agent.py

import os
from dotenv import load_dotenv
from supabase import create_client
from supabase.client import Client
from langgraph.graph import START, StateGraph, MessagesState
from langgraph.prebuilt import tools_condition, ToolNode
from langchain_core.messages import SystemMessage, HumanMessage
from langchain_core.tools import tool
from langchain_google_genai import ChatGoogleGenerativeAI
from langchain_groq import ChatGroq
from langchain_huggingface import ChatHuggingFace, HuggingFaceEndpoint, HuggingFaceEmbeddings
from langchain_community.tools.tavily_search import TavilySearchResults
from langchain_community.document_loaders import WikipediaLoader, ArxivLoader
from langchain_community.vectorstores import SupabaseVectorStore
from langchain.tools.retriever import create_retriever_tool

load_dotenv()

# Check environment variables
SUPABASE_URL = os.environ.get("SUPABASE_URL")
SUPABASE_SERVICE_KEY = os.environ.get("SUPABASE_SERVICE_KEY")

print(f"SUPABASE_URL: {SUPABASE_URL[:10]}..." if SUPABASE_URL else "SUPABASE_URL not set")
print(f"SUPABASE_SERVICE_KEY: {SUPABASE_SERVICE_KEY[:10]}..." if SUPABASE_SERVICE_KEY else "SUPABASE_SERVICE_KEY not set")


def get_supabase_client():
    if not SUPABASE_URL or not SUPABASE_SERVICE_KEY:
        raise ValueError("Supabase environment variables are missing.")
    return create_client(SUPABASE_URL, SUPABASE_SERVICE_KEY)


@tool
def multiply(a: int, b: int) -> int:
    """Multiply two integers."""
    return a * b

@tool
def add(a: int, b: int) -> int:
    """Add two integers."""
    return a + b

@tool
def subtract(a: int, b: int) -> int:
    """Subtract b from a."""
    return a - b

@tool
def divide(a: int, b: int) -> float:
    """Divide a by b."""
    if b == 0:
        raise ValueError("Cannot divide by zero.")
    return a / b

@tool
def modulus(a: int, b: int) -> int:
    """Modulo operation."""
    return a % b

@tool
def wiki_search(query: str) -> str:
    """Search Wikipedia for the query and return top results."""
    docs = WikipediaLoader(query=query, load_max_docs=2).load()
    return "\n\n---\n\n".join([doc.page_content for doc in docs])

@tool
def web_search(query: str) -> str:
    """Search the web and return top results."""
    docs = TavilySearchResults(max_results=3).invoke(query=query)
    return "\n\n---\n\n".join([doc.page_content for doc in docs])

@tool
def arvix_search(query: str) -> str:
    """Search Arxiv for the query and return excerpts."""
    docs = ArxivLoader(query=query, load_max_docs=3).load()
    return "\n\n---\n\n".join([doc.page_content[:1000] for doc in docs])


tools = [multiply, add, subtract, divide, modulus, wiki_search, web_search, arvix_search]

with open("system_prompt.txt", "r", encoding="utf-8") as f:
    system_prompt = f.read().strip()

if not system_prompt:
    print("Warning: system_prompt.txt is empty. Using default system prompt.")
    system_prompt = "You are a helpful assistant."

sys_msg = SystemMessage(content=system_prompt)


def build_graph(provider: str = "groq"):
    embeddings = HuggingFaceEmbeddings(model_name="sentence-transformers/all-mpnet-base-v2")
    supabase = get_supabase_client()
    vector_store = SupabaseVectorStore(
        client=supabase,
        embedding=embeddings,
        table_name="documents",
        query_name="match_documents_langchain",
    )
    retriever_tool = create_retriever_tool(
        retriever=vector_store.as_retriever(),
        name="Question Search",
        description="A tool to retrieve similar questions from a vector store.",
    )

    if provider == "google":
        llm = ChatGoogleGenerativeAI(model="gemini-2.0-flash", temperature=0)
    elif provider == "groq":
        llm = ChatGroq(model="qwen-qwq-32b", temperature=0)
    elif provider == "huggingface":
        llm = ChatHuggingFace(
            llm=HuggingFaceEndpoint(
                url="https://api-inference.huggingface.co/models/Meta-DeepLearning/llama-2-7b-chat-hf",
                temperature=0,
            ),
        )
    else:
        raise ValueError("Invalid provider specified")

    llm_with_tools = llm.bind_tools(tools)

    def assistant(state: MessagesState):
        try:
            print("Assistant received messages:")
            for m in state["messages"]:
                print(f"- {m.__class__.__name__}: {m.content[:100]}")

            result = llm_with_tools.invoke(state["messages"])
            print("LLM output message:")
            if hasattr(result, "content"):
                print(result.content[:500])
            else:
                print(result)

            if not result or not getattr(result, "content", None):
                print("Warning: LLM returned empty result or no content.")
                return {"messages": [HumanMessage(content="Sorry, I couldn't generate an answer.")]}
            
            return {"messages": [result]}
        except Exception as e:
            print(f"Error invoking LLM: {e}")
            return {"messages": [HumanMessage(content="Sorry, I encountered an error during processing.")]}

    def retriever(state: MessagesState):
        similar = vector_store.similarity_search(state["messages"][0].content)
        msg = HumanMessage(content=f"Similar question reference:\n\n{similar[0].page_content}")
        return {"messages": [sys_msg] + state["messages"] + [msg]}

    graph = StateGraph(MessagesState)
    graph.add_node("retriever", retriever)
    graph.add_node("assistant", assistant)
    graph.add_node("tools", ToolNode(tools))
    graph.add_edge(START, "retriever")
    graph.add_edge("retriever", "assistant")
    graph.add_conditional_edges("assistant", tools_condition)
    graph.add_edge("tools", "assistant")

    return graph.compile()


if __name__ == "__main__":
    g = build_graph("groq")
    question = "When was Aquinas added to Wikipedia page on double effect?"
    output = g.invoke({"messages": [HumanMessage(content=question)]})
    for msg in output["messages"]:
        print(f"\n[{msg.__class__.__name__}] {msg.content}\n")