Fix multi-month scrape coverage and add duo MUSIC EXCHANGE

- Extend 8 scrapers (liquid-room, shibuya-o, club-quattro, meets-otsuka,
  nishieifuku-jam, fever-shindaita, fad-yokohama, and new duo-music-exchange)
  to fetch 3 calendar months instead of 1-2, covering the full 65-day window
- Add duo MUSIC EXCHANGE scraper (渋谷, ~700 cap, /schedule/YYYY/index_YYYY-MM.html)
- Add npm test: Node.js built-in test runner verifies each scraper fetches
  all required month URLs via mocked fetch (10 tests, no extra deps)

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

1 files changed, 62 insertions, 48 deletions

diff --git a/app/scrapers/liquid-room.ts b/app/scrapers/liquid-room.ts
index 1eeade6..a1265c8 100644
--- a/app/scrapers/liquid-room.ts
+++ b/app/scrapers/liquid-room.ts
@@ -10,62 +10,76 @@ export const venue: VenueMeta = {
   capacity: 1000,
 };
 
-export const scraper: Scraper = {
-  venue,
-  async scrape(): Promise<EventInput[]> {
-    const res = await fetch("https://www.liquidroom.net/schedule");
-    if (!res.ok) throw new Error(`HTTP ${res.status}`);
-    const html = await res.text();
-    const $ = cheerio.load(html);
-    const events: EventInput[] = [];
+function parseHtml(html: string): EventInput[] {
+  const $ = cheerio.load(html);
+  const events: EventInput[] = [];
 
-    $("article").each((_, el) => {
-      const $el = $(el);
+  $("article").each((_, el) => {
+    const $el = $(el);
 
-      const href = $el.find("a.s_link").attr("href") ?? "";
-      // Date is encoded in the URL: e.g. /schedule/eventname_20260501
-      const dateMatch = href.match(/_(\d{4})(\d{2})(\d{2})$/);
-      if (!dateMatch) return;
-      const date = `${dateMatch[1]}-${dateMatch[2]}-${dateMatch[3]}`;
+    const href = $el.find("a.s_link").attr("href") ?? "";
+    const dateMatch = href.match(/_(\d{4})(\d{2})(\d{2})$/);
+    if (!dateMatch) return;
+    const date = `${dateMatch[1]}-${dateMatch[2]}-${dateMatch[3]}`;
 
-      const h2 = $el.find("h2").first().text().trim();
-      if (!h2) return;
+    const h2 = $el.find("h2").first().text().trim();
+    if (!h2) return;
 
-      const subtitle = $el.find("p.subtitle").first().text().trim();
-      // h2 is the artist/band name; subtitle (if present) is the event title
-      const title = subtitle || h2;
-      const artist = subtitle ? h2 : null;
+    const subtitle = $el.find("p.subtitle").first().text().trim();
+    const title = subtitle || h2;
+    const artist = subtitle ? h2 : null;
 
-      const openTime =
-        $el
-          .find("dl")
-          .filter((_, dl) => $(dl).find("dt").text().includes("OPEN"))
-          .find("dd")
-          .text()
-          .trim()
-          .match(/\d{2}:\d{2}/)?.[0] ?? null;
+    const openTime =
+      $el
+        .find("dl")
+        .filter((_, dl) => $(dl).find("dt").text().includes("OPEN"))
+        .find("dd")
+        .text()
+        .trim()
+        .match(/\d{2}:\d{2}/)?.[0] ?? null;
 
-      const startTime =
-        $el
-          .find("dl")
-          .filter((_, dl) => $(dl).find("dt").text().includes("START"))
-          .find("dd")
-          .text()
-          .trim()
-          .match(/\d{2}:\d{2}/)?.[0] ?? null;
+    const startTime =
+      $el
+        .find("dl")
+        .filter((_, dl) => $(dl).find("dt").text().includes("START"))
+        .find("dd")
+        .text()
+        .trim()
+        .match(/\d{2}:\d{2}/)?.[0] ?? null;
 
-      events.push({
-        venue_id: venue.id,
-        title,
-        artist,
-        date,
-        open_time: openTime,
-        start_time: startTime,
-        image_url: $el.find("div.left img").attr("src") ?? null,
-        source_url: href,
-      });
+    events.push({
+      venue_id: venue.id,
+      title,
+      artist,
+      date,
+      open_time: openTime,
+      start_time: startTime,
+      image_url: $el.find("div.left img").attr("src") ?? null,
+      source_url: href,
     });
+  });
 
-    return events;
+  return events;
+}
+
+export const scraper: Scraper = {
+  venue,
+  async scrape(): Promise<EventInput[]> {
+    const now = new Date();
+    const urls = [0, 1, 2].map((offset) => {
+      const d = new Date(now.getFullYear(), now.getMonth() + offset, 1);
+      return `https://www.liquidroom.net/schedule/${d.getFullYear()}/${String(d.getMonth() + 1).padStart(2, "0")}`;
+    });
+
+    const htmls = await Promise.all(
+      urls.map((url) => fetch(url).then((r) => (r.ok ? r.text() : "")))
+    );
+
+    const seen = new Set<string>();
+    return htmls.flatMap(parseHtml).filter((e) => {
+      if (seen.has(e.source_url ?? e.title)) return false;
+      seen.add(e.source_url ?? e.title);
+      return true;
+    });
   },
 };