a/src/chat/planner_actions/action_modifier.py +++ b/src/chat/planner_actions/action_modifier.py @@ -2,7 +2,7 @@ import random import asyncio import hashlib import time -from typing import List, Any, Dict, TYPE_CHECKING, Tuple +from typing import List, Dict, TYPE_CHECKING, Tuple from src.common.logger import get_logger from src.config.config import global_config, model_config @@ -161,7 +161,7 @@ class ActionModifier: deactivated_actions = [] # 分类处理不同激活类型的actions - llm_judge_actions = {} + llm_judge_actions: Dict[str, ActionInfo] = {} actions_to_check = list(actions_with_info.items()) random.shuffle(actions_to_check) @@ -218,7 +218,7 @@ class ActionModifier: async def _process_llm_judge_actions_parallel( self, - llm_judge_actions: Dict[str, Any], + llm_judge_actions: Dict[str, ActionInfo], chat_content: str = "", ) -> Dict[str, bool]: """ @@ -237,7 +237,7 @@ class ActionModifier: current_time = time.time() results = {} - tasks_to_run = {} + tasks_to_run: Dict[str, ActionInfo] = {} # 检查缓存 for action_name, action_info in llm_judge_actions.items(): diff --git a/src/chat/replyer/default_generator.py b/src/chat/replyer/default_generator.py index 59340914..b858af92 100644 --- a/src/chat/replyer/default_generator.py +++ b/src/chat/replyer/default_generator.py @@ -10,6 +10,7 @@ from src.mais4u.mai_think import mai_thinking_manager from src.common.logger import get_logger from src.common.data_models.database_data_model import DatabaseMessages from src.common.data_models.info_data_model import ActionPlannerInfo +from src.common.data_models.llm_data_model import LLMGenerationDataModel from src.config.config import global_config, model_config from src.llm_models.utils_model import LLMRequest from src.chat.message_receive.message import UserInfo, Seg, MessageRecv, MessageSending @@ -162,7 +163,7 @@ class DefaultReplyer: from_plugin: bool = True, stream_id: Optional[str] = None, reply_message: Optional[DatabaseMessages] = None, - ) -> Tuple[bool, Optional[Dict[str, Any]], Optional[str], Optional[List[int]]]: + ) -> Tuple[bool, LLMGenerationDataModel]: # sourcery skip: merge-nested-ifs """ 回复器 (Replier): 负责生成回复文本的核心逻辑。 @@ -182,6 +183,7 @@ class DefaultReplyer: prompt = None selected_expressions: Optional[List[int]] = None + llm_response = LLMGenerationDataModel() if available_actions is None: available_actions = {} try: @@ -195,10 +197,12 @@ class DefaultReplyer: reply_message=reply_message, reply_reason=reply_reason, ) + llm_response.prompt = prompt + llm_response.selected_expressions = selected_expressions if not prompt: logger.warning("构建prompt失败，跳过回复生成") - return False, None, None, [] + return False, llm_response from src.plugin_system.core.events_manager import events_manager if not from_plugin: @@ -215,12 +219,10 @@ class DefaultReplyer: try: content, reasoning_content, model_name, tool_call = await self.llm_generate_content(prompt) logger.debug(f"replyer生成内容: {content}") - llm_response = { - "content": content, - "reasoning": reasoning_content, - "model": model_name, - "tool_calls": tool_call, - } + llm_response.content = content + llm_response.reasoning = reasoning_content + llm_response.model = model_name + llm_response.tool_calls = tool_call if not from_plugin and not await events_manager.handle_mai_events( EventType.AFTER_LLM, None, prompt, llm_response, stream_id=stream_id ): @@ -230,24 +232,23 @@ class DefaultReplyer: except Exception as llm_e: # 精简报错信息 logger.error(f"LLM 生成失败: {llm_e}") - return False, None, prompt, selected_expressions # LLM 调用失败则无法生成回复 + return False, llm_response # LLM 调用失败则无法生成回复 - return True, llm_response, prompt, selected_expressions + return True, llm_response except UserWarning as uw: raise uw except Exception as e: logger.error(f"回复生成意外失败: {e}") traceback.print_exc() - return False, None, prompt, selected_expressions + return False, llm_response async def rewrite_reply_with_context( self, raw_reply: str = "", reason: str = "", reply_to: str = "", - return_prompt: bool = False, - ) -> Tuple[bool, Optional[str], Optional[str]]: + ) -> Tuple[bool, LLMGenerationDataModel]: """ 表达器 (Expressor): 负责重写和优化回复文本。 @@ -260,6 +261,7 @@ class DefaultReplyer: Returns: Tuple[bool, Optional[str]]: (是否成功, 重写后的回复内容) """ + llm_response = LLMGenerationDataModel() try: with Timer("构建Prompt", {}): # 内部计时器，可选保留 prompt = await self.build_prompt_rewrite_context( @@ -267,29 +269,33 @@ class DefaultReplyer: reason=reason, reply_to=reply_to, ) + llm_response.prompt = prompt content = None reasoning_content = None model_name = "unknown_model" if not prompt: logger.error("Prompt 构建失败，无法生成回复。") - return False, None, None + return False, llm_response try: content, reasoning_content, model_name, _ = await self.llm_generate_content(prompt) logger.info(f"想要表达：{raw_reply}||理由：{reason}||生成回复: {content}\n") + llm_response.content = content + llm_response.reasoning = reasoning_content + llm_response.model = model_name except Exception as llm_e: # 精简报错信息 logger.error(f"LLM 生成失败: {llm_e}") - return False, None, prompt if return_prompt else None # LLM 调用失败则无法生成回复 + return False, llm_response # LLM 调用失败则无法生成回复 - return True, content, prompt if return_prompt else None + return True, llm_response except Exception as e: logger.error(f"回复生成意外失败: {e}") traceback.print_exc() - return False, None, prompt if return_prompt else None + return False, llm_response async def build_relation_info(self, sender: str, target: str): if not global_config.relationship.enable_relationship: @@ -375,9 +381,7 @@ class DefaultReplyer: if global_config.memory.enable_instant_memory: chat_history_str = build_readable_messages( - messages=chat_history, - replace_bot_name=True, - timestamp_mode="normal" + messages=chat_history, replace_bot_name=True, timestamp_mode="normal" ) asyncio.create_task(self.instant_memory.create_and_store_memory(chat_history_str)) @@ -668,16 +672,18 @@ class DefaultReplyer: action_descriptions += chosen_action_descriptions return action_descriptions - + async def build_personality_prompt(self) -> str: 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 = "" - - prompt_personality = f"{global_config.personality.personality_core};{global_config.personality.personality_side}" - return f"你的名字是{bot_name}{bot_nickname}，你{prompt_personality}" + + prompt_personality = ( + f"{global_config.personality.personality_core};{global_config.personality.personality_side}" + ) + return f"你的名字是{bot_name}{bot_nickname}，你{prompt_personality}" async def build_prompt_reply_context( self, @@ -875,17 +881,12 @@ class DefaultReplyer: raw_reply: str, reason: str, reply_to: str, - reply_message: Optional[Dict[str, Any]] = None, ) -> str: # sourcery skip: merge-else-if-into-elif, remove-redundant-if chat_stream = self.chat_stream chat_id = chat_stream.stream_id is_group_chat = bool(chat_stream.group_info) - if reply_message: - sender = reply_message.get("sender", "") - target = reply_message.get("target", "") - else: - sender, target = self._parse_reply_target(reply_to) + sender, target = self._parse_reply_target(reply_to) # 添加情绪状态获取 if global_config.mood.enable_mood: @@ -908,7 +909,7 @@ class DefaultReplyer: ) # 并行执行2个构建任务 - (expression_habits_block, _), relation_info, personality_prompt = await asyncio.gather( + (expression_habits_block, _), relation_info, personality_prompt = await asyncio.gather( self.build_expression_habits(chat_talking_prompt_half, target), self.build_relation_info(sender, target), self.build_personality_prompt(), diff --git a/src/common/data_models/llm_data_model.py b/src/common/data_models/llm_data_model.py new file mode 100644 index 00000000..1d5b75e0 --- /dev/null +++ b/src/common/data_models/llm_data_model.py @@ -0,0 +1,16 @@ +from dataclasses import dataclass +from typing import Optional, List, Tuple, TYPE_CHECKING, Any + +from . import BaseDataModel +if TYPE_CHECKING: + from src.llm_models.payload_content.tool_option import ToolCall + +@dataclass +class LLMGenerationDataModel(BaseDataModel): + content: Optional[str] = None + reasoning: Optional[str] = None + model: Optional[str] = None + tool_calls: Optional[List["ToolCall"]] = None + prompt: Optional[str] = None + selected_expressions: Optional[List[int]] = None + reply_set: Optional[List[Tuple[str, Any]]] = None \ No newline at end of file diff --git a/src/plugin_system/apis/generator_api.py b/src/plugin_system/apis/generator_api.py index 49e78e95..257c60fa 100644 --- a/src/plugin_system/apis/generator_api.py +++ b/src/plugin_system/apis/generator_api.py @@ -21,6 +21,7 @@ from src.plugin_system.base.component_types import ActionInfo if TYPE_CHECKING: from src.common.data_models.info_data_model import ActionPlannerInfo from src.common.data_models.database_data_model import DatabaseMessages + from src.common.data_models.llm_data_model import LLMGenerationDataModel install(extra_lines=3) @@ -85,11 +86,9 @@ async def generate_reply( enable_tool: bool = False, enable_splitter: bool = True, enable_chinese_typo: bool = True, - return_prompt: bool = False, request_type: str = "generator_api", from_plugin: bool = True, - return_expressions: bool = False, -) -> Tuple[bool, List[Tuple[str, Any]], Optional[str], Optional[List[int]]]: +) -> Tuple[bool, Optional["LLMGenerationDataModel"]]: """生成回复 Args: @@ -117,7 +116,7 @@ async def generate_reply( replyer = get_replyer(chat_stream, chat_id, request_type=request_type) if not replyer: logger.error("[GeneratorAPI] 无法获取回复器") - return False, [], None, None + return False, None if not extra_info and action_data: extra_info = action_data.get("extra_info", "") @@ -126,7 +125,7 @@ async def generate_reply( reply_reason = action_data.get("reason", "") # 调用回复器生成回复 - success, llm_response_dict, prompt, selected_expressions = await replyer.generate_reply_with_context( + success, llm_response = await replyer.generate_reply_with_context( extra_info=extra_info, available_actions=available_actions, chosen_actions=chosen_actions, @@ -138,43 +137,27 @@ async def generate_reply( ) if not success: logger.warning("[GeneratorAPI] 回复生成失败") - return False, [], None, None - assert llm_response_dict is not None, "llm_response_dict不应为None" # 虽然说不会出现llm_response为空的情况 - if content := llm_response_dict.get("content", ""): + return False, None + if content := llm_response.content: reply_set = process_human_text(content, enable_splitter, enable_chinese_typo) else: reply_set = [] + llm_response.reply_set = reply_set logger.debug(f"[GeneratorAPI] 回复生成成功，生成了 {len(reply_set)} 个回复项") - # if return_prompt: - # if return_expressions: - # return success, reply_set, prompt, selected_expressions - # else: - # return success, reply_set, prompt, None - # else: - # if return_expressions: - # return success, reply_set, (None, selected_expressions) - # else: - # return success, reply_set, None - return ( - success, - reply_set, - prompt if return_prompt else None, - selected_expressions if return_expressions else None, - ) + return success, llm_response except ValueError as ve: raise ve except UserWarning as uw: logger.warning(f"[GeneratorAPI] 中断了生成: {uw}") - return False, [], None, None + return False, None except Exception as e: logger.error(f"[GeneratorAPI] 生成回复时出错: {e}") logger.error(traceback.format_exc()) - return False, [], None, None - + return False, None async def rewrite_reply( chat_stream: Optional[ChatStream] = None, @@ -185,9 +168,8 @@ async def rewrite_reply( raw_reply: str = "", reason: str = "", reply_to: str = "", - return_prompt: bool = False, request_type: str = "generator_api", -) -> Tuple[bool, List[Tuple[str, Any]], Optional[str]]: +) -> Tuple[bool, Optional["LLMGenerationDataModel"]]: """重写回复 Args: @@ -210,7 +192,7 @@ async def rewrite_reply( replyer = get_replyer(chat_stream, chat_id, request_type=request_type) if not replyer: logger.error("[GeneratorAPI] 无法获取回复器") - return False, [], None + return False, None logger.info("[GeneratorAPI] 开始重写回复") @@ -221,29 +203,28 @@ async def rewrite_reply( reply_to = reply_to or reply_data.get("reply_to", "") # 调用回复器重写回复 - success, content, prompt = await replyer.rewrite_reply_with_context( + success, llm_response = await replyer.rewrite_reply_with_context( raw_reply=raw_reply, reason=reason, reply_to=reply_to, - return_prompt=return_prompt, ) reply_set = [] - if content: + if success and llm_response and (content := llm_response.content): reply_set = process_human_text(content, enable_splitter, enable_chinese_typo) - + llm_response.reply_set = reply_set if success: logger.info(f"[GeneratorAPI] 重写回复成功，生成了 {len(reply_set)} 个回复项") else: logger.warning("[GeneratorAPI] 重写回复失败") - return success, reply_set, prompt if return_prompt else None + return success, llm_response except ValueError as ve: raise ve except Exception as e: logger.error(f"[GeneratorAPI] 重写回复时出错: {e}") - return False, [], None + return False, None def process_human_text(content: str, enable_splitter: bool, enable_chinese_typo: bool) -> List[Tuple[str, Any]]: diff --git a/src/plugin_system/core/events_manager.py b/src/plugin_system/core/events_manager.py index 1c99cb7a..2472d831 100644 --- a/src/plugin_system/core/events_manager.py +++ b/src/plugin_system/core/events_manager.py @@ -1,6 +1,6 @@ import asyncio import contextlib -from typing import List, Dict, Optional, Type, Tuple, Any, Coroutine +from typing import List, Dict, Optional, Type, Tuple, Any, TYPE_CHECKING from src.chat.message_receive.message import MessageRecv from src.chat.message_receive.chat_stream import get_chat_manager @@ -9,6 +9,9 @@ from src.plugin_system.base.component_types import EventType, EventHandlerInfo, from src.plugin_system.base.base_events_handler import BaseEventHandler from .global_announcement_manager import global_announcement_manager +if TYPE_CHECKING: + from src.common.data_models.llm_data_model import LLMGenerationDataModel + logger = get_logger("events_manager") @@ -47,7 +50,7 @@ class EventsManager: event_type: EventType, message: Optional[MessageRecv] = None, llm_prompt: Optional[str] = None, - llm_response: Optional[Dict[str, Any]] = None, + llm_response: Optional["LLMGenerationDataModel"] = None, stream_id: Optional[str] = None, action_usage: Optional[List[str]] = None, ) -> Optional[MaiMessages]: @@ -97,7 +100,7 @@ class EventsManager: event_type: EventType, message: Optional[MessageRecv] = None, llm_prompt: Optional[str] = None, - llm_response: Optional[Dict[str, Any]] = None, + llm_response: Optional["LLMGenerationDataModel"] = None, stream_id: Optional[str] = None, action_usage: Optional[List[str]] = None, ) -> bool: @@ -175,16 +178,16 @@ class EventsManager: return False def _transform_event_message( - self, message: MessageRecv, llm_prompt: Optional[str] = None, llm_response: Optional[Dict[str, Any]] = None + self, message: MessageRecv, llm_prompt: Optional[str] = None, llm_response: Optional["LLMGenerationDataModel"] = None ) -> MaiMessages: """转换事件消息格式""" # 直接赋值部分内容 transformed_message = MaiMessages( llm_prompt=llm_prompt, - llm_response_content=llm_response.get("content") if llm_response else None, - llm_response_reasoning=llm_response.get("reasoning") if llm_response else None, - llm_response_model=llm_response.get("model") if llm_response else None, - llm_response_tool_call=llm_response.get("tool_calls") if llm_response else None, + llm_response_content=llm_response.content if llm_response else None, + llm_response_reasoning=llm_response.reasoning if llm_response else None, + llm_response_model=llm_response.model if llm_response else None, + llm_response_tool_call=llm_response.tool_calls if llm_response else None, raw_message=message.raw_message, additional_data=message.message_info.additional_config or {}, ) @@ -228,7 +231,7 @@ class EventsManager: return transformed_message def _build_message_from_stream( - self, stream_id: str, llm_prompt: Optional[str] = None, llm_response: Optional[Dict[str, Any]] = None + self, stream_id: str, llm_prompt: Optional[str] = None, llm_response: Optional["LLMGenerationDataModel"] = None ) -> MaiMessages: """从流ID构建消息""" chat_stream = get_chat_manager().get_stream(stream_id) @@ -240,7 +243,7 @@ class EventsManager: self, stream_id: str, llm_prompt: Optional[str] = None, - llm_response: Optional[Dict[str, Any]] = None, + llm_response: Optional["LLMGenerationDataModel"] = None, action_usage: Optional[List[str]] = None, ) -> MaiMessages: """没有message对象时进行转换""" @@ -249,10 +252,10 @@ class EventsManager: return MaiMessages( stream_id=stream_id, llm_prompt=llm_prompt, - llm_response_content=(llm_response.get("content") if llm_response else None), - llm_response_reasoning=(llm_response.get("reasoning") if llm_response else None), - llm_response_model=llm_response.get("model") if llm_response else None, - llm_response_tool_call=(llm_response.get("tool_calls") if llm_response else None), + llm_response_content=(llm_response.content if llm_response else None), + llm_response_reasoning=(llm_response.reasoning if llm_response else None), + llm_response_model=(llm_response.model if llm_response else None), + llm_response_tool_call=(llm_response.tool_calls if llm_response else None), is_group_message=(not (not chat_stream.group_info)), is_private_message=(not chat_stream.group_info), action_usage=action_usage, From 2bb4ec4ae24589cada3f256b5087e5cea2fa6b8b Mon Sep 17 00:00:00 2001 From: UnCLAS-Prommer Date: Fri, 22 Aug 2025 23:43:26 +0800 Subject: [PATCH 2/2] =?UTF-8?q?=E5=88=A0=E6=8E=89=E9=BB=98=E8=AE=A4?= =?UTF-8?q?=E7=9A=84Pro=E9=98=B2=E6=AD=A2=E5=86=8D=E6=9D=A5=E8=87=AA?= =?UTF-8?q?=E5=8A=A8=E9=99=8D=E7=BA=A7=E9=97=AE=E9=A2=98?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- template/model_config_template.toml | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/template/model_config_template.toml b/template/model_config_template.toml index e3b041e6..9ef0887c 100644 --- a/template/model_config_template.toml +++ b/template/model_config_template.toml @@ -1,5 +1,5 @@ [inner] -version = "1.4.0" +version = "1.4.1" # 配置文件版本号迭代规则同bot_config.toml @@ -40,14 +40,14 @@ price_out = 8.0 # 输出价格（用于API调用统计，单 #force_stream_mode = true # 强制流式输出模式（若模型不支持非流式输出，请取消该注释，启用强制流式输出，若无该字段，默认值为false） [[models]] -model_identifier = "Pro/deepseek-ai/DeepSeek-V3" +model_identifier = "deepseek-ai/DeepSeek-V3" name = "siliconflow-deepseek-v3" api_provider = "SiliconFlow" price_in = 2.0 price_out = 8.0 [[models]] -model_identifier = "Pro/deepseek-ai/DeepSeek-R1-Distill-Qwen-32B" +model_identifier = "deepseek-ai/DeepSeek-R1-Distill-Qwen-32B" name = "deepseek-r1-distill-qwen-32b" api_provider = "SiliconFlow" price_in = 4.0