langchain/templates/retrieval-agent/retrieval_agent/chain.py

import os
from typing import List, Tuple

from langchain.agents import AgentExecutor
from langchain.agents.format_scratchpad import format_to_openai_function_messages
from langchain.agents.output_parsers import OpenAIFunctionsAgentOutputParser
from langchain.chat_models import AzureChatOpenAI
from langchain.prompts import ChatPromptTemplate, MessagesPlaceholder
from langchain.pydantic_v1 import BaseModel, Field
from langchain.schema.messages import AIMessage, HumanMessage
from langchain.tools import ArxivQueryRun
from langchain.tools.render import format_tool_to_openai_function
from langchain.utilities import ArxivAPIWrapper


class ArxivInput(BaseModel):
    query: str = Field(description="search query to look up")


# Create the tool
arxiv_tool = ArxivQueryRun(api_wrapper=ArxivAPIWrapper(), args_schema=ArxivInput)
tools = [arxiv_tool]
llm = AzureChatOpenAI(
    temperature=0,
    deployment_name=os.environ["AZURE_OPENAI_DEPLOYMENT_NAME"],
    openai_api_base=os.environ["AZURE_OPENAI_API_BASE"],
    openai_api_version=os.environ["AZURE_OPENAI_API_VERSION"],
    openai_api_key=os.environ["AZURE_OPENAI_API_KEY"],
)
assistant_system_message = """You are a helpful research assistant. \
Lookup relevant information as needed."""
prompt = ChatPromptTemplate.from_messages(
    [
        ("system", assistant_system_message),
        MessagesPlaceholder(variable_name="chat_history"),
        ("user", "{input}"),
        MessagesPlaceholder(variable_name="agent_scratchpad"),
    ]
)

llm_with_tools = llm.bind(functions=[format_tool_to_openai_function(t) for t in tools])


def _format_chat_history(chat_history: List[Tuple[str, str]]):
    buffer = []
    for human, ai in chat_history:
        buffer.append(HumanMessage(content=human))
        buffer.append(AIMessage(content=ai))
    return buffer


agent = (
    {
        "input": lambda x: x["input"],
        "chat_history": lambda x: _format_chat_history(x["chat_history"]),
        "agent_scratchpad": lambda x: format_to_openai_function_messages(
            x["intermediate_steps"]
        ),
    }
    | prompt
    | llm_with_tools
    | OpenAIFunctionsAgentOutputParser()
)


class AgentInput(BaseModel):
    input: str
    chat_history: List[Tuple[str, str]] = Field(
        ..., extra={"widget": {"type": "chat", "input": "input", "output": "output"}}
    )


agent_executor = AgentExecutor(agent=agent, tools=tools, verbose=True).with_types(
    input_type=AgentInput
)