190 lines
7.4 KiB
Python
190 lines
7.4 KiB
Python
"""Define the configurable parameters for the agent."""
|
||
|
||
import uuid
|
||
from dataclasses import MISSING, dataclass, field, fields
|
||
from typing import Annotated, get_args, get_origin
|
||
|
||
from yuxi import config as sys_config
|
||
|
||
|
||
@dataclass(kw_only=True)
|
||
class BaseContext:
|
||
"""
|
||
定义一个基础 Context 供 各类 graph 继承
|
||
|
||
配置优先级:
|
||
1. 运行时配置(RunnableConfig)：最高优先级，直接从函数参数传入
|
||
2. 类默认配置：最低优先级，类中定义的默认值
|
||
"""
|
||
|
||
def update(self, data: dict):
|
||
"""更新配置字段"""
|
||
for key, value in data.items():
|
||
if hasattr(self, key):
|
||
setattr(self, key, value)
|
||
|
||
thread_id: str = field(
|
||
default_factory=lambda: str(uuid.uuid4()),
|
||
metadata={"name": "线程ID", "configurable": False, "description": "用来唯一标识一个对话线程"},
|
||
)
|
||
|
||
user_id: str = field(
|
||
default_factory=lambda: str(uuid.uuid4()),
|
||
metadata={"name": "用户ID", "configurable": False, "description": "用来唯一标识一个用户"},
|
||
)
|
||
|
||
system_prompt: Annotated[str, {"__template_metadata__": {"kind": "prompt"}}] = field(
|
||
default="You are a helpful assistant.",
|
||
metadata={"name": "系统提示词", "description": "用来描述智能体的角色和行为"},
|
||
)
|
||
|
||
model: Annotated[str, {"__template_metadata__": {"kind": "llm"}}] = field(
|
||
default=sys_config.default_model,
|
||
metadata={
|
||
"name": "智能体模型",
|
||
"options": [],
|
||
"description": "智能体的驱动模型，建议选择 Agent 能力较强的模型，不建议使用小参数模型。",
|
||
},
|
||
)
|
||
|
||
tools: Annotated[list[str], {"__template_metadata__": {"kind": "tools"}}] = field(
|
||
default_factory=lambda: ["ask_user_question", "tavily_search"],
|
||
metadata={
|
||
"name": "工具",
|
||
"description": "内置的工具。",
|
||
},
|
||
)
|
||
|
||
knowledges: Annotated[list[str], {"__template_metadata__": {"kind": "knowledges"}}] = field(
|
||
default_factory=list,
|
||
metadata={
|
||
"name": "知识库",
|
||
"description": "知识库列表，可以在左侧知识库页面中创建知识库。",
|
||
"type": "list", # Explicitly mark as list type for frontend if needed
|
||
},
|
||
)
|
||
|
||
mcps: Annotated[list[str], {"__template_metadata__": {"kind": "mcps"}}] = field(
|
||
default_factory=list,
|
||
metadata={
|
||
"name": "MCP服务器",
|
||
"options": [],
|
||
"description": (
|
||
"MCP服务器列表，建议使用支持 SSE 的 MCP 服务器，"
|
||
"如果需要使用 uvx 或 npx 运行的服务器，也请在项目外部启动 MCP 服务器，并在项目中配置 MCP 服务器。"
|
||
),
|
||
},
|
||
)
|
||
|
||
skills: Annotated[list[str], {"__template_metadata__": {"kind": "skills"}}] = field(
|
||
default_factory=list,
|
||
metadata={
|
||
"name": "Skills",
|
||
"options": [],
|
||
"description": "可选技能列表（由超级管理员维护）。运行时仅挂载并只读暴露选中的 "
|
||
"skills。技能依赖的工具和 MCP 服务器也会被自动挂载。",
|
||
"type": "list",
|
||
},
|
||
)
|
||
|
||
subagents_model: Annotated[str, {"__template_metadata__": {"kind": "llm"}}] = field(
|
||
default=sys_config.default_model,
|
||
metadata={
|
||
"name": "子智能体的默认模型",
|
||
"description": "为所有子智能体设置默认模型，可在各子智能体配置中单独覆盖。",
|
||
},
|
||
)
|
||
|
||
subagents: Annotated[list[str], {"__template_metadata__": {"kind": "subagents"}}] = field(
|
||
default_factory=list,
|
||
metadata={
|
||
"name": "子智能体",
|
||
"options": [],
|
||
"description": "可选子智能体列表。为空表示不启用任何 SubAgent。但依然会启用一个 general-purpose 的子智能体",
|
||
"type": "list",
|
||
},
|
||
)
|
||
|
||
summary_threshold: int = field(
|
||
default=100,
|
||
metadata={
|
||
"name": "上下文摘要触发阈值 (KB)",
|
||
"description": "当上下文大小超过该值时，启用摘要功能以优化上下文使用。单位为 KB，默认值为 100KB。",
|
||
"type": "number",
|
||
},
|
||
)
|
||
|
||
@classmethod
|
||
def get_configurable_items(cls):
|
||
"""实现一个可配置的参数列表，在 UI 上配置时使用"""
|
||
configurable_items = {}
|
||
for f in fields(cls):
|
||
if f.init and not f.metadata.get("hide", False):
|
||
if f.metadata.get("configurable", True):
|
||
# 处理类型信息
|
||
field_type = f.type
|
||
type_name = cls._get_type_name(field_type)
|
||
|
||
# 提取 Annotated 的元数据
|
||
template_metadata = cls._extract_template_metadata(field_type)
|
||
|
||
options = f.metadata.get("options", [])
|
||
if callable(options):
|
||
options = options()
|
||
|
||
configurable_items[f.name] = {
|
||
"type": f.metadata.get("type", type_name),
|
||
"name": f.metadata.get("name", f.name),
|
||
"options": options,
|
||
"default": f.default
|
||
if f.default is not MISSING
|
||
else f.default_factory()
|
||
if f.default_factory is not MISSING
|
||
else None,
|
||
"description": f.metadata.get("description", ""),
|
||
"template_metadata": template_metadata, # Annotated 的额外元数据
|
||
}
|
||
|
||
return configurable_items
|
||
|
||
@classmethod
|
||
def _get_type_name(cls, field_type) -> str:
|
||
"""获取类型名称，处理 Annotated 类型"""
|
||
# 检查是否是 Annotated 类型
|
||
if get_origin(field_type) is not None:
|
||
# 处理泛型类型如 list[str], Annotated[str, {...}]
|
||
origin = get_origin(field_type)
|
||
if hasattr(origin, "__name__"):
|
||
if origin.__name__ == "Annotated":
|
||
# Annotated 类型，获取真实类型
|
||
args = get_args(field_type)
|
||
if args:
|
||
return cls._get_type_name(args[0]) # 递归处理真实类型
|
||
return origin.__name__
|
||
else:
|
||
return str(origin)
|
||
elif hasattr(field_type, "__name__"):
|
||
return field_type.__name__
|
||
else:
|
||
return str(field_type)
|
||
|
||
@classmethod
|
||
def _extract_template_metadata(cls, field_type) -> dict:
|
||
"""从 Annotated 类型中提取模板元数据"""
|
||
if get_origin(field_type) is not None:
|
||
origin = get_origin(field_type)
|
||
if hasattr(origin, "__name__") and origin.__name__ == "Annotated":
|
||
args = get_args(field_type)
|
||
if len(args) > 1:
|
||
# 查找包含 __template_metadata__ 的字典
|
||
for metadata in args[1:]:
|
||
if isinstance(metadata, dict) and "__template_metadata__" in metadata:
|
||
return metadata["__template_metadata__"]
|
||
return {}
|
||
|
||
def update_from_dict(self, data: dict):
|
||
"""从字典更新配置字段"""
|
||
for key, value in data.items():
|
||
if hasattr(self, key):
|
||
setattr(self, key, value)
|