Add robust company slug for job-offer blacklist matching

Company matching relied on normText(), which kept legal-form suffixes,
spacing and umlaut spelling — so "Bosch GmbH", "Bosch AG" and "bosch gmbh"
did not match the same entry, making the blacklist unreliable.

Introduce firmaSlug(): transliterate German umlauts, strip diacritics and
trailing legal-form tokens (GmbH/AG/SE/KG/…), then kebab-join. Add a
firma_slug field to jobangebote_blacklist (schema + ALTER/backfill migration)
and match on it for typ firma/firma_stelle/auto, falling back to firma_norm
for legacy rows. POST /joboffers rejects blacklisted offers via matchBlacklist
automatically since offerSignature now carries the slug.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

lib/blacklist.js

@@ -8,9 +8,10 @@
 // relevant to its typ are set:
 //   'url'          -> url_norm            (exact normalized URL)
 //   'domain'       -> domain              (whole host, e.g. a job board)
-//   'firma'        -> firma_norm          (block a company entirely)
-//   'firma_stelle' -> firma_norm + stelle_norm (+ optional ort_norm)
-//   'auto'         -> url_norm / quelle+external_id / firma_norm+stelle_norm
+//   'firma'        -> firma_slug          (block a company entirely; slug is a
+//                     legal-form/umlaut-robust key, falls back to firma_norm)
+//   'firma_stelle' -> firma_slug + stelle_norm (+ optional ort_norm)
+//   'auto'         -> url_norm / quelle+external_id / firma_slug+stelle_norm
 //                     (created on delete; matches on ANY stored signature)
 
 const TYPES = ['url', 'domain', 'firma', 'firma_stelle', 'auto'];
@@ -72,6 +73,40 @@ function normalizeUrl(raw) {
   return host + pathname + qs;
 }
 
+// Legal-form / entity suffixes stripped from a company name before slugging,
+// so "Bosch GmbH", "Bosch AG", "Bosch GmbH & Co. KG" and "Bosch" all collapse
+// to the same slug. Matched as whole space-separated tokens at the end.
+const LEGAL_FORM_TOKENS = new Set([
+  'gmbh', 'ggmbh', 'mbh', 'ug', 'haftungsbeschraenkt', 'ag', 'kg', 'kgaa',
+  'ohg', 'gbr', 'se', 'ek', 'eg', 'ev', 'partg', 'partmbb', 'co', 'cie',
+  'inc', 'incorporated', 'llc', 'ltd', 'limited', 'plc', 'corp', 'corporation',
+  'company', 'sa', 'sarl', 'sas', 'bv', 'nv', 'oy', 'ab', 'as', 'aps', 'srl', 'spa',
+]);
+
+// A robust matching key for a company name. Beyond normText it transliterates
+// German umlauts (ä→ae …), strips diacritics, drops trailing legal-form tokens
+// and joins the rest with hyphens. This is what the blacklist compares on so
+// that entity form, spacing and umlaut spelling no longer cause misses.
+function firmaSlug(s) {
+  let t = unescapeInput(s)
+    .toLowerCase()
+    .replace(/ä/g, 'ae').replace(/ö/g, 'oe').replace(/ü/g, 'ue').replace(/ß/g, 'ss')
+    .normalize("NFKD").replace(/[\u0300-\u036f]/g, "")
+    .replace(/\((?:[mwdfax](?:\s*\/\s*[mwdfax])*)\)/gi, ' ')
+    .replace(/[^a-z0-9]+/g, ' ')
+    .replace(/\s+/g, ' ')
+    .trim();
+  if (!t) return '';
+  const words = t.split(' ');
+  const kept = words.slice();
+  // Drop legal-form tokens from the end (e.g. "gmbh & co kg" → "gmbh co kg"),
+  // but never strip everything away.
+  while (kept.length > 1 && LEGAL_FORM_TOKENS.has(kept[kept.length - 1])) {
+    kept.pop();
+  }
+  return (kept.length ? kept : words).join('-');
+}
+
 // Normalise a company / job-title string for fuzzy identity matching:
 // unescape, lowercase, drop gender markers like "(m/w/d)", collapse anything
 // non-alphanumeric to single spaces.
@@ -95,11 +130,19 @@ function offerSignature(offer) {
     external_id: o.external_id != null && String(o.external_id).trim() !== ''
       ? String(o.external_id).trim() : null,
     firma_norm: normText(o.firma),
+    firma_slug: firmaSlug(o.firma),
     stelle_norm: normText(o.stelle),
     ort_norm: normText(o.ort),
   };
 }
 
+// Company identity match: prefer the robust slug, fall back to firma_norm for
+// legacy rows written before firma_slug existed (or not yet backfilled).
+function firmaMatch(row, sig) {
+  if (row.firma_slug && sig.firma_slug) return row.firma_slug === sig.firma_slug;
+  return !!row.firma_norm && row.firma_norm === sig.firma_norm;
+}
+
 // Does one blacklist row block the given signature?
 function rowMatches(row, sig) {
   switch (row.typ) {
@@ -108,9 +151,9 @@ function rowMatches(row, sig) {
     case 'domain':
       return !!row.domain && row.domain === sig.domain;
     case 'firma':
-      return !!row.firma_norm && row.firma_norm === sig.firma_norm;
+      return firmaMatch(row, sig);
     case 'firma_stelle':
-      return !!row.firma_norm && row.firma_norm === sig.firma_norm
+      return firmaMatch(row, sig)
         && !!row.stelle_norm && row.stelle_norm === sig.stelle_norm
         && (!row.ort_norm || row.ort_norm === sig.ort_norm);
     case 'auto':
@@ -118,8 +161,8 @@ function rowMatches(row, sig) {
       if (row.url_norm && sig.url_norm && row.url_norm === sig.url_norm) return true;
       if (row.quelle && row.external_id && sig.external_id
         && row.quelle === sig.quelle && row.external_id === sig.external_id) return true;
-      if (row.firma_norm && row.stelle_norm
-        && row.firma_norm === sig.firma_norm && row.stelle_norm === sig.stelle_norm
+      if ((row.firma_slug || row.firma_norm) && row.stelle_norm
+        && firmaMatch(row, sig) && row.stelle_norm === sig.stelle_norm
         && (!row.ort_norm || row.ort_norm === sig.ort_norm)) return true;
       return false;
   }
@@ -137,7 +180,7 @@ function matchBlacklist(rows, offer) {
 // Column set for INSERT — keeps server.js and api.js in lock-step.
 const COLUMNS = [
   'typ', 'url_norm', 'domain', 'quelle', 'external_id',
-  'firma_norm', 'stelle_norm', 'ort_norm', 'firma', 'stelle', 'quelle_url', 'grund',
+  'firma_norm', 'firma_slug', 'stelle_norm', 'ort_norm', 'firma', 'stelle', 'quelle_url', 'grund',
 ];
 
 function emptyEntry() {
@@ -170,6 +213,7 @@ function buildManualEntry(input) {
     const w = wert || inp.firma || '';
     e.firma = w || null;
     e.firma_norm = normText(w);
+    e.firma_slug = firmaSlug(w) || null;
     if (!e.firma_norm) return null;
   } else if (typ === 'firma_stelle') {
     const f = inp.firma || '';
@@ -177,6 +221,7 @@ function buildManualEntry(input) {
     e.firma = f || null;
     e.stelle = s || null;
     e.firma_norm = normText(f);
+    e.firma_slug = firmaSlug(f) || null;
     e.stelle_norm = normText(s);
     e.ort_norm = inp.ort ? normText(inp.ort) : null;
     if (!e.firma_norm || !e.stelle_norm) return null;
@@ -196,6 +241,7 @@ function buildAutoEntry(offer, grund) {
   e.quelle = sig.external_id ? sig.quelle : null;
   e.external_id = sig.external_id || null;
   e.firma_norm = sig.firma_norm || null;
+  e.firma_slug = sig.firma_slug || null;
   e.stelle_norm = sig.stelle_norm || null;
   e.ort_norm = sig.ort_norm || null;
   e.firma = offer && offer.firma ? offer.firma : null;
@@ -210,6 +256,7 @@ module.exports = {
   normalizeUrl,
   urlDomain,
   normText,
+  firmaSlug,
   offerSignature,
   rowMatches,
   matchBlacklist,

lib/openapi.js

@@ -938,6 +938,7 @@ function buildOpenApiSpec(baseUrl = '') {
             quelle: { type: 'string', nullable: true },
             external_id: { type: 'string', nullable: true },
             firma_norm: { type: 'string', nullable: true },
+            firma_slug: { type: 'string', nullable: true, description: 'Rechtsform-/Umlaut-robuster Firmen-Slug; primärer Abgleich für typ=firma/firma_stelle/auto.' },
             stelle_norm: { type: 'string', nullable: true },
             ort_norm: { type: 'string', nullable: true },
             firma: { type: 'string', nullable: true },