feat: Screenshot-Spalte im Dashboard — Full-Page CDP Screenshots

- worker.py: _take_screenshot() via Chrome CDP (JPEG 55%, max 3000px)
- worker.py: alle Scraper geben (results, screenshot_b64) Tuple zurück
- agent.py: screenshot_b64 in API-Response enthalten
- scheduler.py: speichere_screenshot() Funktion + Verknüpfung mit prices
- db.py: screenshots-Tabelle + screenshot_id FK in prices
- web.py: /api/screenshot/<id> Endpoint (base64→JPEG Response)
- web.py: 📷 Button in Preistabelle → Lightbox mit Full-Page Screenshot

Co-authored-by: Cursor <cursoragent@cursor.com>

hub/src/db.py

@@ -83,14 +83,29 @@ def init_db():
             von TEXT,
             nach TEXT,
             booking_url TEXT,
+            screenshot_id INTEGER,
+            scraped_at TEXT DEFAULT (datetime('now'))
+        )
+    """)
+    for col_sql in [
+        "ALTER TABLE prices ADD COLUMN booking_url TEXT",
+        "ALTER TABLE prices ADD COLUMN screenshot_id INTEGER",
+    ]:
+        try:
+            c.execute(col_sql)
+        except Exception:
+            pass
+
+    c.execute("""
+        CREATE TABLE IF NOT EXISTS screenshots (
+            id INTEGER PRIMARY KEY AUTOINCREMENT,
+            job_id INTEGER,
+            node TEXT,
+            scanner TEXT,
+            screenshot_b64 TEXT,
             scraped_at TEXT DEFAULT (datetime('now'))
         )
     """)
-    # Spalte nachrüsten falls DB schon existiert
-    try:
-        c.execute("ALTER TABLE prices ADD COLUMN booking_url TEXT")
-    except Exception:
-        pass
 
     c.execute("""
         CREATE TABLE IF NOT EXISTS analyses (

hub/src/scheduler.py

@@ -76,13 +76,16 @@ def dispatch_job(node, job, tage_override=None):
             timeout=300
         )
         if r.status_code == 200:
-            results = r.json().get("results", [])
+            data           = r.json()
+            results        = data.get("results", [])
+            screenshot_b64 = data.get("screenshot_b64", "")
             via_label = f" via {job.get('via','')}" if job.get('via') else ""
             log(f"{node['name']}: {len(results)} Preise ← {job['scanner']}"
                 f"{' ['+job.get('airline_filter','')+']' if job.get('airline_filter') else ''}"
                 f"{via_label}"
                 f"{' +'+str(tage_override)+'T' if tage_override else ''}")
-            speichere_preise(results, node["name"], job)
+            screenshot_id = speichere_screenshot(screenshot_b64, node["name"], job)
+            speichere_preise(results, node["name"], job, screenshot_id)
             return True
         else:
             log(f"{node['name']}: Fehler {r.status_code} bei {job['scanner']}", "ERROR")
@@ -93,19 +96,39 @@ def dispatch_job(node, job, tage_override=None):
         return False
 
 
-def speichere_preise(results, node_name, job):
+def speichere_screenshot(screenshot_b64, node_name, job):
+    """Speichert Screenshot in DB, gibt screenshot_id zurück (oder None)."""
+    if not screenshot_b64:
+        return None
+    try:
+        conn = get_conn()
+        cur = conn.execute("""
+            INSERT INTO screenshots (job_id, node, scanner, screenshot_b64)
+            VALUES (?, ?, ?, ?)
+        """, (job["id"], node_name, job["scanner"], screenshot_b64))
+        screenshot_id = cur.lastrowid
+        conn.commit()
+        conn.close()
+        return screenshot_id
+    except Exception as e:
+        log(f"Screenshot-Speicher-Fehler: {e}", "WARN")
+        return None
+
+
+def speichere_preise(results, node_name, job, screenshot_id=None):
     conn = get_conn()
     for r in results:
         conn.execute("""
             INSERT INTO prices
-            (job_id, scanner, node, preis, waehrung, airline, abflug, ankunft, von, nach, booking_url)
+            (job_id, scanner, node, preis, waehrung, airline, abflug, ankunft, von, nach, booking_url, screenshot_id)
-            VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?)
+            VALUES (?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?, ?)
         """, (
             job["id"], r.get("scanner", job["scanner"]), node_name,
             r["preis"], r.get("waehrung", "EUR"), r.get("airline", ""),
             r.get("abflug", ""), r.get("ankunft", ""),
             job["von"], job["nach"],
-            r.get("booking_url", "")
+            r.get("booking_url", ""),
+            screenshot_id,
         ))
     conn.commit()
     conn.close()

hub/src/web.py

@@ -1,7 +1,8 @@
 import os
+import base64
 import threading
 from datetime import datetime
-from flask import Flask, jsonify, request, render_template_string
+from flask import Flask, jsonify, request, render_template_string, Response, abort
 from flask_cors import CORS
 from db import init_db, get_conn, log
 from scheduler import scraping_lauf
@@ -143,11 +144,24 @@ OVERVIEW_HTML = BASE_HTML.replace("{% block content %}{% endblock %}", """
 <div class="card">
   <h2>Alle Preise heute (Detail)</h2>
   <table>
-    <thead><tr><th>Anbieter</th><th>Node</th><th>Preis</th><th>Plausibilität</th><th>Abflug</th><th>Rückflug</th><th>Buchen</th></tr></thead>
+    <thead><tr><th>Anbieter</th><th>Node</th><th>Preis</th><th>Plausibilität</th><th>Abflug</th><th>Rückflug</th><th>Buchen</th><th>Screenshot</th></tr></thead>
     <tbody id="preise-tbody"></tbody>
   </table>
 </div>
 
+<!-- Screenshot Lightbox -->
+<div id="ss-modal" onclick="document.getElementById('ss-modal').style.display='none'"
+     style="display:none;position:fixed;top:0;left:0;width:100%;height:100%;background:rgba(0,0,0,0.92);z-index:9999;overflow:auto;cursor:zoom-out">
+  <div style="text-align:center;padding:1rem">
+    <button onclick="document.getElementById('ss-modal').style.display='none'"
+            style="background:#ef4444;color:white;border:none;padding:0.4rem 1rem;border-radius:6px;cursor:pointer;margin-bottom:0.5rem;font-size:0.9rem">
+      ✕ Schließen
+    </button>
+    <div id="ss-info" style="color:#94a3b8;font-size:0.8rem;margin-bottom:0.5rem"></div>
+    <img id="ss-img" src="" style="max-width:100%;border-radius:8px;box-shadow:0 0 40px rgba(0,0,0,0.8)" onclick="event.stopPropagation()">
+  </div>
+</div>
+
 <div class="card">
   <h2>Nodes</h2>
   <table>
@@ -260,6 +274,12 @@ async function ladeUebersicht() {
     const gesamtHtml = isMulticity
       ? `<strong style="color:${warn?'#fbbf24':'#a78bfa'}">${p.preis} €</strong><br><span style="font-size:0.75rem;color:#64748b">∑ ~${Math.round(p.preis)+HOTEL_HKG} € inkl. Hotel</span>`
       : `<strong style="color:${warn?'#fbbf24':'#34d399'}">${p.preis} €</strong>`;
+    const ssBtn = p.screenshot_id
+      ? `<button onclick="zeigeScreenshot(${p.screenshot_id},'${p.scanner} · ${p.node} · ${p.abflug||''}')"
+             style="background:#1e3a5f;border:1px solid #2563eb;color:#93c5fd;padding:0.2rem 0.5rem;border-radius:5px;cursor:pointer;font-size:0.8rem">
+           📷
+         </button>`
+      : '<span style="color:#334155;font-size:0.75rem">—</span>';
     return `<tr${isMulticity?' style="background:rgba(99,102,241,0.06);border-left:3px solid #6366f1"':''}>
       <td>${scannerLabel}</td>
       <td style="font-size:0.8rem;color:#64748b">${p.node}</td>
@@ -268,6 +288,7 @@ async function ladeUebersicht() {
       <td style="font-size:0.85rem">${p.abflug||'—'}</td>
       <td style="font-size:0.85rem">${p.ankunft||'—'}</td>
       <td>${buchBtn}</td>
+      <td>${ssBtn}</td>
     </tr>`;
   }).join('') || '<tr><td colspan="7" style="color:#475569;text-align:center">Noch keine Daten heute</td></tr>';
 
@@ -312,6 +333,18 @@ async function pruefeScanStatus() {
   }
 }
 
+function zeigeScreenshot(id, label) {
+  const modal = document.getElementById('ss-modal');
+  const img   = document.getElementById('ss-img');
+  const info  = document.getElementById('ss-info');
+  img.src = '';
+  info.textContent = 'Lade Screenshot...';
+  modal.style.display = 'block';
+  img.onload = () => { info.textContent = label || ''; };
+  img.onerror = () => { info.textContent = '⚠ Screenshot nicht verfügbar'; };
+  img.src = `/api/screenshot/${id}`;
+}
+
 async function manuellScrapen() {
   const r = await fetch('/api/scrape/now', {method:'POST'});
   const d = await r.json();
@@ -537,6 +570,20 @@ def api_scrape_status():
     return jsonify({"running": _lauf_aktiv})
 
 
+@app.route("/api/screenshot/<int:screenshot_id>")
+def api_screenshot(screenshot_id):
+    conn = get_conn()
+    row = conn.execute(
+        "SELECT screenshot_b64 FROM screenshots WHERE id=?", (screenshot_id,)
+    ).fetchone()
+    conn.close()
+    if not row or not row["screenshot_b64"]:
+        abort(404)
+    img_data = base64.b64decode(row["screenshot_b64"])
+    return Response(img_data, mimetype="image/jpeg",
+                    headers={"Cache-Control": "public, max-age=3600"})
+
+
 # ─── Seiten ────────────────────────────────────────────────────────────────────
 
 @app.route("/")

node/src/agent.py

@@ -37,14 +37,16 @@ def job():
           f"Umstieg {layover_min}-{layover_max}min | max {max_flugzeit_h}h/{max_stops} Stopps")
 
     try:
-        results = scrape(scanner, von, nach, tage, aufenthalt, trip_type, kabine,
+        results, screenshot_b64 = scrape(
-                         gepaeck, airline_filter, layover_min, layover_max,
+            scanner, von, nach, tage, aufenthalt, trip_type, kabine,
-                         max_flugzeit_h, max_stops, via, stopover_min_h, stopover_max_h)
+            gepaeck, airline_filter, layover_min, layover_max,
-        print(f"[{NODE_NAME}] {len(results)} Preise gefunden")
+            max_flugzeit_h, max_stops, via, stopover_min_h, stopover_max_h)
+        print(f"[{NODE_NAME}] {len(results)} Preise | Screenshot: {len(screenshot_b64)//1024} KB")
         return jsonify({
-            "results": results,
+            "results":       results,
-            "node": NODE_NAME,
+            "node":          NODE_NAME,
-            "count": len(results)
+            "count":         len(results),
+            "screenshot_b64": screenshot_b64,
         })
     except Exception as e:
         print(f"[{NODE_NAME}] Fehler: {e}")

node/src/worker.py

@@ -9,6 +9,11 @@ def scrape(scanner, von, nach, tage=30, aufenthalt_tage=60,
            layover_min=120, layover_max=300,
            max_flugzeit_h=22, max_stops=2,
            via="", stopover_min_h=20, stopover_max_h=30):
+    """
+    Gibt (results, screenshot_b64) zurück.
+    results = Liste von Preis-Dicts
+    screenshot_b64 = JPEG Full-Page Screenshot als base64-String (leer wenn Fehler)
+    """
     dispatcher = {
         "google_flights":   scrape_google_flights,
         "kayak":            scrape_kayak,
@@ -26,6 +31,27 @@ def scrape(scanner, von, nach, tage=30, aufenthalt_tage=60,
               airline_filter, layover_min, layover_max, max_flugzeit_h, max_stops)
 
 
+def _take_screenshot(sb):
+    """Full-Page Screenshot via CDP (JPEG 55%, max 3000px). Gibt base64-String zurück."""
+    try:
+        result = sb.driver.execute_cdp_cmd("Page.captureScreenshot", {
+            "format": "jpeg",
+            "quality": 55,
+            "captureBeyondViewport": True,
+            "clip": {"x": 0, "y": 0, "width": 1280, "height": 3000, "scale": 0.75},
+        })
+        data = result.get("data", "")
+        if data:
+            print(f"[Screenshot] OK — {len(data)//1024} KB base64")
+        return data
+    except Exception as e:
+        print(f"[Screenshot] CDP-Fehler: {e}")
+        try:
+            return sb.driver.get_screenshot_as_base64()
+        except Exception:
+            return ""
+
+
 def _booking_url_google(von, nach, abflug, rueck, kc):
     # Hash-Fragment wird von headless Chrome ignoriert → tfs-Parameter nutzen
     if rueck:
@@ -254,7 +280,8 @@ def scrape_google_flights(von, nach, tage=30, aufenthalt_tage=60,
 
         results = [r for r in results if r["preis"] > 400]
         print(f"[GF] Ergebnis: {[r['preis'] for r in results[:5]]}")
-    return results[:10]
+        screenshot_b64 = _take_screenshot(sb)
+    return results[:10], screenshot_b64
 
 
 def scrape_kayak(von, nach, tage=30, aufenthalt_tage=60,
@@ -324,7 +351,8 @@ def scrape_kayak(von, nach, tage=30, aufenthalt_tage=60,
 
         airline_label = f" [{airline_filter}]" if airline_filter else ""
         print(f"[KY{airline_label}] Ergebnis: {[r['preis'] for r in results[:5]]}")
-    return results[:10]
+        screenshot_b64 = _take_screenshot(sb)
+    return results[:10], screenshot_b64
 
 
 def scrape_trip(von, nach, tage=30, aufenthalt_tage=60,
@@ -388,7 +416,8 @@ def scrape_trip(von, nach, tage=30, aufenthalt_tage=60,
                 results.append(r)
 
         print(f"[TR] Ergebnis: {[r['preis'] for r in results[:5]]}")
-    return results[:10]
+        screenshot_b64 = _take_screenshot(sb)
+    return results[:10], screenshot_b64
 
 
 def _booking_url_kayak_multicity(von, nach, via, abflug, via_datum, rueck, kc, bags=1, airline=""):
@@ -488,7 +517,8 @@ def scrape_kayak_multicity(von, nach, tage=30, aufenthalt_tage=60,
                     results.append(r)
 
         print(f"[MC{airline_label}] Ergebnis: {[r['preis'] for r in results[:5]]}")
-    return results[:10]
+        screenshot_b64 = _take_screenshot(sb)
+    return results[:10], screenshot_b64
 
 
 def scrape_skyscanner(von, nach, tage=30, aufenthalt_tage=60,
@@ -498,4 +528,4 @@ def scrape_skyscanner(von, nach, tage=30, aufenthalt_tage=60,
                       max_flugzeit_h=22, max_stops=2):
     """Skyscanner hat starken Bot-Schutz — übersprungen."""
     print("[SS] Skyscanner übersprungen (Bot-Detection)")
-    return []
+    return [], ""