Add Dockerfile, Nomad spec, Gitea CI/CD, requirements and health endpoint

app.py

@@ -0,0 +1,278 @@
+#!/usr/bin/env python3
+"""
+BlaaAi — AI-powered DBA listing analyzer
+FastAPI backend
+
+Usage:
+  uvicorn app:app --reload --port 8000
+"""
+
+import json, os, smtplib
+from concurrent.futures import ThreadPoolExecutor
+from datetime import datetime, timezone
+from email.mime.multipart import MIMEMultipart
+from email.mime.text import MIMEText
+from pathlib import Path
+from typing import Optional
+
+from fastapi import BackgroundTasks, FastAPI, HTTPException
+from fastapi.responses import HTMLResponse
+from fastapi.staticfiles import StaticFiles
+from fastapi.templating import Jinja2Templates
+from pydantic import BaseModel
+from starlette.requests import Request
+
+from fetch_dba import (
+    DATA_DIR, ITEM_CACHE, create_search, detect_domain, enrich_listings,
+    fetch_page, find_new, listings_file, load_meta, load_seen,
+    save_seen, seen_file, list_searches,
+)
+from score import CRITERIA, METRICS_FILE, score_listings
+
+app       = FastAPI(title="BlaaAi")
+app.mount("/static", StaticFiles(directory="static"), name="static")
+templates = Jinja2Templates(directory="templates")
+executor  = ThreadPoolExecutor(max_workers=4)
+
+
+# ── Meta helpers ──────────────────────────────────────────────────────────────
+
+def update_meta(search_id: str, **kwargs) -> dict:
+    p    = DATA_DIR / search_id / "meta.json"
+    meta = json.loads(p.read_text())
+    meta.update(kwargs)
+    p.write_text(json.dumps(meta, ensure_ascii=False, indent=2))
+    return meta
+
+
+# ── Background pipeline ───────────────────────────────────────────────────────
+
+def run_fetch_and_score(search_id: str, prefs: str = "") -> None:
+    """Synchronous fetch + score pipeline — runs in thread pool."""
+    try:
+        meta       = load_meta(search_id)
+        search_url = meta["url"]
+
+        # ── Fetch ──────────────────────────────────────────────────────────────
+        update_meta(search_id, status="fetching")
+        sf   = seen_file(search_id)
+        lf   = listings_file(search_id)
+        seen = load_seen(sf)
+
+        listings, _ = fetch_page(search_url, 1)
+        new_items   = find_new(listings, seen)
+
+        if new_items:
+            new_items = enrich_listings(new_items)
+            existing  = json.loads(lf.read_text()) if lf.exists() else []
+            existing.extend(new_items)
+            lf.write_text(json.dumps(existing, ensure_ascii=False, indent=2))
+            save_seen(sf, seen | {l["id"] for l in listings})
+
+        # ── Score ──────────────────────────────────────────────────────────────
+        update_meta(search_id, status="scoring")
+        items    = json.loads(lf.read_text()) if lf.exists() else []
+        domain   = detect_domain(search_url)
+        criteria = CRITERIA[domain]
+
+        score_listings(items, criteria, prefs, force=False, source_file=lf)
+
+        update_meta(
+            search_id,
+            status="ready",
+            listing_count=len(items),
+            last_scored_at=datetime.now().isoformat(timespec="seconds"),
+            prefs=prefs,
+        )
+
+    except Exception as e:
+        update_meta(search_id, status="error", error=str(e))
+        raise
+
+
+# ── Pydantic models ───────────────────────────────────────────────────────────
+
+class NewSearchRequest(BaseModel):
+    url:   str
+    prefs: Optional[str] = ""
+
+
+class EmailRequest(BaseModel):
+    email: str
+
+
+# ── Routes ────────────────────────────────────────────────────────────────────
+
+@app.get("/health")
+async def health():
+    return {
+        "status": "healthy",
+        "timestamp": datetime.now(timezone.utc).isoformat(),
+        "version": os.getenv("BUILD_VERSION", "unknown"),
+        "commit": os.getenv("GIT_COMMIT", "unknown")[:7],
+    }
+
+
+@app.get("/", response_class=HTMLResponse)
+async def index(request: Request):
+    return templates.TemplateResponse(request, "index.html")
+
+
+@app.get("/search/{search_id}", response_class=HTMLResponse)
+async def search_view(request: Request, search_id: str):
+    return templates.TemplateResponse(request, "index.html", {"search_id": search_id})
+
+
+@app.post("/api/searches")
+async def create(body: NewSearchRequest, background_tasks: BackgroundTasks):
+    search_id = create_search(body.url)
+    update_meta(search_id, status="queued", prefs=body.prefs or "")
+    background_tasks.add_task(run_fetch_and_score, search_id, body.prefs or "")
+    return {"id": search_id, "status": "queued"}
+
+
+@app.get("/api/searches")
+async def get_all():
+    return list_searches()
+
+
+@app.get("/api/searches/{search_id}")
+async def get_search(search_id: str):
+    try:
+        meta = load_meta(search_id)
+    except FileNotFoundError:
+        raise HTTPException(404, "Søgning ikke fundet")
+
+    lf = listings_file(search_id)
+    if lf.exists():
+        items  = json.loads(lf.read_text())
+        ranked = sorted(
+            [i for i in items if i.get("ai_score") is not None],
+            key=lambda x: x["ai_score"],
+            reverse=True,
+        )
+        meta["listings"]      = ranked
+        meta["listing_count"] = len(items)
+        meta["scored_count"]  = len(ranked)
+        if "status" not in meta and ranked:
+            meta["status"] = "ready"
+
+    if "status" not in meta:
+        meta["status"] = "ready" if meta.get("listing_count", 0) > 0 else "unknown"
+
+    return meta
+
+
+@app.post("/api/searches/{search_id}/rescore")
+async def rescore(search_id: str, body: NewSearchRequest, background_tasks: BackgroundTasks):
+    try:
+        load_meta(search_id)
+    except FileNotFoundError:
+        raise HTTPException(404, "Søgning ikke fundet")
+
+    update_meta(search_id, status="queued", prefs=body.prefs or "")
+    background_tasks.add_task(run_fetch_and_score, search_id, body.prefs or "")
+    return {"id": search_id, "status": "queued"}
+
+
+@app.get("/metrics")
+async def get_metrics():
+    global_metrics = {}
+    if METRICS_FILE.exists():
+        global_metrics = json.loads(METRICS_FILE.read_text())
+
+    # Per-search breakdown
+    searches = []
+    if DATA_DIR.exists():
+        for d in sorted(DATA_DIR.iterdir(), key=lambda p: p.stat().st_mtime, reverse=True):
+            mf = d / "metrics.json"
+            if mf.exists():
+                searches.append(json.loads(mf.read_text()))
+
+    # Item cache stats
+    cached_items = len(list(ITEM_CACHE.glob("*.json"))) if ITEM_CACHE.exists() else 0
+
+    return {**global_metrics, "item_cache_size": cached_items, "searches": searches}
+
+
+@app.post("/api/searches/{search_id}/email")
+async def send_email(search_id: str, body: EmailRequest):
+    try:
+        meta = load_meta(search_id)
+    except FileNotFoundError:
+        raise HTTPException(404, "Søgning ikke fundet")
+
+    if meta.get("status") != "ready":
+        raise HTTPException(400, "Analysen er ikke færdig endnu")
+
+    lf = listings_file(search_id)
+    if not lf.exists():
+        raise HTTPException(400, "Ingen resultater at sende")
+
+    items  = json.loads(lf.read_text())
+    ranked = sorted(
+        [i for i in items if i.get("ai_score") is not None],
+        key=lambda x: x["ai_score"],
+        reverse=True,
+    )[:10]
+
+    html = _build_email_html(ranked, meta)
+    _send_email(body.email, "🔍 Dine DBA-resultater fra BlaaAi", html)
+    return {"status": "sent", "to": body.email}
+
+
+# ── Email helpers ─────────────────────────────────────────────────────────────
+
+def _build_email_html(ranked: list[dict], meta: dict) -> str:
+    rows = ""
+    for i, item in enumerate(ranked, 1):
+        score  = item.get("ai_score", 0)
+        bar    = "█" * int(score) + "░" * (10 - int(score))
+        reason = item.get("ai_reason", "")
+        warn   = f'<p style="color:#dc2626;font-size:13px">⚠️ {item["ai_warnings"]}</p>' if item.get("ai_warnings") else ""
+        rows += f"""
+        <tr>
+          <td style="padding:16px;border-bottom:1px solid #e5e7eb">
+            <strong>#{i} [{score}] {item['name']}</strong><br>
+            <span style="color:#6b7280">{item.get('price_dkk','?')} DKK</span>
+            <span style="font-family:monospace;color:#6366f1;margin-left:8px">{bar}</span><br>
+            <p style="margin:6px 0;font-size:14px">{reason}</p>
+            {warn}
+            <a href="{item['url']}" style="color:#6366f1;font-size:13px">Se annonce →</a>
+          </td>
+        </tr>"""
+
+    return f"""
+    <html><body style="font-family:sans-serif;max-width:600px;margin:0 auto;padding:20px">
+      <h1 style="color:#1f2937">🔍 Dine DBA-resultater</h1>
+      <p style="color:#6b7280">Søgning: <a href="{meta['url']}">{meta['url'][:60]}…</a></p>
+      <table style="width:100%;border-collapse:collapse">{rows}</table>
+      <p style="color:#9ca3af;font-size:12px;margin-top:24px">Leveret af BlaaAi</p>
+    </body></html>"""
+
+
+def _send_email(to: str, subject: str, html: str) -> None:
+    host = os.environ.get("SMTP_HOST", "")
+    user = os.environ.get("SMTP_USER", "")
+    pwd  = os.environ.get("SMTP_PASS", "")
+    port = int(os.environ.get("SMTP_PORT", "587"))
+
+    if not host:
+        raise HTTPException(503, "Email ikke konfigureret (SMTP_HOST mangler)")
+
+    msg                    = MIMEMultipart("alternative")
+    msg["Subject"]         = subject
+    msg["From"]            = user
+    msg["To"]              = to
+    msg.attach(MIMEText(html, "html"))
+
+    with smtplib.SMTP(host, port) as s:
+        s.starttls()
+        s.login(user, pwd)
+        s.sendmail(user, to, msg.as_string())
+
+
+if __name__ == "__main__":
+    import uvicorn
+    port = int(os.getenv("PORT", "8000"))
+    uvicorn.run("app:app", host="0.0.0.0", port=port, reload=True)