# app.py
import time
import random
import gradio as gr

# -------------------------
# Minimal HyperLayer demo
# - mode: choose demo behavior
# - latency: simulate zero / low / higher latency for demo
# - examples: quick tryouts
# -------------------------

def simulate_processing(prompt: str, mode: str, latency_ms: int):
    """
    Simulate HyperLayer AI agent processing.
    - prompt: user input
    - mode: 'Explain', 'Summarize', 'Detect Intent', 'Tokenize'
    - latency_ms: simulated processing time in milliseconds
    """
    # simulate processing delay
    time.sleep(max(0, latency_ms) / 1000.0)

    # create deterministic-ish fake outputs for demo
    if not prompt.strip():
        return "Please enter a prompt to see simulated HyperLayer output."

    base = f"[Mode: {mode}]"

    if mode == "Explain":
        out = f"{base} Explanation: HyperLayer interprets your input and returns a concise, technical summary.\n\nInput: {prompt}\n\nSummary: {prompt[:120]}... (simulated explanation)"
    elif mode == "Summarize":
        out = f"{base} Summary: {prompt[:140]}... (simulated short summary)"
    elif mode == "Detect Intent":
        intents = ["query_price", "execute_trade", "get_balance", "unknown"]
        detected = random.choice(intents)
        out = f"{base} Detected intent = `{detected}` (confidence: {random.uniform(0.6,0.99):.2f})"
    elif mode == "Tokenize":
        tokens = prompt.split()
        out = f"{base} Tokens ({len(tokens)}): " + ", ".join(tokens[:20])
    else:
        out = f"{base} Echo: {prompt}"

    # simulate metadata block
    meta = f"\n\n---\nSimulated latency: {latency_ms} ms • Node: x402-demo-01 • timestamp: {time.strftime('%Y-%m-%d %H:%M:%S')}"
    return out + meta


# Gradio UI details
title = "🛰️ HyperLayer (x402) — Demo Playground"
description = """
**HyperLayer demo** — lightweight simulation of zero-latency AI agent responses on a Solana-native infra.
Use the controls to pick a mode, set simulated latency and try sample prompts.  
*This Space shows a demo-only simulation for presentation / hackathon purposes.*
"""

with gr.Blocks(title=title) as demo:
    gr.Markdown(f"# {title}\n\n{description}")

    with gr.Row():
        with gr.Column(scale=2):
            prompt = gr.Textbox(lines=4, label="Prompt / Data stream", placeholder="Type a request for the AI agent...")
            mode = gr.Dropdown(choices=["Explain", "Summarize", "Detect Intent", "Tokenize", "Echo"], value="Explain", label="Mode")
            latency = gr.Slider(minimum=0, maximum=2000, step=50, value=50, label="Simulated latency (ms)")
            run_btn = gr.Button("Run Demo")
            examples = gr.Examples(
                examples=[
                    ["Evaluate arbitrage opportunities between market A and B", "Explain", 50],
                    ["Summarize last 24h orderbook activity for X token", "Summarize", 120],
                    ["transfer 0.5 SOL to 0xabc...z", "Detect Intent", 30],
                    ["tokenize: buy 100 sell 50", "Tokenize", 0],
                ],
                inputs=[prompt, mode, latency],
                label="Try examples"
            )
        with gr.Column(scale=1):
            output = gr.Textbox(lines=12, label="Agent Output (simulated)")
            gr.Markdown("**Notes**\n- This is a demo to showcase how real-time agent responses would look.\n- Replace simulation with real RPC / model calls for production.")

    def run(prompt_text, mode_val, latency_val):
        return simulate_processing(prompt_text, mode_val, int(latency_val))

    run_btn.click(fn=run, inputs=[prompt, mode, latency], outputs=[output])

    # add default examples
    prompt.submit(fn=run, inputs=[prompt, mode, latency], outputs=[output])

if __name__ == "__main__":
    demo.launch(server_name="0.0.0.0", server_port=7860, share=False)
    from transformers import AutoModelForCausalLM, AutoTokenizer
import gradio as gr


model_id = "x402_hyperlayer_model"

tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(model_id)

def chat(prompt):
    inputs = tokenizer(prompt, return_tensors="pt")
    outputs = model.generate(**inputs, max_new_tokens=100)
    return tokenizer.decode(outputs[0], skip_special_tokens=True)

demo = gr.Interface(fn=chat, inputs="text", outputs="text", title="X402 HyperLayer Model")
demo.launch()