import os
import gradio as gr
from langchain_google_genai.chat_models import ChatGoogleGenerativeAI

# Set the path to the service account key
os.environ["GOOGLE_APPLICATION_CREDENTIALS"] = "./firm-catalyst-437006-s4-407500537db5.json"

# Initialize the LLM
llm = ChatGoogleGenerativeAI(model="gemini-1.5-pro")

# Chat function
def chat_with_gemini(user_input, chat_history=[]):
    try:
        # Prepare the prompt and include chat history
        conversation = "\n".join([f"User: {msg[0]}\nBot: {msg[1]}" for msg in chat_history])
        conversation += f"\nUser: {user_input}\nBot:"
        
        # Get response from the model
        response = llm.predict(conversation)  # Assuming 'predict' method is used for response generation
        
        # Update chat history
        chat_history.append((user_input, response))
        return response, chat_history
    except Exception as e:
        return f"Error: {str(e)}", chat_history

# Create Gradio interface
iface = gr.Interface(
    fn=chat_with_gemini,
    inputs=["text", "state"],  # Use "state" to maintain chat history
    outputs=["text", "state"],  # Return updated chat history
    title="Chatbot with Gemini 1.5",
    description="Ask me anything!",
    theme="default",  # Optional: Specify a theme if desired
    live=True
)

# Launch the interface with debugging enabled
iface.launch(debug=True)