app.py

import gradio as gr
import asyncio

from ingest.registry import get_all_adapters
from ingest.cluster import semantic_clusters
from ingest.export import export_results_zip
from ingest.health import check_health

ADAPTERS = get_all_adapters()

async def federated_search(query: str):
    results = []
    for adapter in ADAPTERS.values():
        try:
            res = await adapter.search(query)
            results.extend(res)
        except Exception:
            continue
    return results

def run_search(query):
    if not query.strip():
        return [], []

    results = asyncio.run(federated_search(query))

    rows = [
        [r["source"], r["title"], r["url"], r["snippet"]]
        for r in results
    ]

    texts = [r["snippet"] for r in results]
    clusters = semantic_clusters(texts) if texts else []

    return rows, clusters

def export_zip_handler(table):
    results = []
    for row in table:
        results.append({
            "source": row[0],
            "title": row[1],
            "url": row[2],
            "snippet": row[3],
        })
    return export_results_zip(results)

with gr.Blocks() as app:
    gr.Markdown(
        "# **Federal FOIA Intelligence Search**\n"
        "### Public Electronic Reading Rooms Only\n\n"
        "Live, robots-compliant search across U.S. government FOIA libraries."
    )

    query = gr.Textbox(label="Search term", placeholder="e.g. UAP, procurement, surveillance")
    search_btn = gr.Button("Search")

    results_table = gr.Dataframe(
        headers=["Agency", "Title", "URL", "Snippet"],
        interactive=False,
        wrap=True
    )

    cluster_output = gr.State()

    search_btn.click(
        fn=run_search,
        inputs=query,
        outputs=[results_table, cluster_output]
    )

    gr.Markdown("### Export")
    export_btn = gr.Button("Export Results (ZIP)")
    zip_file = gr.File(label="Download ZIP")

    export_btn.click(
        fn=export_zip_handler,
        inputs=results_table,
        outputs=zip_file
    )

    gr.Markdown("### Source Health")
    for adapter in ADAPTERS.values():
        health = check_health(adapter)
        gr.Markdown(
            f"- **{health['source']}**: {health['status']} "
            f"({health['latency_ms']} ms)"
        )

app.launch()