import gradio as gr
from datetime import datetime

# Simple demo logic for exploring Hugging Face Spaces
# Later we can replace this with Whisper + LLM APIs

def analyze_call(user_name, issue_type, customer_message):
    timestamp = datetime.now().strftime("%Y-%m-%d %H:%M:%S")

    ai_reply = (
        f"Hello {user_name}, thanks for contacting CryptoVoIP support. "
        f"I understand your issue is related to '{issue_type}'. "
        f"Based on your message: '{customer_message}', our AI suggests checking SIP registration, "
        f"network latency, and RTP firewall rules first."
    )

    summary = (
        f"Call Summary\n"
        f"- Customer: {user_name}\n"
        f"- Category: {issue_type}\n"
        f"- Time: {timestamp}\n"
        f"- Recommended next step: Verify FreeSWITCH logs and RTP media path."
    )

    return ai_reply, summary


with gr.Blocks(title="CryptoVoIP AI Support Demo") as demo:
    gr.Markdown("# 📞 CryptoVoIP AI Support Demo")
    gr.Markdown("Use this as your **first Hugging Face Space** to explore live AI web apps.")

    with gr.Row():
        user_name = gr.Textbox(label="Customer Name", placeholder="Enter customer name")
        issue_type = gr.Dropdown(
            ["SIP Registration", "One-way Audio", "Call Drop", "Video Call", "Billing"],
            label="Issue Type"
        )

    customer_message = gr.Textbox(
        label="Customer Problem",
        lines=5,
        placeholder="Describe the telecom / VoIP issue..."
    )

    run_btn = gr.Button("Analyze with AI")

    ai_reply = gr.Textbox(label="AI Suggested Response", lines=4)
    summary = gr.Textbox(label="Call Summary", lines=6)

    run_btn.click(
        fn=analyze_call,
        inputs=[user_name, issue_type, customer_message],
        outputs=[ai_reply, summary]
    )


demo.launch()