Sync Parliament Atom fallback

src/sources.ts

@@ -155,8 +155,97 @@ function parseParliamentDate(text: string): string | null {
   return Number.isNaN(value.getTime()) ? null : value.toISOString();
 }
 
+async function fetchParliamentAtom(source: SourceDefinition): Promise<NormalizedItem[]> {
+  const feedUrl = "https://www3.parliament.nz/en/channellisting/rss/9343/e37f168d-4521-4437-a6fd-6bb7ff54e800";
+  const response = await fetch(feedUrl, {
+    headers: {
+      accept: "application/atom+xml,application/xml,text/xml;q=0.9,*/*;q=0.5",
+      "user-agent": "Mozilla/5.0 (compatible; ArrowstackMediaIntelligence/0.1; +https://media.arrowstack.dev)"
+    },
+    redirect: "follow",
+    signal: AbortSignal.timeout(20000)
+  });
+  if (!response.ok) {
+    throw new Error(`Parliament Atom feed HTTP ${response.status} ${response.statusText}`);
+  }
+
+  const xml = await response.text();
+  if (/radware page/i.test(xml) || /<title>\s*radware page\s*<\/title>/i.test(xml)) {
+    throw new Error(`Parliament Atom feed blocked by Radware (bytes=${xml.length})`);
+  }
+
+  const parser = new XMLParser({ ignoreAttributes: false, attributeNamePrefix: "@_" });
+  const parsed = parser.parse(xml);
+  let list = asArray<any>(
+    parsed?.feed?.entry ??
+    parsed?.rss?.channel?.item ??
+    parsed?.["atom:feed"]?.["atom:entry"]
+  );
+  if (!list.length) list = rawFeedItems(xml);
+
+  const items = list.slice(0, 30).map((entry: any, index: number) => {
+    const title = cleanXmlText(
+      entry.title?.["#text"] ??
+      entry["atom:title"]?.["#text"] ??
+      entry.title ??
+      entry["atom:title"] ??
+      "Parliament item"
+    );
+    const summary = cleanXmlText(
+      entry.summary?.["#text"] ??
+      entry["atom:summary"]?.["#text"] ??
+      entry.summary ??
+      entry["atom:summary"] ??
+      entry.description ??
+      entry.content ??
+      ""
+    );
+    const linkValue = entry.link ?? entry["atom:link"];
+    const url = firstLink(linkValue) || feedUrl;
+    const publishedAt =
+      cleanXmlText(
+        entry.updated ??
+        entry["atom:updated"] ??
+        entry.published ??
+        entry["atom:published"] ??
+        entry.pubDate ??
+        ""
+      ) || null;
+
+    return {
+      sourceId: "parliament" as const,
+      externalId: String(
+        entry.id?.["#text"] ??
+        entry["atom:id"]?.["#text"] ??
+        entry.id ??
+        entry["atom:id"] ??
+        entry.guid ??
+        url ??
+        index
+      ),
+      title,
+      summary,
+      url,
+      publishedAt,
+      topics: ["parliament", "select-committee", "evidence"],
+      raw: entry
+    };
+  }).filter((item: NormalizedItem) => item.title && item.url);
+
+  if (!items.length) {
+    throw new Error(
+      `Parliament Atom feed contained no entries (bytes=${xml.length}, prefix=${xml.slice(0, 120).replace(/\s+/g, " ")})`
+    );
+  }
+
+  return items;
+}
+
 async function fetchParliament(): Promise<NormalizedItem[]> {
   const source = SOURCES.find((s) => s.id === "parliament")!;
+  let htmlFailure = "";
+
+  try {
     const response = await fetch(source.url, {
       headers: {
         accept: "text/html,application/xhtml+xml",
@@ -167,10 +256,14 @@ async function fetchParliament(): Promise<NormalizedItem[]> {
       signal: AbortSignal.timeout(20000)
     });
     if (!response.ok) {
-    throw new Error(`${response.status} ${response.statusText} for ${source.url}`);
+      throw new Error(`${response.status} ${response.statusText}`);
     }
 
     const html = await response.text();
+    if (/radware page/i.test(html) || /<title>\s*radware page\s*<\/title>/i.test(html)) {
+      throw new Error(`blocked by Radware (bytes=${html.length})`);
+    }
+
     const $ = load(html);
     const seen = new Set<string>();
     const items: NormalizedItem[] = [];
@@ -228,12 +321,7 @@ async function fetchParliament(): Promise<NormalizedItem[]> {
       const container = $(element).closest("tr").length
         ? $(element).closest("tr")
         : $(element).closest("li, article, .list-item, div");
-    addItem(
-      href,
-      $(element).text(),
-      container.text(),
-      index
-    );
+      addItem(href, $(element).text(), container.text(), index);
     });
 
     if (!items.length) {
@@ -246,14 +334,20 @@ async function fetchParliament(): Promise<NormalizedItem[]> {
       }
     }
 
-  if (!items.length) {
-    const pageTitle = $("title").text().replace(/\s+/g, " ").trim();
-    throw new Error(
-      `Parliament evidence listing returned no document links (bytes=${html.length}, title=${pageTitle || "unknown"})`
-    );
+    if (items.length) return items;
+    htmlFailure = `HTML listing contained no document links (bytes=${html.length})`;
+  } catch (error) {
+    htmlFailure = error instanceof Error ? error.message : String(error);
   }
 
-  return items;
+  try {
+    return await fetchParliamentAtom(source);
+  } catch (error) {
+    const atomFailure = error instanceof Error ? error.message : String(error);
+    throw new Error(
+      `Parliament unavailable from Arrowstack. HTML: ${htmlFailure || "unknown failure"}; Atom: ${atomFailure}`
+    );
+  }
 }
 
 interface SitemapEntry {