app.py

# app.py
import gradio as gr
from genesis.pipeline import research_once, DEMO_QUERIES

def run_pipeline_ui(query):
    if not query.strip():
        return "Please enter a research topic.", "", None, None
    report = research_once(query)
    summary = report["summary"]
    citations_md = "\n".join(f"- [{c['type']}]({c['url']})" for c in report["citations"]) or "No citations found."
    return summary, citations_md, report["visual_image_url"], report["audio_url"]

with gr.Blocks(theme=gr.themes.Soft(primary_hue="green", secondary_hue="blue"), css="""
    .demo-btn {border-radius: 25px; padding: 8px 14px; font-weight: bold;}
    .demo-btn:hover {background-color: #28a745 !important; color: white !important;}
""") as demo:
    
    gr.Markdown(
        """
        # 🧬 GENESIS-AI — Synthetic Biology Research Engine
        Ask about **CRISPR**, drug design, biosensors, metabolic engineering, and more.  
        _Your AI co-pilot for disruptive biotech discovery._
        """,
        elem_id="title"
    )
    
    with gr.Row():
        query_box = gr.Textbox(label="Enter Your Research Topic", placeholder="e.g., Design a CRISPR-based living therapeutic for triple-negative breast cancer", scale=4)
        run_btn = gr.Button("🚀 Run Research", variant="primary", scale=1)
    
    with gr.Row():
        gr.Markdown("**Or try one of these expert-curated demo queries:**")
    
    with gr.Row():
        for q in DEMO_QUERIES:
            gr.Button(q, elem_classes="demo-btn").click(
                fn=run_pipeline_ui,
                inputs=gr.Textbox(value=q, visible=False),
                outputs=["summary_out", "cites_out", "img_out", "audio_out"]
            )

    with gr.Tab("📄 Summary"):
        summary_out = gr.Textbox(label="Executive Summary", lines=15, interactive=False)
    with gr.Tab("🔗 Citations"):
        cites_out = gr.Markdown()
    with gr.Tab("🖼 Diagram"):
        img_out = gr.Image(type="filepath", label="Generated Research Diagram")
    with gr.Tab("🔊 Narration"):
        audio_out = gr.Audio(type="filepath", label="AI Narration")

    # Main run button event
    run_btn.click(
        fn=run_pipeline_ui,
        inputs=query_box,
        outputs=[summary_out, cites_out, img_out, audio_out]
    )

demo.launch()