SEO blogpost

server/index.js

@@ -37,6 +37,7 @@ const {
   issueToken,
   verifyJwt,
 } = require('./lib/auth');
+const { ensureScanEventSchema, recordScanEvent } = require('./lib/scanEvents');
 const {
   PlantImportValidationError,
   ensurePlantSchema,
@@ -959,6 +960,12 @@ app.post('/v1/scan', guestScanLimiter, async (request, response) => {
     };
 
     await storeEndpointResponse(db, endpointId, payload);
+    await recordScanEvent(db, {
+      scanType: 'identify',
+      species: result?.botanicalName || result?.name || null,
+      confidence: result?.confidence,
+      locale: language,
+    });
     response.status(200).json(payload);
   } catch (error) {
     console.error(`Scan error for user ${userId}:`, error);
@@ -1139,6 +1146,13 @@ app.post('/v1/health-check', async (request, response) => {
     };
 
     await storeEndpointResponse(db, endpointId, payload);
+    await recordScanEvent(db, {
+      scanType: 'health',
+      diagnosis: analysis?.likelyIssues?.[0]?.title || null,
+      confidence: analysis?.likelyIssues?.[0]?.confidence,
+      healthStatus: analysis?.status || null,
+      locale: language,
+    });
     response.status(200).json(payload);
   } catch (error) {
     const payload = toApiErrorPayload(error);
@@ -1946,6 +1960,7 @@ const start = async () => {
   await ensurePlantSchema(db);
   await ensureBillingSchema(db);
   await ensureAuthSchema(db);
+  await ensureScanEventSchema(db);
   await seedBootstrapCatalogIfNeeded();
   if (isStorageConfigured()) {
     await ensureStorageBucketWithRetry().catch((err) => console.warn('MinIO bucket setup failed:', err.message));

server/lib/scanEvents.js

@@ -0,0 +1,105 @@
+const { run, all } = require('./postgres');
+
+/**
+ * Anonymes Scan-Logging für aggregierte Auswertungen (Content, Katalogpflege).
+ *
+ * Bewusst OHNE user_id, Bildreferenz oder Idempotency-Key: Die Tabelle ist von
+ * Anfang an nicht personenbeziehbar, damit später nichts nachträglich
+ * anonymisiert werden muss. Wer wann gescannt hat, steht weiterhin in
+ * billing_accounts bzw. PostHog — hier geht es nur um das Was.
+ *
+ * Schreibfehler dürfen einen Scan nie scheitern lassen, siehe recordScanEvent.
+ */
+const ensureScanEventSchema = async (db) => {
+  await run(
+    db,
+    `CREATE TABLE IF NOT EXISTS scan_events (
+      id BIGSERIAL PRIMARY KEY,
+      scan_type TEXT NOT NULL,
+      species TEXT,
+      confidence DOUBLE PRECISION,
+      diagnosis TEXT,
+      health_status TEXT,
+      locale TEXT,
+      created_at TIMESTAMPTZ NOT NULL DEFAULT NOW()
+    )`,
+  );
+
+  await run(db, 'CREATE INDEX IF NOT EXISTS idx_scan_events_created_at ON scan_events (created_at DESC)');
+  await run(db, 'CREATE INDEX IF NOT EXISTS idx_scan_events_species ON scan_events (LOWER(species))');
+  await run(db, 'CREATE INDEX IF NOT EXISTS idx_scan_events_scan_type ON scan_events (scan_type)');
+};
+
+const trimOrNull = (value) => {
+  if (typeof value !== 'string') return null;
+  const trimmed = value.trim();
+  return trimmed ? trimmed.slice(0, 200) : null;
+};
+
+const toNumberOrNull = (value) => {
+  const numeric = Number(value);
+  return Number.isFinite(numeric) ? numeric : null;
+};
+
+/**
+ * Best effort. Ein fehlgeschlagenes Log darf den Scan des Nutzers nicht
+ * beeinflussen — deshalb wird hier geloggt und geschluckt, nicht geworfen.
+ */
+const recordScanEvent = async (db, event) => {
+  try {
+    await run(
+      db,
+      `INSERT INTO scan_events (scan_type, species, confidence, diagnosis, health_status, locale)
+       VALUES ($1, $2, $3, $4, $5, $6)`,
+      [
+        trimOrNull(event?.scanType) || 'unknown',
+        trimOrNull(event?.species),
+        toNumberOrNull(event?.confidence),
+        trimOrNull(event?.diagnosis),
+        trimOrNull(event?.healthStatus),
+        trimOrNull(event?.locale),
+      ],
+    );
+  } catch (error) {
+    console.warn('scan_events write failed (ignored):', error.message);
+  }
+};
+
+/**
+ * Aggregat für spätere Content-Auswertungen. Erst ab nennenswerter Menge
+ * aussagekräftig — die Mindestmenge bewusst im Aufrufer prüfen, nicht hier.
+ */
+const getScanEventStats = async (db, { since = null, limit = 100 } = {}) => {
+  const params = [];
+  let whereClause = '';
+  if (since) {
+    params.push(since);
+    whereClause = 'WHERE created_at >= $1';
+  }
+  params.push(limit);
+
+  const topSpecies = await all(
+    db,
+    `SELECT species, COUNT(*)::int AS scans, ROUND(AVG(confidence)::numeric, 3) AS avg_confidence
+     FROM scan_events
+     ${whereClause}${whereClause ? ' AND' : 'WHERE'} species IS NOT NULL
+     GROUP BY species
+     ORDER BY scans DESC
+     LIMIT $${params.length}`,
+    params,
+  );
+
+  const totals = await all(
+    db,
+    `SELECT scan_type, COUNT(*)::int AS scans FROM scan_events ${whereClause} GROUP BY scan_type`,
+    since ? [since] : [],
+  );
+
+  return { topSpecies, totals };
+};
+
+module.exports = {
+  ensureScanEventSchema,
+  getScanEventStats,
+  recordScanEvent,
+};