src/streamlit_app.py

import os
import io
import streamlit as st
from dotenv import load_dotenv
from PIL import Image
import google.generativeai as genai
from langgraph.graph import StateGraph, END
from typing import TypedDict, List, Union

# ---------------------------
# Load API Key
# ---------------------------
load_dotenv()
API_KEY = os.getenv("GOOGLE_API_KEY")
genai.configure(api_key=API_KEY)

model = genai.GenerativeModel("gemini-2.0-flash")

# ---------------------------
# State Definition
# ---------------------------
class ChatState(TypedDict):
    user_input: str
    image: Union[Image.Image, None]
    raw_response: str
    final_response: str
    chat_history: List[dict]


# ---------------------------
# LangGraph Nodes
# ---------------------------
def input_node(state: ChatState) -> ChatState:
    return state


def processing_node(state: ChatState) -> ChatState:
    parts = [state["user_input"]]
    if state["image"]:
        parts.append(state["image"])

    try:
        chat = model.start_chat(history=[])
        resp = chat.send_message(parts)
        state["raw_response"] = resp.text
    except Exception as e:
        state["raw_response"] = f"Error: {e}"

    return state


def checking_node(state: ChatState) -> ChatState:
    raw = state["raw_response"]

    # Remove unnecessary lines from Gemini responses
    if raw.startswith("Sure!") or "The image shows" in raw:
        lines = raw.split("\n")
        filtered = [
            line for line in lines
            if not line.startswith("Sure!") and "The image shows" not in line
        ]
        final = "\n".join(filtered).strip()
        state["final_response"] = final
    else:
        state["final_response"] = raw

    # Save to session chat history
    st.session_state.chat_history.append({"role": "user", "content": state["user_input"]})
    st.session_state.chat_history.append({"role": "model", "content": state["final_response"]})

    return state


# ---------------------------
# Build the LangGraph
# ---------------------------
builder = StateGraph(ChatState)
builder.add_node("input", input_node)
builder.add_node("processing", processing_node)
builder.add_node("checking", checking_node)

builder.set_entry_point("input")
builder.add_edge("input", "processing")
builder.add_edge("processing", "checking")
builder.add_edge("checking", END)

graph = builder.compile()

# ---------------------------
# Streamlit UI Setup
# ---------------------------
st.set_page_config(page_title="Math Chatbot", layout="centered")
st.title("Math Chatbot")

# Initialize session state
if "chat_history" not in st.session_state:
    st.session_state.chat_history = []

# Display chat history
for msg in st.session_state.chat_history:
    with st.chat_message(msg["role"]):
        st.markdown(msg["content"])

# ---------------------------
# Sidebar
# ---------------------------
with st.sidebar:
    st.header("Options")
    if st.button("New Chat"):
        st.session_state.chat_history = []
        st.rerun()

# ---------------------------
# Chat Input Form
# ---------------------------
with st.form("chat_form", clear_on_submit=True):
    user_input = st.text_input("Your message:", placeholder="Ask your math problem here")

    uploaded_file = st.file_uploader("Upload an image", type=["jpg", "jpeg", "png"])
    
    submitted = st.form_submit_button("Send")

    if submitted:
        # Load image safely
        image = None
        if uploaded_file:
            try:
                image = Image.open(io.BytesIO(uploaded_file.read()))
            except Exception as e:
                st.error(f"Error loading image: {e}")
                st.stop()

        # Prepare state
        input_state = {
            "user_input": user_input,
            "image": image,
            "raw_response": "",
            "final_response": "",
            "chat_history": st.session_state.chat_history,
        }

        # Run LangGraph
        output = graph.invoke(input_state)

        # Show model response
        with st.chat_message("model"):
            st.markdown(output["final_response"])