app.py

from utils import _set_env


_set_env("OPENAI_API_KEY")


from utils import *

def create_graph():
    from langgraph.graph import StateGraph, START, END
    from langgraph.prebuilt import ToolNode, tools_condition


    ## ADD TRACKING
    response_model = init_chat_model("gpt-4o", temperature=0)
    grader_model = init_chat_model("gpt-4o", temperature=0)


    workflow = StateGraph(MessagesState)

    # Define the nodes we will cycle between
    workflow.add_node(generate_query_or_respond)
    workflow.add_node("retrieve", ToolNode([retriever_tool]))
    workflow.add_node(rewrite_question)
    workflow.add_node(generate_answer)

    workflow.add_edge(START, "generate_query_or_respond")

    # Decide whether to retrieve
    workflow.add_conditional_edges(
        "generate_query_or_respond",
        # Assess LLM decision (call `retriever_tool` tool or respond to the user)
        tools_condition,
        {
            # Translate the condition outputs to nodes in our graph
            "tools": "retrieve",
            END: END,
        },
    )

    # Edges taken after the `action` node is called.
    workflow.add_conditional_edges(
        "retrieve",
        # Assess agent decision
        grade_documents,
    )
    workflow.add_edge("generate_answer", END)
    workflow.add_edge("rewrite_question", "generate_query_or_respond")

    # Compile
    graph = workflow.compile()

    return graph




from langchain.schema import AIMessage, HumanMessage  
import gradio as gr
from langchain.chat_models import init_chat_model

## ADD TRACKING
response_model = init_chat_model("gpt-4o", temperature=0)
grader_model = init_chat_model("gpt-4o", temperature=0)

graph = create_graph()

def predict(message, history):
    history_langchain_format = []
    for msg in history:
        if msg['role'] == "user":
            history_langchain_format.append(HumanMessage(content=msg['content']))
        elif msg['role'] == "assistant":
            history_langchain_format.append(AIMessage(content=msg['content']))
    history_langchain_format.append(HumanMessage(content=message))

    
    gpt_response = graph.invoke(history_langchain_format)

    
    return gpt_response.content

iface = gr.ChatInterface(
    predict,
    api_name="chat",
)

iface.launch()