import gradio as gr
import os
from langchain_groq import ChatGroq
from langchain_core.messages import HumanMessage, AIMessage, SystemMessage

# 1. Initialize the Chat Model
# We use the specific Groq integration as described in LangChain's provider docs
llm = ChatGroq(
    groq_api_key=os.environ.get("GROQ_API"),
    model_name="llama-3.3-70b-versatile",
    temperature=0.7
)

# 2. Define the Chat Function
def chat_function(message, history):
    # The docs recommend using a list of specific message types
    conversation_messages = []
    
    # Add a System Message to define behavior
    conversation_messages.append(SystemMessage(content="You are a helpful assistant."))
    
    # Reconstruct history from Gradio's list
    for user_text, ai_text in history:
        if user_text:
            conversation_messages.append(HumanMessage(content=user_text))
        if ai_text:
            conversation_messages.append(AIMessage(content=ai_text))
            
    # Add the user's latest message
    conversation_messages.append(HumanMessage(content=message))
    
    # Invoke the model with the full list
    response = llm.invoke(conversation_messages)
    
    # Return the text content of the response
    return response.content

# 3. Launch the Interface
demo = gr.ChatInterface(
    fn=chat_function,
    title="🤖 Groq Chatbot",
    description="A simple chatbot using LangChain's message history structures."
)

if __name__ == "__main__":
    demo.launch()