diff --git a/docs/latest/advanced/agents-config.md b/docs/latest/advanced/agents-config.md index 37637e6e..31ab9955 100644 --- a/docs/latest/advanced/agents-config.md +++ b/docs/latest/advanced/agents-config.md @@ -77,7 +77,8 @@ class MyAgentContext(BaseContext): ```python from src.agents.common import BaseContext, gen_tool_info -from src.agents.common.tools import get_buildin_tools +from src.agents.common.toolkits.buildin import calculator, query_knowledge_graph +from src.agents.common.toolkits.buildin.tools import _create_tavily_search from src.agents.common.toolkits.mysql import get_mysql_tools @dataclass(kw_only=True) @@ -86,7 +87,9 @@ class ReporterContext(BaseContext): default_factory=lambda: [t.name for t in get_mysql_tools()], metadata={ "name": "工具", - "options": lambda: gen_tool_info(get_buildin_tools() + get_mysql_tools()), + "options": lambda: gen_tool_info( + [calculator, query_knowledge_graph, _create_tavily_search()] + get_mysql_tools() + ), "description": "包含内置工具和 MySQL 工具包。", }, ) diff --git a/server/routers/chat_router.py b/server/routers/chat_router.py index 49d786d9..acc3575d 100644 --- a/server/routers/chat_router.py +++ b/server/routers/chat_router.py @@ -551,8 +551,7 @@ async def save_agent_config( if "knowledges" in config and config["knowledges"]: # 获取用户有权访问的知识库名称 try: - user_info = {"role": current_user.role, "department_id": current_user.department_id} - accessible_databases = await knowledge_base.get_databases_by_user(user_info) + accessible_databases = await knowledge_base.get_databases_by_user_id(current_user.user_id) accessible_kb_names = { db.get("name") for db in accessible_databases.get("databases", []) if db.get("name") } diff --git a/server/routers/knowledge_router.py b/server/routers/knowledge_router.py index 2b06a966..90a01d00 100644 --- a/server/routers/knowledge_router.py +++ b/server/routers/knowledge_router.py @@ -89,8 +89,7 @@ async def _ensure_database_not_dify(db_id: str, operation: str) -> None: async def get_databases(current_user: User = Depends(get_admin_user)): """获取所有知识库（根据用户权限过滤）""" try: - user_info = {"role": current_user.role, "department_id": current_user.department_id} - return await knowledge_base.get_databases_by_user(user_info) + return await knowledge_base.get_databases_by_user_id(current_user.user_id) except Exception as e: logger.error(f"获取数据库列表失败 {e}, {traceback.format_exc()}") return {"message": f"获取数据库列表失败 {e}", "databases": []} @@ -180,8 +179,7 @@ async def create_database( async def get_accessible_databases(current_user: User = Depends(get_required_user)): """获取当前用户有权访问的知识库列表（用于智能体配置）""" try: - user_info = {"role": current_user.role, "department_id": current_user.department_id} - databases = await knowledge_base.get_databases_by_user(user_info) + databases = await knowledge_base.get_databases_by_user_id(current_user.user_id) accessible = [ { @@ -664,10 +662,10 @@ async def batch_delete_documents(db_id: str, file_ids: list[str] = Body(...), cu """批量删除文档或文件夹""" logger.debug(f"BATCH DELETE documents {file_ids} in {db_id}") await _ensure_database_not_dify(db_id, "批量文档删除") - + deleted_count = 0 failed_items = [] - + for doc_id in file_ids: try: file_meta_info = await knowledge_base.get_file_basic_info(db_id, doc_id) @@ -700,7 +698,7 @@ async def batch_delete_documents(db_id: str, file_ids: list[str] = Body(...), cu if deleted_count == 0: raise HTTPException(status_code=400, detail=f"批量删除失败: 所有 {len(failed_items)} 个文件均未删除。") return { - "message": f"部分删除成功: 已删除 {deleted_count} 个文件，失败 {len(failed_items)} 个", + "message": f"部分删除成功: 已删除 {deleted_count} 个文件，失败 {len(failed_items)} 个", "deleted_count": deleted_count, "failed_items": failed_items } diff --git a/server/routers/mindmap_router.py b/server/routers/mindmap_router.py index 012bb8b1..b2f13fb1 100644 --- a/server/routers/mindmap_router.py +++ b/server/routers/mindmap_router.py @@ -286,8 +286,7 @@ async def get_databases_overview(current_user: User = Depends(get_admin_user)): 知识库列表 """ try: - user_info = {"role": current_user.role, "department_id": current_user.department_id} - databases = await knowledge_base.get_databases_by_user(user_info) + databases = await knowledge_base.get_databases_by_user_id(current_user.user_id) # databases["databases"] 是一个列表，每个元素已经包含了基本信息 db_list_raw = databases.get("databases", []) diff --git a/src/agents/chatbot/graph.py b/src/agents/chatbot/graph.py index 984d5a64..c29973fc 100644 --- a/src/agents/chatbot/graph.py +++ b/src/agents/chatbot/graph.py @@ -12,6 +12,7 @@ from src.agents.common.middlewares import ( RuntimeConfigMiddleware, save_attachments_to_fs, ) +from src.agents.common.middlewares.knowledge_base_middleware import KnowledgeBaseMiddleware from src.services.mcp_service import get_tools_from_all_servers @@ -43,7 +44,8 @@ class ChatbotAgent(BaseAgent): middleware=[ save_attachments_to_fs, # 附件注入提示词 FilesystemMiddleware(backend=_create_fs_backend), # 文件系统后端 - RuntimeConfigMiddleware(extra_tools=all_mcp_tools), # 运行时配置应用（模型/工具/知识库/MCP/提示词） + KnowledgeBaseMiddleware(), # 知识库工具 + RuntimeConfigMiddleware(extra_tools=all_mcp_tools), # 运行时配置应用（模型/工具/MCP/提示词） ModelRetryMiddleware(), # 模型重试中间件 TodoListMiddleware(), PatchToolCallsMiddleware(), diff --git a/src/agents/common/__init__.py b/src/agents/common/__init__.py index dd71683e..f5a9abba 100644 --- a/src/agents/common/__init__.py +++ b/src/agents/common/__init__.py @@ -7,7 +7,6 @@ from src.agents.common.models import load_chat_model from src.agents.common.state import BaseState # Tools - 核心工具函数 -from src.agents.common.toolkits.buildin import get_buildin_tools from src.agents.common.toolkits.utils import gen_tool_info # MCP - Agent 层统一入口（自动过滤 disabled_tools） @@ -21,7 +20,6 @@ __all__ = [ # Model utilities "load_chat_model", # Core tools - "get_buildin_tools", "gen_tool_info", # Core MCP "get_enabled_mcp_tools", diff --git a/src/agents/common/context.py b/src/agents/common/context.py index 4a609c3e..13dd6ae7 100644 --- a/src/agents/common/context.py +++ b/src/agents/common/context.py @@ -12,9 +12,6 @@ from src import config as sys_config from src.services.mcp_service import get_mcp_server_names from src.utils import logger -from .toolkits.buildin import get_buildin_tools -from .toolkits.utils import gen_tool_info - @dataclass(kw_only=True) class BaseContext: @@ -66,7 +63,6 @@ class BaseContext: default_factory=list, metadata={ "name": "工具", - "options": lambda: gen_tool_info(get_buildin_tools()), "description": "内置的工具。", }, ) diff --git a/src/agents/common/middlewares/knowledge_base_middleware.py b/src/agents/common/middlewares/knowledge_base_middleware.py new file mode 100644 index 00000000..489807db --- /dev/null +++ b/src/agents/common/middlewares/knowledge_base_middleware.py @@ -0,0 +1,26 @@ +"""知识库中间件 - 提供通用知识库工具""" +from collections.abc import Callable +from typing import Any + +from langchain.agents.middleware import AgentMiddleware, ModelRequest, ModelResponse +from langchain.tools.tool_node import ToolCallRequest + +from src.agents.common.toolkits.kbs import get_common_kb_tools +from src.utils.logging_config import logger + + +class KnowledgeBaseMiddleware(AgentMiddleware): + """知识库中间件 - 提供通用知识库工具 + + 提供 3 个通用工具： + - list_kbs: 列出用户可访问的知识库 + - get_mindmap: 获取指定知识库的思维导图 + - query_kb: 在指定知识库中检索 + """ + + def __init__(self): + super().__init__() + # 预加载通用知识库工具 + self.kb_tools = get_common_kb_tools() + self.tools = self.kb_tools + logger.debug(f"Initialized KnowledgeBaseMiddleware with {len(self.kb_tools)} tools") \ No newline at end of file diff --git a/src/agents/common/middlewares/runtime_config_middleware.py b/src/agents/common/middlewares/runtime_config_middleware.py index ea217edc..40db900b 100644 --- a/src/agents/common/middlewares/runtime_config_middleware.py +++ b/src/agents/common/middlewares/runtime_config_middleware.py @@ -12,8 +12,7 @@ from langchain_core.messages import SystemMessage, ToolMessage from langgraph.types import Command from src.agents.common import load_chat_model -from src.agents.common.toolkits.buildin import get_buildin_tools -from src.agents.common.toolkits.kbs import get_kb_based_tools +from src.agents.common.toolkits import get_all_tool_instances from src.services.mcp_service import get_enabled_mcp_tools from src.services.skill_resolver import ( SkillSessionSnapshot, @@ -48,10 +47,9 @@ class RuntimeConfigState(AgentState): class RuntimeConfigMiddleware(AgentMiddleware): - """运行时配置中间件 - 应用模型/工具/知识库/MCP/提示词配置 + """运行时配置中间件 - 应用模型/工具/MCP/提示词配置 - 注意：所有可能用到的知识库工具必须在初始化时预加载并注册到 self.tools - 运行时根据配置从 self.tools 中筛选工具，不能动态添加新工具 + 知识库工具已移至独立的 KnowledgeBaseMiddleware 支持自定义上下文字段名称，以便在不同场景（如主智能体/子智能体）使用不同的配置字段 """ @@ -106,10 +104,10 @@ class RuntimeConfigMiddleware(AgentMiddleware): self.tools: list[Any] = [] # 预加载工具列表（仅当启用工具覆盖时） + # 注意：知识库工具已移至独立的 KnowledgeBaseMiddleware if self.enable_tools_override: - self.kb_tools = get_kb_based_tools() - self.buildin_tools = get_buildin_tools() - self.tools = self.kb_tools + self.buildin_tools + (extra_tools or []) + self.base_tools = get_all_tool_instances() + self.tools = self.base_tools + (extra_tools or []) elif extra_tools: logger.warning( "RuntimeConfigMiddleware: extra_tools 参数已提供，但 enable_tools_override=False，" @@ -239,13 +237,7 @@ class RuntimeConfigMiddleware(AgentMiddleware): continue logger.warning(f"RuntimeConfigMiddleware: tool dependency not found, skip: {tool_name}") - # 2. 知识库工具 - knowledges = getattr(context, self.knowledges_context_name, None) - if knowledges: - kb_tools = get_kb_based_tools(db_names=knowledges) - selected_tools.extend(kb_tools) - - # 3. MCP 工具（使用统一入口，自动过滤 disabled_tools） + # 2. MCP 工具（使用统一入口，自动过滤 disabled_tools） mcps = getattr(context, self.mcps_context_name, None) or [] all_mcp_names: list[str] = [] for server_name in mcps: diff --git a/src/agents/common/toolkits/__init__.py b/src/agents/common/toolkits/__init__.py index 5a1be101..1eacd85a 100644 --- a/src/agents/common/toolkits/__init__.py +++ b/src/agents/common/toolkits/__init__.py @@ -4,51 +4,23 @@ from .registry import ( get_all_extra_metadata, get_all_tool_instances, get_extra_metadata, - register_tool, tool, ) # 工具获取函数 -from .buildin import get_buildin_tools -from .kbs import get_kb_based_tools -from src.services.mcp_service import get_enabled_mcp_tools +from .kbs import get_common_kb_tools - -async def get_tools_from_context(context, extra_tools=None): - """从上下文配置中获取工具列表""" - # 1. 基础工具 (从 context.tools 中筛选) - all_basic_tools = get_buildin_tools() + (extra_tools or []) - selected_tools = [] - - if context.tools: - # 创建工具映射表 - tools_map = {t.name: t for t in all_basic_tools} - for tool_name in context.tools: - if tool_name in tools_map: - selected_tools.append(tools_map[tool_name]) - - # 2. 知识库工具 - if context.knowledges: - kb_tools = get_kb_based_tools(db_names=context.knowledges) - selected_tools.extend(kb_tools) - - # 3. MCP 工具（使用统一入口，自动过滤 disabled_tools） - if context.mcps: - for server_name in context.mcps: - mcp_tools = await get_enabled_mcp_tools(server_name) - selected_tools.extend(mcp_tools) - - return selected_tools +# 触发各模块的 @tool 装饰器执行，自动注册工具 +from . import buildin +from . import debug +from . import mysql __all__ = [ - "register_tool", "get_extra_metadata", "get_all_extra_metadata", "get_all_tool_instances", "ToolExtraMetadata", "tool", - "get_buildin_tools", - "get_kb_based_tools", - "get_tools_from_context", + "get_common_kb_tools", ] diff --git a/src/agents/common/toolkits/buildin/__init__.py b/src/agents/common/toolkits/buildin/__init__.py index 1321c601..3c8bb64a 100644 --- a/src/agents/common/toolkits/buildin/__init__.py +++ b/src/agents/common/toolkits/buildin/__init__.py @@ -1,7 +1,7 @@ # buildin 工具包 -from .tools import get_buildin_tools, query_knowledge_graph +from .tools import calculator, query_knowledge_graph __all__ = [ - "get_buildin_tools", + "calculator", "query_knowledge_graph", ] diff --git a/src/agents/common/toolkits/buildin/tools.py b/src/agents/common/toolkits/buildin/tools.py index 1efed55e..1d741d6e 100644 --- a/src/agents/common/toolkits/buildin/tools.py +++ b/src/agents/common/toolkits/buildin/tools.py @@ -1,33 +1,48 @@ -import os import traceback -import uuid from typing import Annotated, Any -import requests -from langgraph.types import interrupt - from src import config, graph_base from src.agents.common.toolkits import tool -from src.storage.minio import aupload_file_to_minio +from src.agents.common.toolkits.registry import _all_tool_instances, _extra_registry, ToolExtraMetadata from src.utils import logger -# Lazy initialization for TavilySearch (only when TAVILY_API_KEY is available) +# Lazy initialization for TavilySearch (only when API key is available) _tavily_search_instance = None -def get_tavily_search(): - """Get TavilySearch instance lazily, only when API key is available.""" +def _create_tavily_search(): + """Create and register TavilySearch tool with metadata.""" global _tavily_search_instance - if _tavily_search_instance is None and config.enable_web_search: + if _tavily_search_instance is None: from langchain_tavily import TavilySearch _tavily_search_instance = TavilySearch() - _tavily_search_instance.metadata = {"name": "Tavily 网页搜索", "category": "buildin", "tags": ["搜索"]} - # 即使没有配置 API_KEY 也返回实例，调用时会出错 return _tavily_search_instance +# 注册 TavilySearch 工具（延迟初始化） +def _register_tavily_tool(): + """Register TavilySearch tool with extra metadata.""" + tavily_instance = _create_tavily_search() + # 手动注册到全局注册表 + _extra_registry["tavily_search"] = ToolExtraMetadata( + category="buildin", + tags=["搜索"], + display_name="Tavily 网页搜索", + ) + # 添加到工具实例列表 + _all_tool_instances.append(tavily_instance) + + +# 模块加载时注册 +if config.enable_web_search: + try: + _register_tavily_tool() + except Exception as e: + logger.warning(f"Failed to register TavilySearch tool: {e}") + + @tool(category="buildin", tags=["计算"], display_name="计算器") def calculator(a: float, b: float, operation: str) -> float: """计算器：对给定的2个数字进行基本数学运算""" @@ -49,81 +64,6 @@ def calculator(a: float, b: float, operation: str) -> float: raise -@tool(category="buildin", tags=["图片", "测试"], display_name="文生图测试") -async def text_to_img_demo(text: str) -> str: - """【测试用】使用模型生成图片， 会返回图片的URL""" - - url = "https://api.siliconflow.cn/v1/images/generations" - - payload = { - "model": "Qwen/Qwen-Image", - "prompt": text, - } - headers = {"Authorization": f"Bearer {os.getenv('SILICONFLOW_API_KEY')}", "Content-Type": "application/json"} - - try: - response = requests.post(url, json=payload, headers=headers) - response_json = response.json() - except Exception as e: - logger.error(f"Failed to generate image with: {e}") - raise ValueError(f"Image generation failed: {e}") - - try: - image_url = response_json["images"][0]["url"] - except (KeyError, IndexError, TypeError) as e: - logger.error(f"Failed to parse image URL from response: {e}, {response_json=}") - raise ValueError(f"Image URL extraction failed: {e}") - - # 2. Upload to MinIO (Simplified) - response = requests.get(image_url) - file_data = response.content - - file_name = f"{uuid.uuid4()}.jpg" - image_url = await aupload_file_to_minio( - bucket_name="generated-images", file_name=file_name, data=file_data, file_extension="jpg" - ) - logger.info(f"Image uploaded. URL: {image_url}") - return image_url - - -@tool(category="debug", tags=["内置", "审批"], display_name="人工审批") -def get_approved_user_goal( - operation_description: str, -) -> dict: - """ - 请求人工审批，在执行重要操作前获得人类确认。 - - Args: - operation_description: 需要审批的操作描述，例如 "调用知识库工具" - Returns: - dict: 包含审批结果的字典，格式为 {"approved": bool, "message": str} - """ - # 构建详细的中断信息 - interrupt_info = { - "question": "是否批准以下操作？", - "operation": operation_description, - } - - # 触发人工审批 - is_approved = interrupt(interrupt_info) - - # 返回审批结果 - if is_approved: - result = { - "approved": True, - "message": f"✅ 操作已批准：{operation_description}", - } - print(f"✅ 人工审批通过: {operation_description}") - else: - result = { - "approved": False, - "message": f"❌ 操作被拒绝：{operation_description}", - } - print(f"❌ 人工审批被拒绝: {operation_description}") - - return result - - KG_QUERY_DESCRIPTION = """ 使用这个工具可以查询知识图谱中包含的三元组信息。 关键词（query），使用可能帮助回答这个问题的关键词进行查询，不要直接使用用户的原始输入去查询。 @@ -144,18 +84,3 @@ def query_knowledge_graph(query: Annotated[str, "The keyword to query knowledge except Exception as e: logger.error(f"Knowledge graph query error: {e}, {traceback.format_exc()}") return f"知识图谱查询失败: {str(e)}" - - -def get_buildin_tools() -> list: - """获取内置工具列表""" - static_tools = [ - query_knowledge_graph, - get_approved_user_goal, - calculator, - text_to_img_demo, - ] - - # 始终添加 tavily_search，无论是否配置 API_KEY（调用时会出错） - static_tools.append(get_tavily_search()) - - return static_tools diff --git a/src/agents/common/toolkits/debug/__init__.py b/src/agents/common/toolkits/debug/__init__.py new file mode 100644 index 00000000..8e9ade69 --- /dev/null +++ b/src/agents/common/toolkits/debug/__init__.py @@ -0,0 +1,7 @@ +# debug 工具包 +from .tools import get_approved_user_goal, text_to_img_demo + +__all__ = [ + "get_approved_user_goal", + "text_to_img_demo", +] diff --git a/src/agents/common/toolkits/debug/tools.py b/src/agents/common/toolkits/debug/tools.py new file mode 100644 index 00000000..8dfb1467 --- /dev/null +++ b/src/agents/common/toolkits/debug/tools.py @@ -0,0 +1,86 @@ +import os +import uuid +from typing import Any + +import requests +from langgraph.types import interrupt + +from src import config +from src.agents.common.toolkits import tool +from src.storage.minio import aupload_file_to_minio +from src.utils import logger + + +@tool(category="debug", tags=["图片", "测试"], display_name="文生图测试") +async def text_to_img_demo(text: str) -> str: + """【测试用】使用模型生成图片， 会返回图片的URL""" + + url = "https://api.siliconflow.cn/v1/images/generations" + + payload = { + "model": "Qwen/Qwen-Image", + "prompt": text, + } + headers = {"Authorization": f"Bearer {os.getenv('SILICONFLOW_API_KEY')}", "Content-Type": "application/json"} + + try: + response = requests.post(url, json=payload, headers=headers) + response_json = response.json() + except Exception as e: + logger.error(f"Failed to generate image with: {e}") + raise ValueError(f"Image generation failed: {e}") + + try: + image_url = response_json["images"][0]["url"] + except (KeyError, IndexError, TypeError) as e: + logger.error(f"Failed to parse image URL from response: {e}, {response_json=}") + raise ValueError(f"Image URL extraction failed: {e}") + + # 2. Upload to MinIO (Simplified) + response = requests.get(image_url) + file_data = response.content + + file_name = f"{uuid.uuid4()}.jpg" + image_url = await aupload_file_to_minio( + bucket_name="generated-images", file_name=file_name, data=file_data, file_extension="jpg" + ) + logger.info(f"Image uploaded. URL: {image_url}") + return image_url + + +@tool(category="debug", tags=["内置", "审批"], display_name="人工审批") +def get_approved_user_goal( + operation_description: str, +) -> dict: + """ + 请求人工审批，在执行重要操作前获得人类确认。 + + Args: + operation_description: 需要审批的操作描述，例如 "调用知识库工具" + Returns: + dict: 包含审批结果的字典，格式为 {"approved": bool, "message": str} + """ + # 构建详细的中断信息 + interrupt_info = { + "question": "是否批准以下操作？", + "operation": operation_description, + } + + # 触发人工审批 + is_approved = interrupt(interrupt_info) + + # 返回审批结果 + if is_approved: + result = { + "approved": True, + "message": f"✅ 操作已批准：{operation_description}", + } + print(f"✅ 人工审批通过: {operation_description}") + else: + result = { + "approved": False, + "message": f"❌ 操作被拒绝：{operation_description}", + } + print(f"❌ 人工审批被拒绝: {operation_description}") + + return result diff --git a/src/agents/common/toolkits/kbs/__init__.py b/src/agents/common/toolkits/kbs/__init__.py index 531dcc91..d771ed2b 100644 --- a/src/agents/common/toolkits/kbs/__init__.py +++ b/src/agents/common/toolkits/kbs/__init__.py @@ -1,3 +1,5 @@ -from .tools import get_kb_based_tools +from .tools import ( + get_common_kb_tools, +) -__all__ = ["get_kb_based_tools"] +__all__ = ["get_common_kb_tools"] diff --git a/src/agents/common/toolkits/kbs/tools.py b/src/agents/common/toolkits/kbs/tools.py index 2cf5c0fa..26654658 100644 --- a/src/agents/common/toolkits/kbs/tools.py +++ b/src/agents/common/toolkits/kbs/tools.py @@ -1,159 +1,224 @@ """知识库工具模块""" import inspect -import traceback from typing import Any -from langchain_core.tools import StructuredTool +from langchain_core.tools import tool +from langgraph.prebuilt.tool_node import ToolRuntime from pydantic import BaseModel, Field from src import knowledge_base from src.utils import logger -class KnowledgeRetrieverModel(BaseModel): - query_text: str | None = Field( - default=None, +# ========== 通用知识库工具函数 ========== + + +class ListKBsInput(BaseModel): + """列出用户可访问的知识库输入模型""" + # Langchain 的 runtime 注入机制要求必须有参数 + dummy: str = Field(default="", description="Dummy parameter - ignore") # Add this + + +@tool(args_schema=ListKBsInput) +async def list_kbs(dummy: str, runtime: ToolRuntime) -> str: # Now has 2 params + """列出当前用户可访问的知识库列表 + + 返回用户基于权限可访问的知识库名称列表。这个列表是根据用户的角色和部门信息过滤后的结果， + 但不包括用户在当前对话中未启用的知识库。 + + Returns: + 用户可访问的知识库名称列表（字符串格式） + """ + # 从 runtime.context 获取用户信息 + runtime_context = runtime.context + user_id = getattr(runtime_context, "user_id", None) + if not user_id: + return "无法获取用户信息" + + # 打印 runtime—context 中的所有信息以进行调试 + logger.debug(f"Runtime context: {runtime_context.__dict__}") + + # 获取用户在当前对话中启用的知识库列表 + enabled_kb_names = getattr(runtime_context, "knowledges", []) or [] + + # 获取用户可访问的知识库列表（包含名称和描述） + try: + result = await knowledge_base.get_databases_by_raw_id(user_id) + all_kbs = result.get("databases", []) + except Exception as e: + logger.error(f"获取用户知识库列表失败: {e}") + return f"获取知识库列表失败: {str(e)}" + + all_kb_names = [kb["name"] for kb in all_kbs] + + logger.debug(f"用户 {user_id} 可访问的知识库列表: {all_kb_names}") + logger.debug(f"用户 {user_id} 当前对话启用的知识库列表: {enabled_kb_names}") + + # 与启用的知识库取交集 + available_kbs = [kb for kb in all_kbs if kb["name"] in enabled_kb_names] + + if not available_kbs: + return "当前没有可访问的知识库" + + # 格式化输出（包含名称和描述） + kb_list = [] + for kb in available_kbs: + name = kb.get("name", "") + desc = kb.get("description") or "无描述" + kb_list.append({"name": name, "description": desc}) + + return kb_list + + +class GetMindmapInput(BaseModel): + """获取思维导图输入模型""" + + kb_name: str = Field( + description="知识库名称，用于指定要获取思维导图的知识库" + ) + + +@tool(args_schema=GetMindmapInput) +async def get_mindmap(kb_name: str, runtime: ToolRuntime) -> str: + """获取指定知识库的思维导图结构 + + 当用户想要了解知识库的整体结构、文件分类、知识架构时使用此工具。 + 返回知识库的思维导图层级结构。 + + Args: + kb_name: 知识库名称 + + Returns: + 知识库的思维导图结构（文本格式） + """ + if not kb_name: + return "请提供知识库名称" + + # 获取所有检索器 + retrievers = knowledge_base.get_retrievers() + + # 查找对应的知识库 + target_db_id = None + target_info = None + for db_id, info in retrievers.items(): + if info["name"] == kb_name: + target_db_id = db_id + target_info = info + break + + if not target_db_id: + return f"知识库 '{kb_name}' 不存在" + + try: + from src.repositories.knowledge_base_repository import KnowledgeBaseRepository + + kb_repo = KnowledgeBaseRepository() + kb = await kb_repo.get_by_id(target_db_id) + + if kb is None: + return f"知识库 {target_info['name']} 不存在" + + mindmap_data = kb.mindmap + + if not mindmap_data: + return f"知识库 {target_info['name']} 还没有生成思维导图。" + + # 将思维导图数据转换为文本格式 + def mindmap_to_text(node, level=0): + """递归将思维导图JSON转换为层级文本""" + indent = " " * level + text = f"{indent}- {node.get('content', '')}\n" + for child in node.get("children", []): + text += mindmap_to_text(child, level + 1) + return text + + mindmap_text = f"知识库 {target_info['name']} 的思维导图结构：\n\n" + mindmap_text += mindmap_to_text(mindmap_data) + + return mindmap_text + + except Exception as e: + logger.error(f"获取思维导图失败: {e}") + return f"获取思维导图失败: {str(e)}" + + +class QueryKBInput(BaseModel): + """知识库检索输入模型""" + + kb_name: str = Field( + description="知识库名称，用于指定要在哪个知识库中检索" + ) + query_text: str = Field( description=( - "查询的关键词，查询的时候，应该尽量以可能帮助回答这个问题的关键词进行查询，不要直接使用用户的原始输入去查询。" + "查询的关键词，查询的时候，应该尽量以可能帮助回答这个问题的关键词进行查询，" + "不要直接使用用户的原始输入去查询。" ) ) - operation: str = Field( - default="search", - description=( - "操作类型：'search' 表示检索知识库内容，'get_mindmap' 表示获取知识库的思维导图结构。" - "当用户询问知识库的整体结构、文件分类、知识架构时，使用 'get_mindmap'。" - "当用户需要查询具体内容时，使用 'search'。" - ), - ) - - -class CommonKnowledgeRetriever(KnowledgeRetrieverModel): - """Common knowledge retriever model.""" - file_name: str | None = Field( default=None, description=( - "（非必要不启用此参数，留空即可）当操作类型为 'search' 且已经读取思维导图之后，可以指定文件关键词，支持模糊匹配。\n" + "（非必要不启用此参数，留空即可）当已经读取思维导图之后，可以指定文件关键词，支持模糊匹配。\n" "仅当检索结果过多且不相关，需要进一步缩小范围时使用。" ) ) -def get_kb_based_tools(db_names: list[str] | None = None) -> list: - """获取所有知识库基于的工具""" - # 获取所有知识库 - kb_tools = [] +@tool(args_schema=QueryKBInput) +async def query_kb(kb_name: str, query_text: str, file_name: str | None = None, runtime: ToolRuntime = None) -> Any: + """在指定知识库中检索内容 + + 当用户需要查询具体内容时使用此工具。根据关键词在知识库中检索相关文档片段。 + + Args: + kb_name: 知识库名称 + query_text: 查询的关键词 + file_name: （可选）文件名称过滤 + + Returns: + 检索结果 + """ + if not kb_name: + return "请提供知识库名称" + if not query_text: + return "请提供查询内容" + + # 获取所有检索器 retrievers = knowledge_base.get_retrievers() - if db_names is None: - db_ids = None - else: - db_ids = [kb_id for kb_id, kb in retrievers.items() if kb["name"] in db_names] - def _create_retriever_wrapper(db_id: str, retriever_info: dict[str, Any]): - """创建检索器包装函数的工厂函数，避免闭包变量捕获问题""" + # 查找对应的知识库 + target_info = None + for db_id, info in retrievers.items(): + if info["name"] == kb_name: + target_info = info + break - async def async_retriever_wrapper( - query_text: str, operation: str = "search", file_name: str | None = None - ) -> Any: - """异步检索器包装函数，支持检索和获取思维导图""" + if not target_info: + return f"知识库 '{kb_name}' 不存在" - # 获取思维导图 - if operation == "get_mindmap": - try: - logger.debug(f"Getting mindmap for database {db_id}") + try: + retriever = target_info["retriever"] + kwargs = {} + if file_name: + kwargs["file_name"] = file_name - from src.repositories.knowledge_base_repository import KnowledgeBaseRepository + if inspect.iscoroutinefunction(retriever): + result = await retriever(query_text, **kwargs) + else: + result = retriever(query_text, **kwargs) - kb_repo = KnowledgeBaseRepository() - kb = await kb_repo.get_by_id(db_id) + return result - if kb is None: - return f"知识库 {retriever_info['name']} 不存在" + except Exception as e: + logger.error(f"检索失败: {e}") + return f"检索失败: {str(e)}" - mindmap_data = kb.mindmap - if not mindmap_data: - return f"知识库 {retriever_info['name']} 还没有生成思维导图。" +def get_common_kb_tools() -> list: + """获取通用知识库工具列表 - # 将思维导图数据转换为文本格式，便于AI理解 - def mindmap_to_text(node, level=0): - """递归将思维导图JSON转换为层级文本""" - indent = " " * level - text = f"{indent}- {node.get('content', '')}\n" - for child in node.get("children", []): - text += mindmap_to_text(child, level + 1) - return text - - mindmap_text = f"知识库 {retriever_info['name']} 的思维导图结构：\n\n" - mindmap_text += mindmap_to_text(mindmap_data) - - logger.debug(f"Successfully retrieved mindmap for {db_id}") - return mindmap_text - - except Exception as e: - logger.error(f"Error getting mindmap for {db_id}: {e}") - return f"获取思维导图失败: {str(e)}" - - # 默认：检索知识库 - retriever = retriever_info["retriever"] - try: - logger.debug(f"Retrieving from database {db_id} with query: {query_text}") - kwargs = {} - if file_name: - kwargs["file_name"] = file_name - - if inspect.iscoroutinefunction(retriever): - result = await retriever(query_text, **kwargs) - else: - result = retriever(query_text, **kwargs) - logger.debug(f"Retrieved {len(result) if isinstance(result, list) else 'N/A'} results from {db_id}") - return result - except Exception as e: - logger.error(f"Error in retriever {db_id}: {e}") - return f"检索失败: {str(e)}" - - return async_retriever_wrapper - - for db_id, retrieve_info in retrievers.items(): - if db_ids is not None and db_id not in db_ids: - continue - - try: - # 构建工具描述 - description = ( - f"使用 {retrieve_info['name']} 知识库的多功能工具。\n" - f"知识库描述：{retrieve_info['description'] or '没有描述。'}\n\n" - f"支持的操作：\n" - f"1. 'search' - 检索知识库内容：根据关键词查询相关文档片段\n" - f"2. 'get_mindmap' - 获取思维导图：查看知识库的整体结构和文件分类\n\n" - f"使用建议：\n" - f"- 需要查询具体内容时，使用 operation='search'\n" - f"- 想了解知识库结构、文件分类时，使用 operation='get_mindmap'" - ) - - # 使用工厂函数创建检索器包装函数，避免闭包问题 - retriever_wrapper = _create_retriever_wrapper(db_id, retrieve_info) - - safename = retrieve_info["name"].replace(" ", "_")[:20] - - args_schema = KnowledgeRetrieverModel - if retrieve_info["metadata"]["kb_type"] in ["milvus"]: - args_schema = CommonKnowledgeRetriever - - # 使用 StructuredTool.from_function 创建异步工具 - tool = StructuredTool.from_function( - coroutine=retriever_wrapper, - name=safename, - description=description, - args_schema=args_schema, - metadata=retrieve_info["metadata"] | {"tag": ["knowledgebase"]}, - ) - - kb_tools.append(tool) - # logger.debug(f"Successfully created tool {tool_id} for database {db_id}") - - except Exception as e: - logger.error(f"Failed to create tool for database {db_id}: {e}, \n{traceback.format_exc()}") - continue - - return kb_tools + 返回 3 个通用工具： + - list_kbs: 列出用户可访问的知识库 + - get_mindmap: 获取指定知识库的思维导图 + - query_kb: 在指定知识库中检索 + """ + return [list_kbs, get_mindmap, query_kb] diff --git a/src/agents/common/toolkits/registry.py b/src/agents/common/toolkits/registry.py index e7ed03c9..ee0eee6f 100644 --- a/src/agents/common/toolkits/registry.py +++ b/src/agents/common/toolkits/registry.py @@ -19,36 +19,6 @@ _extra_registry: dict[str, ToolExtraMetadata] = {} _all_tool_instances: list = [] -def register_tool( - category: str, - tags: list[str] = None, - display_name: str = "", - icon: str = "", -): - """装饰器：注册工具附加元数据 - - 使用方式: - @register_tool(category="buildin", tags=["内置"], display_name="计算器") - def calculator(...): - ... - """ - - def decorator(func_or_tool: Callable) -> Callable: - # 获取工具名称（优先用 metadata.name） - tool_name = getattr(func_or_tool, "name", None) or getattr( - func_or_tool, "__name__", "" - ) - _extra_registry[tool_name] = ToolExtraMetadata( - category=category, - tags=tags or [], - display_name=display_name, - icon=icon, - ) - return func_or_tool - - return decorator - - def get_extra_metadata(tool_name: str) -> ToolExtraMetadata | None: """获取工具附加元数据""" return _extra_registry.get(tool_name) diff --git a/src/agents/deep_agent/graph.py b/src/agents/deep_agent/graph.py index 59461f9d..fdca8b6e 100644 --- a/src/agents/deep_agent/graph.py +++ b/src/agents/deep_agent/graph.py @@ -11,8 +11,9 @@ from langchain.agents.middleware import ( from src.agents.common import BaseAgent, load_chat_model from src.agents.common.backends import create_agent_composite_backend +from src.agents.common.middlewares.knowledge_base_middleware import KnowledgeBaseMiddleware from src.agents.common.middlewares import RuntimeConfigMiddleware, SummaryOffloadMiddleware, save_attachments_to_fs -from src.agents.common.toolkits.buildin.tools import get_tavily_search +from src.agents.common.toolkits.buildin.tools import _create_tavily_search from src.services.mcp_service import get_tools_from_all_servers from src.utils import logger @@ -79,19 +80,16 @@ class DeepAgent(BaseAgent): async def get_tools(self): """返回 Deep Agent 的专用工具""" - tools = [] - tavily_search = get_tavily_search() - if tavily_search: - tools.append(tavily_search) + from src import config + + tools = [] + if config.enable_web_search: + tavily = _create_tavily_search() + if tavily: + tools.append(tavily) - # # Assert that search tool is available for DeepAgent - # assert tools, ( - # "DeepAgent requires at least one search tool. " - # "Please configure TAVILY_API_KEY environment variable to enable web search." - # ) if not tools: logger.warning("No search tools configured, DeepAgent will work without web search") - tools = [] return tools async def get_graph(self, **kwargs): @@ -159,6 +157,7 @@ class DeepAgent(BaseAgent): save_attachments_to_fs, # 附件注入提示词 TodoListMiddleware(), PatchToolCallsMiddleware(), + KnowledgeBaseMiddleware(), # 知识库工具 subagents_middleware, summary_middleware, # 工具调用限制：tavily_search 总调用最多 20 次 diff --git a/src/knowledge/manager.py b/src/knowledge/manager.py index a9dc4fee..ac4f4a65 100644 --- a/src/knowledge/manager.py +++ b/src/knowledge/manager.py @@ -7,6 +7,7 @@ from src.knowledge.chunking.ragflow_like.presets import ( ensure_chunk_defaults_in_additional_params, ) from src.knowledge.factory import KnowledgeBaseFactory +from src.storage.postgres.models_business import User from src.utils import logger from src.utils.datetime_utils import utc_isoformat @@ -243,19 +244,46 @@ class KnowledgeBaseManager: return user_department_id in accessible_departments - async def get_databases_by_user(self, user: dict) -> dict: - """根据用户权限获取知识库列表 + async def get_databases_by_raw_id(self, user_id: int) -> dict: + """根据用户ID获取知识库列表（原始ID版本，兼容旧接口）""" + from src.repositories.user_repository import UserRepository - Args: - user: 用户信息字典，包含 role 和 department_id + # 通过数据库获取用户信息 + user_repo = UserRepository() + user: User | None = await user_repo.get_by_id(id=int(user_id)) + if not user: + logger.warning(f"User not found: {user_id}") + return {"databases": []} + return await self.get_databases_by_user(user) + + + async def get_databases_by_user_id(self, user_id: str) -> dict: + """根据用户ID获取知识库列表（字符串ID版本）""" + from src.repositories.user_repository import UserRepository + + # 通过数据库获取用户信息 + user_repo = UserRepository() + user: User | None = await user_repo.get_by_user_id(user_id) + if not user: + logger.warning(f"User not found: {user_id}") + return {"databases": []} + return await self.get_databases_by_user(user) + + async def get_databases_by_user(self, user: User) -> dict: + """根据用户权限获取知识库列表""" + + # 构建用户信息字典 + user_info = { + "role": user.role, + "department_id": user.department_id, + } + + logger.info(f"Getting databases for user {user.id} with role {user.role} and department {user.department_id}") - Returns: - 过滤后的知识库列表 - """ all_databases = (await self.get_databases()).get("databases", []) # 超级管理员可以看到所有知识库 - if user.get("role") == "superadmin": + if user_info.get("role") == "superadmin": return {"databases": all_databases} filtered_databases = [] @@ -265,7 +293,7 @@ class KnowledgeBaseManager: if not db_id: continue - if await self.check_accessible(user, db_id): + if await self.check_accessible(user_info, db_id): filtered_databases.append(db) return {"databases": filtered_databases} diff --git a/src/services/chat_stream_service.py b/src/services/chat_stream_service.py index 1ac94927..9a697034 100644 --- a/src/services/chat_stream_service.py +++ b/src/services/chat_stream_service.py @@ -379,28 +379,6 @@ async def stream_agent_chat( # 先构建 langgraph_config langgraph_config = {"configurable": {"thread_id": thread_id, "user_id": user_id}} - # 注意：LangGraph 会自动从 checkpointer 恢复 state（包括 attachments 和 files） - # 无需手动加载或传递 - - # 根据用户权限过滤知识库 - requested_knowledge_names = input_context["agent_config"].get("knowledges") - logger.info(f"Requesting knowledges: {requested_knowledge_names}") - if requested_knowledge_names and isinstance(requested_knowledge_names, list) and requested_knowledge_names: - user_info = {"role": "user", "department_id": department_id} - accessible_databases = await knowledge_base.get_databases_by_user(user_info) - accessible_kb_names = { - db.get("name") - for db in accessible_databases.get("databases", []) - if isinstance(db, dict) and db.get("name") - } - logger.info(f"Accessible knowledges: {accessible_kb_names}") - - filtered_knowledge_names = [kb for kb in requested_knowledge_names if kb in accessible_kb_names] - blocked_knowledge_names = [kb for kb in requested_knowledge_names if kb not in accessible_kb_names] - if blocked_knowledge_names: - logger.warning(f"用户 {user_id} 无权访问知识库: {blocked_knowledge_names}, 已自动过滤") - input_context["agent_config"]["knowledges"] = filtered_knowledge_names - full_msg = None accumulated_content = [] async for msg, metadata in agent.stream_messages(messages, input_context=input_context): diff --git a/src/services/skill_service.py b/src/services/skill_service.py index c253c9d1..9a479b8d 100644 --- a/src/services/skill_service.py +++ b/src/services/skill_service.py @@ -78,13 +78,6 @@ def validate_skill_slug(slug: str) -> str: raise ValueError("无效 skill slug") return normalized - -def _get_buildin_tool_names() -> list[str]: - from src.agents.common.toolkits.buildin import get_buildin_tools - - return [tool.name for tool in get_buildin_tools()] - - def get_skills_root_dir() -> Path: root = Path(sys_config.save_dir) / "skills" root.mkdir(parents=True, exist_ok=True) diff --git a/test/test_runtime_config_middleware_skills.py b/test/test_runtime_config_middleware_skills.py index 67977d44..1f7332a9 100644 --- a/test/test_runtime_config_middleware_skills.py +++ b/test/test_runtime_config_middleware_skills.py @@ -242,11 +242,6 @@ async def test_awrap_tool_call_denies_invisible_skill(): @pytest.mark.asyncio async def test_model_call_injects_dependency_tools_and_mcps_after_activation(monkeypatch: pytest.MonkeyPatch): - monkeypatch.setattr( - runtime_middleware, - "get_buildin_tools", - lambda: [_FakeTool(name="calculator"), _FakeTool(name="dep-tool")], - ) monkeypatch.setattr(runtime_middleware, "get_kb_based_tools", lambda db_names=None: []) snapshot = _build_snapshot( diff --git a/web/src/components/ToolCallingResult/ToolCallRenderer.vue b/web/src/components/ToolCallingResult/ToolCallRenderer.vue index 6310c79b..918d78ab 100644 --- a/web/src/components/ToolCallingResult/ToolCallRenderer.vue +++ b/web/src/components/ToolCallingResult/ToolCallRenderer.vue @@ -5,11 +5,17 @@ - - + + + + + + + + - + @@ -21,31 +27,31 @@ - + - + - + - + - + - + - + - + - + @@ -54,11 +60,11 @@ + + diff --git a/web/src/components/ToolCallingResult/tools/KnowledgeBaseTool.vue b/web/src/components/ToolCallingResult/tools/KnowledgeBaseTool.vue deleted file mode 100644 index f39f8b38..00000000 --- a/web/src/components/ToolCallingResult/tools/KnowledgeBaseTool.vue +++ /dev/null @@ -1,127 +0,0 @@ - - - - - diff --git a/web/src/components/ToolCallingResult/tools/ListKbsTool.vue b/web/src/components/ToolCallingResult/tools/ListKbsTool.vue new file mode 100644 index 00000000..8ce294d9 --- /dev/null +++ b/web/src/components/ToolCallingResult/tools/ListKbsTool.vue @@ -0,0 +1,92 @@ + + + + + diff --git a/web/src/components/ToolCallingResult/tools/QueryKbTool.vue b/web/src/components/ToolCallingResult/tools/QueryKbTool.vue new file mode 100644 index 00000000..7306ab8c --- /dev/null +++ b/web/src/components/ToolCallingResult/tools/QueryKbTool.vue @@ -0,0 +1,70 @@ + + + + +