import gradio as gr
from transformers import pipeline, AutoTokenizer, AutoModelForSeq2SeqLM

# Load Ethical Inquiry and Learning Support Model
ethics_tokenizer = AutoTokenizer.from_pretrained("google/flan-t5-base")
ethics_model = AutoModelForSeq2SeqLM.from_pretrained("google/flan-t5-base")
ethics_pipeline = pipeline("text2text-generation", model=ethics_model, tokenizer=ethics_tokenizer)

# Query Classification
def classify_query(query):
    """Classify the query into Ethical Inquiry or Learning Support."""
    if "ethics" in query or "privacy" in query:
        return "ethical_inquiry"
    else:
        return "learning_support"

# Process Query
def handle_query(query):
    """Route the query to the appropriate model and generate a response."""
    task = classify_query(query)

    if task == "ethical_inquiry":
        try:
            # Ethical guidance response
            response = ethics_pipeline(query)
            return f"Ethical Inquiry Response: {response[0]['generated_text']}"
        except Exception as e:
            return f"Error in Ethical Inquiry: {e}"

    else:
        try:
            # Learning support or general question response
            response = ethics_pipeline(query)
            return f"Learning Support Response: {response[0]['generated_text']}"
        except Exception as e:
            return f"Error in Learning Support: {e}"

# Gradio Interface
def chatbot(query):
    return handle_query(query)

# Deploy with Gradio
interface = gr.Interface(
    fn=chatbot,
    inputs="text",
    outputs="text",
    title="BioSphere AI Chatbot",
    description="A chatbot for Ethical Guidance and Learning Support in Biotech.",
)

# Add Gemmini API Key Integration
def deploy_with_gemmini(api_key):
    print(f"Deploying using Gemmini API Key: {api_key}")
    interface.launch()

# Replace 'your_api_key' with your actual Gemmini API key
gemmini_api_key = "your_api_key"
deploy_with_gemmini(gemmini_api_key)