agent.py

import os
from typing import TypedDict, Annotated
from dotenv import load_dotenv
from langchain_core.messages import AnyMessage, HumanMessage, AIMessage, SystemMessage
from langgraph.prebuilt import ToolNode
from langgraph.graph import START, StateGraph, MessagesState
from langgraph.prebuilt import tools_condition
from langchain_huggingface import HuggingFaceEndpoint, ChatHuggingFace
from langchain_google_genai import ChatGoogleGenerativeAI
from langchain_groq import ChatGroq
from langchain_openai import ChatOpenAI
import requests
from tools import *

# load api key
load_dotenv()


def buildAgent(provider="google"):
    # load the system prompt from the file
    with open("system_prompt.txt", "r", encoding="utf-8") as f:
        system_prompt = f.read()
    print(system_prompt)

    # System message
    sys_msg = SystemMessage(content=system_prompt)

    # Generate the chat interface, including the tools
    if provider == "huggingface":
        llm = ChatHuggingFace(
            llm=HuggingFaceEndpoint(repo_id="Qwen/Qwen2.5-Coder-32B-Instruct"),
        )
    elif provider == "groq":
        llm = ChatGroq(model="qwen-qwq-32b")
    elif provider == "google":
        llm = ChatGoogleGenerativeAI(model="gemini-2.0-flash-exp")
    elif provider == "openrouter":
        llm = ChatOpenAI(
            base_url="https://openrouter.ai/api/v1",
            model="google/gemini-2.0-flash-001",
            api_key=os.getenv("OPENROUTER_API_KEY"),
        )
    else:
        raise ValueError("Invalid provider. Choose 'groq' or 'huggingface'.")

    agent_tools = [
        multiply,
        add,
        subtract,
        divide,
        modulus,
        power,
        square_root,
        web_search,
        wiki_search,
        arxiv_search,
        download_file,
    ]

    chat_with_tools = llm.bind_tools(agent_tools)

    # nodes
    def assistant(state: MessagesState):
        return {
            "messages": [chat_with_tools.invoke(state["messages"])],
        }

    # todo add rag
    def retriever(state: MessagesState):
        """Retriever node"""
        # Handle the case when no similar questions are found
        return {"messages": state["messages"]}

    ## The graph
    builder = StateGraph(MessagesState)
    # Define nodes: these do the work
    builder.add_node("retriever", retriever)
    builder.add_node("assistant", assistant)
    builder.add_node("tools", ToolNode(agent_tools))
    # Define edges: these determine how the control flow moves
    builder.add_edge(START, "retriever")
    builder.add_edge("retriever", "assistant")
    builder.add_conditional_edges(
        "assistant",
        # If the latest message requires a tool, route to tools
        # Otherwise, provide a direct response
        tools_condition,
    )
    builder.add_edge("tools", "assistant")
    return builder.compile()


if __name__ == "__main__":
    # random_question_url = "https://agents-course-unit4-scoring.hf.space/random-question"
    # response = requests.get(random_question_url, timeout=15)
    # questions_data = response.json()
    # question = questions_data.get("question")
    question = "How many studio albums were published by Mercedes Sosa between 2000 and 2009 (included)? You can use the latest 2022 version of english wikipedia."
    graph = buildAgent(provider="google")
    messages = [HumanMessage(content=question)]
    print(messages)
    messages = graph.invoke({"messages": messages})
    for m in messages["messages"]:
        m.pretty_print()