"""一个旨在进行对话并使用工具的代理程序。"""
from __future__ import annotations
from typing import Any, List, Optional, Sequence
from langchain_core._api import deprecated
from langchain_core.callbacks import BaseCallbackManager
from langchain_core.language_models import BaseLanguageModel
from langchain_core.prompts import PromptTemplate
from langchain_core.pydantic_v1 import Field
from langchain_core.tools import BaseTool
from langchain.agents.agent import Agent, AgentOutputParser
from langchain.agents.agent_types import AgentType
from langchain.agents.conversational.output_parser import ConvoOutputParser
from langchain.agents.conversational.prompt import FORMAT_INSTRUCTIONS, PREFIX, SUFFIX
from langchain.agents.utils import validate_tools_single_input
from langchain.chains import LLMChain
[docs]@deprecated("0.1.0", alternative="create_react_agent", removal="0.3.0")
class ConversationalAgent(Agent):
"""一个除了使用工具外还能进行对话的代理程序。"""
ai_prefix: str = "AI"
"""用于AI输出之前的前缀。"""
output_parser: AgentOutputParser = Field(default_factory=ConvoOutputParser)
"""代理的输出解析器。"""
@classmethod
def _get_default_output_parser(
cls, ai_prefix: str = "AI", **kwargs: Any
) -> AgentOutputParser:
return ConvoOutputParser(ai_prefix=ai_prefix)
@property
def _agent_type(self) -> str:
"""返回代理类型的标识符。"""
return AgentType.CONVERSATIONAL_REACT_DESCRIPTION
@property
def observation_prefix(self) -> str:
"""要附加到观测值前面的前缀。"""
return "Observation: "
@property
def llm_prefix(self) -> str:
"""用于在llm调用前附加的前缀。"""
return "Thought:"
[docs] @classmethod
def create_prompt(
cls,
tools: Sequence[BaseTool],
prefix: str = PREFIX,
suffix: str = SUFFIX,
format_instructions: str = FORMAT_INSTRUCTIONS,
ai_prefix: str = "AI",
human_prefix: str = "Human",
input_variables: Optional[List[str]] = None,
) -> PromptTemplate:
"""创建与零次代理相似的提示。
参数：
tools：代理将可以访问的工具列表，用于格式化提示。
prefix：工具列表之前要放置的字符串。
suffix：工具列表之后要放置的字符串。
ai_prefix：AI输出之前要使用的字符串。
human_prefix：人类输出之前要使用的字符串。
input_variables：最终提示将期望的输入变量列表。
返回：
从这里的各部分组装而成的PromptTemplate。
"""
tool_strings = "\n".join(
[f"> {tool.name}: {tool.description}" for tool in tools]
)
tool_names = ", ".join([tool.name for tool in tools])
format_instructions = format_instructions.format(
tool_names=tool_names, ai_prefix=ai_prefix, human_prefix=human_prefix
)
template = "\n\n".join([prefix, tool_strings, format_instructions, suffix])
if input_variables is None:
input_variables = ["input", "chat_history", "agent_scratchpad"]
return PromptTemplate(template=template, input_variables=input_variables)
@classmethod
def _validate_tools(cls, tools: Sequence[BaseTool]) -> None:
super()._validate_tools(tools)
validate_tools_single_input(cls.__name__, tools)