feat(14-02): add RssAdapter.parseFeed + 'rss' normalizer dispatch

Fixture-first RSS parsing (INGEST-04): parses live-captured service.bund.de (pubDate present, numeric-HTML-entity titles) and subreport-elvis (pubDate absent, CDATA titles) feed shapes into RawTenderRecord[] via fast-xml-parser, mirroring the DoeOpenDataAdapter config. SourceType extended with 'rss'; normalize() dispatches 'rss' through the existing normalizeBag() path unchanged (D-04/D-05). Rule 1 fix: fast-xml-parser only decodes the 5 predefined XML entities, not numeric character references — added an explicit decode step so service.bund.de titles ("Übermittlung...") render correctly instead of leaking raw entity syntax. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-07-23 13:17:02 +02:00
parent a47c0c57ee
commit 3a96cbbbe6
7 changed files with 597 additions and 6 deletions
@@ -0,0 +1,105 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<rss version="2.0">
+  <channel>
+    <title>service.bund.de - öffentliche Ausschreibungen</title>
+    <link>http://www.service.bund.de/ausschreibungen</link>
+    <description>RSS-Feed mit aktuellen Ausschreibungen der Vergabestellen der Bundes-, Landes- und Kommunalverwaltung.</description>
+    <language>de-de</language>
+    <pubDate>Thu, 23 Jul 2026 13:06:44 +0200</pubDate>
+    <lastBuildDate>Thu, 23 Jul 2026 13:06:44 +0200</lastBuildDate>
+    <managingEditor>redaktion-support@bva.bund.de</managingEditor>
+    <webMaster>redaktion-support@bva.bund.de</webMaster>
+    <generator>Government Site Builder (GSB7.0)</generator>
+    <image>
+      <url>http://www.service.bund.de/cae/servlet/contentblob/546/normal/66/logo_a.jpg</url>
+      <title>service.bund.de - öffentliche Ausschreibungen</title>
+      <link>http://www.service.bund.de/ausschreibungen</link>
+    </image>
+    <docs>http://blogs.law.harvard.edu/tech/rss</docs>
+    <ttl>60</ttl>
+    <item>
+<title>&#220;bermittlung von Bohrungsdaten in die Bohrpunktkarte Deutschland 4.0 Neuaufstellung des Datenformats</title>
+<link>https://www.service.bund.de/IMPORTE/Ausschreibungen/editor/Bundesanstalt-fuer-Geowissenschaften-und-Rohstoffe/2026/07/6585375.html#track=feed-callforbids</link>
+<description>
+<![CDATA[
+   Erf&uuml;llungsort: <strong> Hannover</strong>
+<br />Vergabestelle: <strong>Bundesanstalt f&#252;r Geowissenschaften und Rohstoffe</strong><br />
+<br />Angebotsfrist: <br />Veröffentlichungsende:  <strong>23.10.2026 11:15</strong>  <br />
+<br />
+]]>
+</description>
+<pubDate>Thu, 23 Jul 2026 11:15:00 +0200</pubDate>
+<guid>https://www.service.bund.de/IMPORTE/Ausschreibungen/editor/Bundesanstalt-fuer-Geowissenschaften-und-Rohstoffe/2026/07/6585375.html</guid>
+    </item>
+<item>
+<title>Vhv 120_26 UFZ - Bench-scale Anaerobic Bioreactor System</title>
+<link>https://www.service.bund.de/IMPORTE/Ausschreibungen/editor/Helmholtz-Zentrum-fuer-Umweltforschung-GmbH/2026/07/6585334.html#track=feed-callforbids</link>
+<description>
+<![CDATA[
+   Erf&uuml;llungsort: <strong>04318 Leipzig</strong>
+<br />Vergabestelle: <strong>Helmholtz-Zentrum f&#252;r Umweltforschung GmbH - UFZ</strong><br />
+<br />Angebotsfrist:  <strong>07.08.2026 10:00</strong> <br />Veröffentlichungsende:  <strong>07.08.2026 10:00</strong>  <br />
+<br />
+]]>
+</description>
+<pubDate>Thu, 23 Jul 2026 11:00:00 +0200</pubDate>
+<guid>https://www.service.bund.de/IMPORTE/Ausschreibungen/editor/Helmholtz-Zentrum-fuer-Umweltforschung-GmbH/2026/07/6585334.html</guid>
+    </item>
+<item>
+<title>Verkehrsleistungen Linienb&#252;ndel 5 &quot;Birgland&quot; (VGN-Linien 475 &#8211; 479)</title>
+<link>https://www.service.bund.de/IMPORTE/Ausschreibungen/editor/Zweckverband-Nahverkehr-Amberg-Sulzbach/2026/07/6585166.html#track=feed-callforbids</link>
+<description>
+<![CDATA[
+   Erf&uuml;llungsort: <strong>92224 Amberg</strong>
+<br />Vergabestelle: <strong>Zweckverband Nahverkehr Amberg-Sulzbach (ZNAS)</strong><br />
+<br />Angebotsfrist:  <strong>31.08.2026 11:00</strong> <br />Veröffentlichungsende:  <strong>31.08.2026 11:00</strong>  <br />
+<br />
+]]>
+</description>
+<pubDate>Thu, 23 Jul 2026 09:00:00 +0200</pubDate>
+<guid>https://www.service.bund.de/IMPORTE/Ausschreibungen/editor/Zweckverband-Nahverkehr-Amberg-Sulzbach/2026/07/6585166.html</guid>
+    </item>
+<item>
+<title>Plattformlifte</title>
+<link>https://www.service.bund.de/IMPORTE/Ausschreibungen/asp/2026/07/196439.html#track=feed-callforbids</link>
+<description>
+<![CDATA[
+   Erf&uuml;llungsort: <strong>86152 Augsburg</strong>
+<br />Vergabestelle: <strong>Stadt Augsburg</strong><br />
+<br />Angebotsfrist:  <strong>27.08.2026 11:30</strong> <br />Veröffentlichungsende:  <strong>27.08.2026 23:59</strong>  <br />
+<br />
+]]>
+</description>
+<pubDate>Thu, 23 Jul 2026 07:28:16 +0200</pubDate>
+<guid>https://www.service.bund.de/IMPORTE/Ausschreibungen/asp/2026/07/196439.html</guid>
+    </item>
+<item>
+<title>Hubbuehnen 1</title>
+<link>https://www.service.bund.de/IMPORTE/Ausschreibungen/asp/2026/07/196438.html#track=feed-callforbids</link>
+<description>
+<![CDATA[
+   Erf&uuml;llungsort: <strong>86152 Augsburg</strong>
+<br />Vergabestelle: <strong>Stadt Augsburg</strong><br />
+<br />Angebotsfrist:  <strong>27.08.2026 11:00</strong> <br />Veröffentlichungsende:  <strong>27.08.2026 23:59</strong>  <br />
+<br />
+]]>
+</description>
+<pubDate>Thu, 23 Jul 2026 07:13:09 +0200</pubDate>
+<guid>https://www.service.bund.de/IMPORTE/Ausschreibungen/asp/2026/07/196438.html</guid>
+    </item>
+<item>
+<title>B 278 Flutgrabenbr&#252;cke bei Borsch, ASB-Nr. 5225 805; Ersatzneubau</title>
+<link>https://www.service.bund.de/IMPORTE/Ausschreibungen/eVergabe/877436.html#track=feed-callforbids</link>
+<description>
+<![CDATA[
+   Erf&uuml;llungsort: <strong> Freistaat Th&#252;ringen, Landkreis Wartburgkreis, B 278 Flutgrabenbr&#252;cke bei Borsch ASB-Nr. 5225 805</strong>
+<br />Vergabestelle: <strong>Th&#252;ringer Landesamt f&#252;r Bau und Verkehr, Stra&#223;enbauverwaltung und Zentrale Dienste</strong><br />
+<br />Angebotsfrist:  <strong>19.08.2026 10:00</strong> <br />Veröffentlichungsende:  <strong>19.08.2026 10:00</strong>  <br />
+<br />
+]]>
+</description>
+<pubDate>Thu, 23 Jul 2026 05:00:00 +0200</pubDate>
+<guid>https://www.service.bund.de/IMPORTE/Ausschreibungen/eVergabe/877436.html</guid>
+    </item>
+  </channel>
+</rss>
@@ -0,0 +1,45 @@
+<?xml version='1.0' encoding='utf-8'?>
+<rss version='2.0'>
+<channel>
+<title><![CDATA[subreport ELViS]]></title>
+<link>https://www.subreport-elvis.de/</link>
+<description><![CDATA[Aktuelle Ausschreibungen von Stadt Neuss, 41460 Neuss]]></description>
+<language>de-de</language>
+<copyright>subreport ELViS</copyright>
+<item>
+<title><![CDATA[E73433797:Objektbezogene Schadensanalysen für Brückenbauwerke im alten Bahndamm bei Minkeln-66-26-48 Objektbezogene Schadensanalysen für Brückenbauwerke im alten Bahndamm bei Minkeln]]></title>
+<description><![CDATA[<table><tr><th style='text-align:left;'>Auftraggeber</th><td>Stadt Neuss, 41460 Neuss</td></tr><tr><th style='text-align:left;'>Vergabenummer</th><td>66-26-48</td></tr><tr><th style='text-align:left;'>Ausschreibungsart</th><td>Öffentliche Ausschreibung nach UVgO</td></tr><tr><th style='text-align:left;'>ELViS-ID</th><td>E73433797</td></tr><tr><th style='text-align:left;'>Ausschreibung</th><td>Objektbezogene Schadensanalysen für Brückenbauwerke im alten Bahndamm bei Minkeln</td></tr><tr><th style='text-align:left;'>Vergabeunterlagen</th><td>66-26-48 Objektbezogene Schadensanalysen für Brückenbauwerke im alten Bahndamm bei Minkeln</td></tr><tr><th style='text-align:left;'>Gültig bis</th><td>27.08.2026 10:00:00</td></tr><tr><th style='text-align:left;'>Angebotsabgabefrist</th><td>27.08.2026 10:00:00</td></tr><tr><th style='text-align:left;'>Öffnungstermin</th><td>27.08.2026 10:00:00</td></tr><tr><th style='text-align:left;'>subreport</th><td><a href='https://www.subreport-elvis.de/browseVerdingungsunterlagen.html#ELVISID:E73433797,558464' target='_new'>zur Vorschau in die Vergabeunterlagen</a></td></tr></table>]]></description>
+<link>https://www.subreport-elvis.de/browseVerdingungsunterlagen.html#ELVISID:E73433797</link>
+<author><![CDATA[Stadt Neuss, vergabe@stadt.neuss.de]]></author>
+<guid isPermaLink="false">E73433797-558464</guid>
+</item>
+<item>
+<title><![CDATA[E81865426:Objektbezogene Schadensanalysen für Brückenbauwerke im alten Bahndamm bei Minkeln-66-26-47 Objektbezogene Schadensanalysen für Brückenbauwerke im alten Bahndamm bei Minkeln]]></title>
+<description><![CDATA[<table><tr><th style='text-align:left;'>Auftraggeber</th><td>Stadt Neuss, 41460 Neuss</td></tr><tr><th style='text-align:left;'>Vergabenummer</th><td>66-26-47</td></tr><tr><th style='text-align:left;'>Ausschreibungsart</th><td>Öffentliche Ausschreibung nach UVgO</td></tr><tr><th style='text-align:left;'>ELViS-ID</th><td>E81865426</td></tr><tr><th style='text-align:left;'>Ausschreibung</th><td>Objektbezogene Schadensanalysen für Brückenbauwerke im alten Bahndamm bei Minkeln</td></tr><tr><th style='text-align:left;'>Vergabeunterlagen</th><td>66-26-47 Objektbezogene Schadensanalysen für Brückenbauwerke im alten Bahndamm bei Minkeln</td></tr><tr><th style='text-align:left;'>Gültig bis</th><td>26.08.2026 10:00:00</td></tr><tr><th style='text-align:left;'>Angebotsabgabefrist</th><td>26.08.2026 10:00:00</td></tr><tr><th style='text-align:left;'>Öffnungstermin</th><td>26.08.2026 10:00:00</td></tr><tr><th style='text-align:left;'>subreport</th><td><a href='https://www.subreport-elvis.de/browseVerdingungsunterlagen.html#ELVISID:E81865426,558461' target='_new'>zur Vorschau in die Vergabeunterlagen</a></td></tr></table>]]></description>
+<link>https://www.subreport-elvis.de/browseVerdingungsunterlagen.html#ELVISID:E81865426</link>
+<author><![CDATA[Stadt Neuss, vergabe@stadt.neuss.de]]></author>
+<guid isPermaLink="false">E81865426-558461</guid>
+</item>
+<item>
+<title><![CDATA[E95748761:Lieferung eines Gerätewagens Logistik 2 (GW-L2)-37-26-44 Lieferung eines Gerätewagens Logistik 2 (GW-L2)]]></title>
+<description><![CDATA[<table><tr><th style='text-align:left;'>Auftraggeber</th><td>Stadt Neuss, 41460 Neuss</td></tr><tr><th style='text-align:left;'>Vergabenummer</th><td>37-26-44</td></tr><tr><th style='text-align:left;'>Ausschreibungsart</th><td>Offenes Verfahren nach VGV</td></tr><tr><th style='text-align:left;'>ELViS-ID</th><td>E95748761</td></tr><tr><th style='text-align:left;'>Ausschreibung</th><td>Lieferung eines Gerätewagens Logistik 2 (GW-L2)</td></tr><tr><th style='text-align:left;'>Vergabeunterlagen</th><td>37-26-44 Lieferung eines Gerätewagens Logistik 2 (GW-L2)</td></tr><tr><th style='text-align:left;'>Gültig bis</th><td>24.08.2026 10:00:00</td></tr><tr><th style='text-align:left;'>Angebotsabgabefrist</th><td>24.08.2026 10:00:00</td></tr><tr><th style='text-align:left;'>Öffnungstermin</th><td>24.08.2026 10:00:00</td></tr><tr><th style='text-align:left;'>subreport</th><td><a href='https://www.subreport-elvis.de/browseVerdingungsunterlagen.html#ELVISID:E95748761,557626' target='_new'>zur Vorschau in die Vergabeunterlagen</a></td></tr></table>]]></description>
+<link>https://www.subreport-elvis.de/browseVerdingungsunterlagen.html#ELVISID:E95748761</link>
+<author><![CDATA[Stadt Neuss, vergabe@stadt.neuss.de]]></author>
+<guid isPermaLink="false">E95748761-557626</guid>
+</item>
+<item>
+<title><![CDATA[E96237741:Lieferung von Lichtmasten, Jahresbedarf 2026/2027-66-26-41 Lieferung von Lichtmasten, Jahresbedarf 2026/2027]]></title>
+<description><![CDATA[<table><tr><th style='text-align:left;'>Auftraggeber</th><td>Stadt Neuss, 41460 Neuss</td></tr><tr><th style='text-align:left;'>Vergabenummer</th><td>66-26-41</td></tr><tr><th style='text-align:left;'>Ausschreibungsart</th><td>Öffentliche Ausschreibung nach UVgO</td></tr><tr><th style='text-align:left;'>ELViS-ID</th><td>E96237741</td></tr><tr><th style='text-align:left;'>Ausschreibung</th><td>Lieferung von Lichtmasten, Jahresbedarf 2026/2027</td></tr><tr><th style='text-align:left;'>Vergabeunterlagen</th><td>66-26-41 Lieferung von Lichtmasten, Jahresbedarf 2026/2027</td></tr><tr><th style='text-align:left;'>Gültig bis</th><td>27.07.2026 10:00:00</td></tr><tr><th style='text-align:left;'>Angebotsabgabefrist</th><td>27.07.2026 10:00:00</td></tr><tr><th style='text-align:left;'>Öffnungstermin</th><td>27.07.2026 10:00:00</td></tr><tr><th style='text-align:left;'>subreport</th><td><a href='https://www.subreport-elvis.de/browseVerdingungsunterlagen.html#ELVISID:E96237741,556451' target='_new'>zur Vorschau in die Vergabeunterlagen</a></td></tr></table>]]></description>
+<link>https://www.subreport-elvis.de/browseVerdingungsunterlagen.html#ELVISID:E96237741</link>
+<author><![CDATA[Stadt Neuss, vergabe@stadt.neuss.de]]></author>
+<guid isPermaLink="false">E96237741-556451</guid>
+</item>
+<item>
+<title><![CDATA[E87453261:Sicherheitsdienstleistungen und Brandschutzwachdienst in städtischen Unterbringungseinrichtungen der Stadt Neuss-53-26-40 Sicherheitsdienstleistungen und Brandschutzwachdienst in städtischen Unterbringungseinrichtungen der Stadt Neuss]]></title>
+<description><![CDATA[<table><tr><th style='text-align:left;'>Auftraggeber</th><td>Stadt Neuss, 41460 Neuss</td></tr><tr><th style='text-align:left;'>Vergabenummer</th><td>53-26-40</td></tr><tr><th style='text-align:left;'>Ausschreibungsart</th><td>Offenes Verfahren nach VGV</td></tr><tr><th style='text-align:left;'>ELViS-ID</th><td>E87453261</td></tr><tr><th style='text-align:left;'>Ausschreibung</th><td>Sicherheitsdienstleistungen und Brandschutzwachdienst in städtischen Unterbringungseinrichtungen der Stadt Neuss</td></tr><tr><th style='text-align:left;'>Vergabeunterlagen</th><td>53-26-40 Sicherheitsdienstleistungen und Brandschutzwachdienst in städtischen Unterbringungseinrichtungen der Stadt Neuss</td></tr><tr><th style='text-align:left;'>Gültig bis</th><td>13.08.2026 10:00:00</td></tr><tr><th style='text-align:left;'>Angebotsabgabefrist</th><td>13.08.2026 10:00:00</td></tr><tr><th style='text-align:left;'>Öffnungstermin</th><td>13.08.2026 10:00:00</td></tr><tr><th style='text-align:left;'>subreport</th><td><a href='https://www.subreport-elvis.de/browseVerdingungsunterlagen.html#ELVISID:E87453261,555142' target='_new'>zur Vorschau in die Vergabeunterlagen</a></td></tr></table>]]></description>
+<link>https://www.subreport-elvis.de/browseVerdingungsunterlagen.html#ELVISID:E87453261</link>
+<author><![CDATA[Stadt Neuss, vergabe@stadt.neuss.de]]></author>
+<guid isPermaLink="false">E87453261-555142</guid>
+</item>
+</channel>
+</rss>
@@ -0,0 +1,189 @@
+import { readFileSync } from 'fs';
+import { join } from 'path';
+import { describe, expect, it } from 'vitest';
+import { RssAdapter } from './rss.adapter';
+
+/**
+ * Real, live-captured RSS fixtures (both fetched directly against the real
+ * endpoints, 2026-07-23) — mirrors cosinex.adapter.spec.ts's fixture/mock
+ * spec style. `service-bund-feed.xml` proves the pubDate-present,
+ * numeric-HTML-entity-title shape; `subreport-elvis-feed.xml` proves the
+ * pubDate-absent, CDATA-title shape (14-RESEARCH.md live captures).
+ */
+const FIXTURES_DIR = join(__dirname, '..', '__fixtures__');
+const SERVICE_BUND_FIXTURE = readFileSync(
+  join(FIXTURES_DIR, 'service-bund-feed.xml'),
+  'utf8',
+);
+const SUBREPORT_ELVIS_FIXTURE = readFileSync(
+  join(FIXTURES_DIR, 'subreport-elvis-feed.xml'),
+  'utf8',
+);
+
+describe('RssAdapter', () => {
+  it('declares sourceType rss and the symbolic rss portal', () => {
+    const adapter = new RssAdapter();
+    expect(adapter.sourceType).toBe('rss');
+    expect(adapter.portals).toEqual(['rss']);
+  });
+
+  describe('parseFeed (pure, fixture-driven)', () => {
+    it('parses service.bund.de items: sourceType/sourcePortal set, pubDate present, numeric HTML entities decoded in title', () => {
+      const adapter = new RssAdapter();
+
+      const records = adapter.parseFeed(SERVICE_BUND_FIXTURE, 'service-bund');
+
+      expect(records.length).toBeGreaterThanOrEqual(1);
+      for (const record of records) {
+        expect(record.sourceType).toBe('rss');
+        expect(record.sourcePortal).toBe('service-bund');
+        expect(record.sourceUrl).toMatch(/^https:\/\/www\.service\.bund\.de\//);
+        expect(record.eformsPayload).toBeNull();
+      }
+
+      // First fixture item's title is numeric-HTML-entity-encoded
+      // (`&#220;bermittlung...`), NOT CDATA — must decode to a real "Ü".
+      const first = records[0];
+      expect(first).toBeDefined();
+      const payload = first!.ocdsPayload as { title: string };
+      expect(payload.title).toMatch(/^Übermittlung von Bohrungsdaten/);
+      expect(payload.title).not.toContain('&#');
+
+      // Every service.bund.de fixture item carries a <pubDate>.
+      expect(records.every((r) => r.publishedAt instanceof Date)).toBe(true);
+    });
+
+    it('uses the item guid as sourceNoticeId for service.bund.de (plain-text guid, no isPermaLink attribute)', () => {
+      const adapter = new RssAdapter();
+      const records = adapter.parseFeed(SERVICE_BUND_FIXTURE, 'service-bund');
+
+      const ids = records.map((r) => r.sourceNoticeId);
+      expect(ids.every((id) => id.startsWith('https://'))).toBe(true);
+      expect(new Set(ids).size).toBe(ids.length);
+    });
+
+    it('parses subreport-elvis items: publishedAt null (no item pubDate), title from CDATA, guid isPermaLink=false extracted as plain text', () => {
+      const adapter = new RssAdapter();
+
+      const records = adapter.parseFeed(
+        SUBREPORT_ELVIS_FIXTURE,
+        'subreport-neuss',
+      );
+
+      expect(records.length).toBeGreaterThanOrEqual(1);
+      for (const record of records) {
+        expect(record.sourceType).toBe('rss');
+        expect(record.sourcePortal).toBe('subreport-neuss');
+        expect(record.publishedAt).toBeNull();
+        expect(record.sourceUrl).toMatch(
+          /^https:\/\/www\.subreport-elvis\.de\//,
+        );
+      }
+
+      const first = records[0];
+      expect(first).toBeDefined();
+      // guid is `<guid isPermaLink="false">E73433797-558464</guid>` —
+      // fast-xml-parser represents this as { '#text': ..., '@_isPermaLink': ... },
+      // extractTagText must pull out the plain '#text' value.
+      expect(first!.sourceNoticeId).toBe('E73433797-558464');
+
+      const payload = first!.ocdsPayload as { title: string };
+      expect(payload.title).toMatch(/^E73433797:/);
+    });
+
+    it('never carries description HTML into the record (V5 — no stored-XSS surface)', () => {
+      const adapter = new RssAdapter();
+      const records = adapter.parseFeed(
+        SUBREPORT_ELVIS_FIXTURE,
+        'subreport-neuss',
+      );
+
+      for (const record of records) {
+        const serialized = JSON.stringify(record.ocdsPayload);
+        expect(serialized).not.toContain('<table>');
+        expect(serialized).not.toContain('<a href');
+      }
+    });
+
+    it('bare-minimum bag: buyerName/procedureType/deadlineAt always null (baseline mapping only, D-04/D-05)', () => {
+      const adapter = new RssAdapter();
+      const records = adapter.parseFeed(SERVICE_BUND_FIXTURE, 'service-bund');
+
+      for (const record of records) {
+        const payload = record.ocdsPayload as {
+          buyerName: string | null;
+          procedureType: string | null;
+          legalFramework: string | null;
+          deadlineAt: string | null;
+        };
+        expect(payload.buyerName).toBeNull();
+        expect(payload.procedureType).toBeNull();
+        expect(payload.legalFramework).toBeNull();
+        expect(payload.deadlineAt).toBeNull();
+      }
+    });
+
+    it('returns [] for empty XML instead of throwing', () => {
+      const adapter = new RssAdapter();
+      expect(adapter.parseFeed('', 'empty-feed')).toEqual([]);
+    });
+
+    it('returns [] for malformed XML instead of throwing', () => {
+      const adapter = new RssAdapter();
+      expect(
+        adapter.parseFeed('<rss><channel><item><title>', 'broken-feed'),
+      ).toEqual([]);
+    });
+
+    it('returns [] for well-formed XML with no <item> at all', () => {
+      const adapter = new RssAdapter();
+      const xml =
+        '<?xml version="1.0"?><rss version="2.0"><channel><title>Empty</title></channel></rss>';
+      expect(adapter.parseFeed(xml, 'no-items-feed')).toEqual([]);
+    });
+
+    it('skips a single item missing <link> without aborting the rest', () => {
+      const adapter = new RssAdapter();
+      const xml = `<?xml version="1.0"?>
+        <rss version="2.0"><channel>
+          <item><title>Ohne Link</title><guid>no-link-guid</guid></item>
+          <item><title>Mit Link</title><link>https://example.invalid/a</link><guid>with-link-guid</guid></item>
+        </channel></rss>`;
+
+      const records = adapter.parseFeed(xml, 'mixed-feed');
+
+      expect(records).toHaveLength(1);
+      expect(records[0]?.sourceNoticeId).toBe('with-link-guid');
+    });
+
+    it('falls back to sha256(link) for sourceNoticeId when guid is absent', () => {
+      const adapter = new RssAdapter();
+      const xml = `<?xml version="1.0"?>
+        <rss version="2.0"><channel>
+          <item><title>No Guid</title><link>https://example.invalid/no-guid</link></item>
+        </channel></rss>`;
+
+      const records = adapter.parseFeed(xml, 'no-guid-feed');
+
+      expect(records).toHaveLength(1);
+      expect(records[0]?.sourceNoticeId).toMatch(/^[a-f0-9]{40}$/);
+    });
+  });
+
+  describe('normalize() dispatch', () => {
+    it("'rss' sourceType routes through TenderNormalizerService.normalizeBag() (proven end-to-end via tender-normalizer.service.spec.ts)", () => {
+      // See tender-normalizer.service.spec.ts's "generic ocdsPayload bag"
+      // describe block for the actual normalize() assertions — this spec
+      // only proves RssAdapter's OWN output shape (parseFeed), not the
+      // normalizer dispatch itself (kept in the normalizer's own spec file
+      // to avoid duplicating TenderNormalizerService test infrastructure).
+      const adapter = new RssAdapter();
+      expect(adapter.sourceType).toBe('rss');
+    });
+  });
+
+  it('never imports or uses axios (native fetch is the sole HTTP client convention)', () => {
+    const source = readFileSync(join(__dirname, 'rss.adapter.ts'), 'utf8');
+    expect(source).not.toMatch(/from ['"]axios['"]/);
+  });
+});
@@ -0,0 +1,202 @@
+import { Injectable, Logger } from '@nestjs/common';
+import { createHash } from 'crypto';
+import { XMLParser } from 'fast-xml-parser';
+import type { RawTenderRecord, SourceType } from '../tender.types';
+import type { TenderSourceAdapter } from './tender-source-adapter.interface';
+
+/**
+ * RssAdapter — INGEST-04. Parses admin-managed RSS feed URLs into
+ * RawTenderRecord[] via the same `fast-xml-parser` configuration
+ * doe-opendata.adapter.ts already uses for eForms-DE XML
+ * (`removeNSPrefix`/`ignoreAttributes`/`attributeNamePrefix` —
+ * 14-RESEARCH.md "RSS parsing").
+ *
+ * `portals: ['rss']` is a SYMBOLIC placeholder, not a real hostname list
+ * (14-RESEARCH.md Pitfall 3): the actual feed hostnames are admin-supplied
+ * RUNTIME data (`TenderRssFeedSource` rows, wired in Plan 14-02 Task 2),
+ * added long after `SourceRegistry.register()`'s DI-boot-time denylist
+ * check runs. The `DENYLISTED_PORTALS` gate therefore provides ZERO
+ * protection for RSS feed URLs — a SEPARATE, save-time hostname/SSRF guard
+ * is enforced in `tender-rss-feed.service.ts` (D-14, Plan 14-02 Task 2, see
+ * threat T-14-02-01). Do not remove that guard under the assumption this
+ * adapter's `portals` array already covers it.
+ *
+ * Live-verified shapes (14-RESEARCH.md, captured 2026-07-23):
+ * - service.bund.de: `<item><pubDate>` present; `<title>` uses raw numeric
+ *   HTML character references (e.g. `&#220;bermittlung...`), NOT CDATA —
+ *   decoded explicitly below (Rule 1: fast-xml-parser only decodes the 5
+ *   predefined XML entities by default, not numeric character refs).
+ * - subreport-elvis: `<item><pubDate>` is ABSENT entirely (`publishedAt`
+ *   is always null for this source); `<title>`/`<description>` are
+ *   CDATA-wrapped (auto-merged to plain strings by fast-xml-parser, no
+ *   decoding needed).
+ *
+ * Baseline field mapping only (title/link/guid) — `buyerName`/
+ * `procedureType`/`deadlineAt` stay null. The optional service.bund.de
+ * `<description>` label-extraction enrichment (Erfüllungsort/
+ * Vergabestelle/Angebotsfrist) is explicitly out of scope for this task
+ * (RESEARCH.md Pattern 3 "optional enhancement"); `<description>` is never
+ * read by this adapter, so no raw HTML fragment is ever carried into a
+ * RawTenderRecord (V5 — no stored-XSS surface, same text-only discipline
+ * as NetServerAdapter/CosinexAdapter).
+ */
+@Injectable()
+export class RssAdapter implements TenderSourceAdapter {
+  readonly sourceType: SourceType = 'rss';
+  /** Symbolic placeholder — see class docstring re: Pitfall 3. */
+  readonly portals = ['rss'] as const;
+
+  private readonly logger = new Logger(RssAdapter.name);
+  private readonly xmlParser = new XMLParser({
+    removeNSPrefix: true,
+    ignoreAttributes: false,
+    attributeNamePrefix: '@_',
+  });
+
+  /**
+   * Placeholder — wired to the real `TenderRssFeedSource` internal fan-out
+   * (native fetch + AbortController per admin-added feed URL) in Plan
+   * 14-02 Task 2. Kept here only so the class satisfies
+   * `TenderSourceAdapter` for this task's fixture-driven `parseFeed` proof.
+   */
+  async fetchTenders(_dayCursor: string): Promise<RawTenderRecord[]> {
+    return [];
+  }
+
+  /**
+   * Pure XML -> RawTenderRecord[] mapping. Never throws (Pitfall 2
+   * discipline, mirrors NetServer/cosinex): malformed/empty XML, or a feed
+   * with no `<channel>`/`<item>` at all, resolves to `[]`. A single item
+   * missing `<link>` is skipped (logger.warn) without aborting the rest.
+   */
+  parseFeed(xml: string, feedLabel: string): RawTenderRecord[] {
+    const fetchedAt = new Date();
+    let items: unknown[];
+
+    try {
+      const parsed = this.xmlParser.parse(xml) as {
+        rss?: { channel?: { item?: unknown | unknown[] } };
+      };
+      const channelItem = parsed.rss?.channel?.item;
+      items = Array.isArray(channelItem)
+        ? channelItem
+        : channelItem
+          ? [channelItem]
+          : [];
+    } catch (error) {
+      this.logger.warn(
+        `RSS feed '${feedLabel}' totally unparsable, returning [] (mirrors NetServer/cosinex Pitfall 2): ${(error as Error).message}`,
+      );
+      return [];
+    }
+
+    const records: RawTenderRecord[] = [];
+
+    for (const item of items) {
+      try {
+        const node = item as Record<string, unknown>;
+        const link = extractTagText(node.link);
+        if (!link) {
+          this.logger.warn(
+            `Skipping RSS item without <link> (feed '${feedLabel}')`,
+          );
+          continue; // no stable URL -> unusable item, skip (Pitfall 2 discipline)
+        }
+
+        const rawTitle = extractTagText(node.title);
+        const title = rawTitle
+          ? decodeNumericEntities(rawTitle).trim()
+          : 'Unbenannte Ausschreibung';
+
+        const guidText = extractTagText(node.guid);
+        const sourceNoticeId =
+          guidText ||
+          createHash('sha256').update(link).digest('hex').slice(0, 40);
+
+        const pubDateRaw = extractTagText(node.pubDate);
+        const publishedAt = parseRssDate(pubDateRaw);
+
+        records.push({
+          sourceType: this.sourceType,
+          sourcePortal: feedLabel,
+          sourceNoticeId,
+          sourceUrl: link,
+          fetchedAt,
+          publishedAt,
+          eformsPayload: null,
+          // RSS has no eForms/OCDS structure; the extracted baseline fields
+          // are carried through this generic bag — same convention as
+          // NetServerAdapter/CosinexAdapter (13-04/13-05) — so
+          // TenderNormalizerService.normalizeBag() maps them without any
+          // RSS-specific normalizer code (D-04/D-05).
+          ocdsPayload: {
+            title,
+            buyerName: null,
+            procedureType: null,
+            legalFramework: null,
+            deadlineAt: null,
+          },
+        });
+      } catch (error) {
+        this.logger.warn(
+          `Skipping unparsable RSS item (feed '${feedLabel}'): ${(error as Error).message}`,
+        );
+      }
+    }
+
+    return records;
+  }
+}
+
+/**
+ * fast-xml-parser represents a plain tag as a raw string/number, and a tag
+ * with attributes (e.g. subreport-elvis's `<guid isPermaLink="false">`) as
+ * `{ '@_attr': ..., '#text': value }` — mirrors
+ * tender-normalizer.service.ts's `textValue()` helper (kept local/
+ * duplicated rather than imported: this is adapter-layer XML-shape
+ * handling, not normalizer-layer field mapping).
+ */
+function extractTagText(node: unknown): string | null {
+  if (node === null || node === undefined) return null;
+  if (typeof node === 'string') return node || null;
+  if (typeof node === 'number') return String(node);
+  if (
+    typeof node === 'object' &&
+    '#text' in (node as Record<string, unknown>)
+  ) {
+    const t = (node as Record<string, unknown>)['#text'];
+    if (t === null || t === undefined) return null;
+    return String(t);
+  }
+  return null;
+}
+
+/**
+ * Rule 1 fix: fast-xml-parser only decodes the 5 predefined XML entities
+ * (`&amp;` `&lt;` `&gt;` `&quot;` `&apos;`) — numeric character references
+ * (`&#220;`, `&#x00DF;`) pass through UNDECODED (confirmed live against
+ * fast-xml-parser v5.10.1, see class docstring). service.bund.de titles use
+ * numeric refs directly (not CDATA-wrapped, unlike subreport-elvis) —
+ * without this decode step, titles would literally show "&#220;bermittlung
+ * ..." to admins/users instead of "Übermittlung...".
+ */
+function decodeNumericEntities(text: string): string {
+  return text
+    .replace(/&#(\d+);/g, (_match, dec: string) =>
+      String.fromCodePoint(Number(dec)),
+    )
+    .replace(/&#x([0-9a-fA-F]+);/g, (_match, hex: string) =>
+      String.fromCodePoint(Number.parseInt(hex, 16)),
+    );
+}
+
+/**
+ * RSS `pubDate` is RFC-822 (`Thu, 23 Jul 2026 11:15:00 +0200`), parseable
+ * directly by `Date`. Absent/malformed -> null (subreport-elvis has no
+ * per-item `pubDate` at all — confirmed live, see class docstring).
+ */
+function parseRssDate(raw: string | null): Date | null {
+  if (!raw) return null;
+  const parsed = new Date(raw);
+  return Number.isNaN(parsed.getTime()) ? null : parsed;
+}
@@ -81,7 +81,11 @@ function bagRecord(
  },
 ): RawTenderRecord {
  const sourcePortal =
-    sourceType === 'cosinex-dtvp' ? 'cosinex-dtvp' : 'tender24';
+    sourceType === 'cosinex-dtvp'
+      ? 'cosinex-dtvp'
+      : sourceType === 'rss'
+        ? 'service-bund'
+        : 'tender24';
  return {
    sourceType,
    sourcePortal,
@@ -317,4 +321,38 @@ describe('TenderNormalizerService', () => {
      expect(normalized.contentHash).toMatch(/^[a-f0-9]{64}$/);
    });
  });
+
+  describe("'rss' sourceType (Phase 14, Plan 02, INGEST-04)", () => {
+    it('maps an rss bag through normalizeBag(): title passes through, cpvDivisions stays empty', () => {
+      const raw = bagRecord('rss', {
+        title: 'Übermittlung von Bohrungsdaten',
+        buyerName: null,
+        procedureType: null,
+        legalFramework: null,
+        deadlineAt: null,
+      });
+
+      const normalized = service.normalize(raw);
+
+      expect(normalized.title).toBe('Übermittlung von Bohrungsdaten');
+      expect(normalized.cpvDivisions).toEqual([]);
+      expect(normalized.cpvCodes).toEqual([]);
+      expect(normalized.region).toBeNull();
+      expect(normalized.bundesland).toBeNull();
+      expect(normalized.buyerName).toBeNull();
+      expect(normalized.procedureType).toBeNull();
+      expect(normalized.deadlineAt).toBeNull();
+      expect(normalized.dedupKey).toBe(
+        `${raw.sourcePortal}:${raw.sourceNoticeId}`,
+      );
+    });
+
+    it('rss title falls back to "Unbenannte Ausschreibung" when empty', () => {
+      const raw = bagRecord('rss', { title: '' });
+
+      const normalized = service.normalize(raw);
+
+      expect(normalized.title).toBe('Unbenannte Ausschreibung');
+    });
+  });
 });
@@ -54,6 +54,7 @@ export class TenderNormalizerService {
    switch (raw.sourceType) {
      case 'ai-netserver':
      case 'cosinex-dtvp':
+      case 'rss':
        return this.normalizeBag(raw);
      case 'doe-opendata':
      default:
@@ -117,10 +118,13 @@ export class TenderNormalizerService {

  /**
   * Generic-bag path (gap closure): NetServer/cosinex-DTVP scraper adapters
-   * carry a flat `{title, buyerName, procedureType, legalFramework,
-   * deadlineAt}` bag in `raw.ocdsPayload` — there is no eForms/OCDS
-   * structure to inspect. `legalFramework` is deliberately NOT mapped:
-   * NormalizedTenderFields has no target field for it.
+   * (and, since Phase 14 Plan 02, the RssAdapter, INGEST-04) carry a flat
+   * `{title, buyerName, procedureType, legalFramework, deadlineAt}` bag in
+   * `raw.ocdsPayload` — there is no eForms/OCDS structure to inspect.
+   * `legalFramework` is deliberately NOT mapped: NormalizedTenderFields has
+   * no target field for it. RSS records always have `buyerName`/
+   * `procedureType`/`deadlineAt` null (baseline title/link/guid mapping
+   * only, D-04/D-05 thin-fields deferral).
   */
  private normalizeBag(raw: RawTenderRecord): NormalizedTenderFields {
    const bag = getBagPayload(raw.ocdsPayload);
@@ -13,8 +13,16 @@
 * (Plan 13-05) adapters can be added without further breaking changes.
 * This union widening is the only breaking type change in Plan 13-01,
 * and it's contained locally (no runtime behavior change here).
+ *
+ * Phase 14, Plan 02 (INGEST-04) adds 'rss': admin-managed RSS feed URLs
+ * (TenderRssFeedSource, D-14), routed through the same generic-bag
+ * normalizer path as 'ai-netserver'/'cosinex-dtvp' (D-04/D-05).
 */
-export type SourceType = 'doe-opendata' | 'ai-netserver' | 'cosinex-dtvp';
+export type SourceType =
+  | 'doe-opendata'
+  | 'ai-netserver'
+  | 'cosinex-dtvp'
+  | 'rss';

 /**
 * A single notice as fetched and lightly parsed from a source, before