1 files changed, 82 insertions, 0 deletions
diff --git a/app/scrapers/shibuya-o.ts b/app/scrapers/shibuya-o.ts
new file mode 100644
index 0000000..1ad8d8c
--- /dev/null
+++ b/app/scrapers/shibuya-o.ts
@@ -0,0 +1,82 @@
+/**
+ * Shibuya O-East / O-West / O-Crest / O-Nest (渋谷)
+ * https://www.shibuya-o.com/schedule/
+ *
+ * The page uses a unified schedule listing for all O venues.
+ */
+import * as cheerio from "cheerio";
+import type { Scraper, VenueMeta } from "./base";
+import type { EventInput } from "~/lib/db.server";
+
+export const venue: VenueMeta = {
+  id: "shibuya-o",
+  name: "渋谷 O-EAST / O-WEST",
+  url: "https://www.shibuya-o.com",
+  area: "渋谷",
+};
+
+export const scraper: Scraper = {
+  venue,
+  async scrape(): Promise<EventInput[]> {
+    const res = await fetch("https://www.shibuya-o.com/schedule/");
+    if (!res.ok) throw new Error(`HTTP ${res.status}`);
+    const html = await res.text();
+    const $ = cheerio.load(html);
+    const events: EventInput[] = [];
+
+    $(".schedule_list li, .c-schedule__item, .event-item").each((_, el) => {
+      const $el = $(el);
+
+      const title = $el.find(".schedule_title, .title, h3").first().text().trim();
+      if (!title) return;
+
+      const rawDate =
+        $el.find(".schedule_date, .date, time").first().text().trim() ||
+        $el.find("time").attr("datetime") ||
+        "";
+      const date = parseJapaneseDate(rawDate);
+      if (!date) return;
+
+      const hall = $el.find(".schedule_hall, .hall, .venue-name").first().text().trim() || null;
+      const timeText = $el.find(".schedule_time, .time").first().text();
+      const openMatch = timeText.match(/OPEN[：: ]*(\d{2}:\d{2})/i);
+      const startMatch = timeText.match(/START[：: ]*(\d{2}:\d{2})/i);
+
+      const detailHref = $el.find("a[href]").first().attr("href") ?? null;
+
+      events.push({
+        venue_id: venue.id,
+        title,
+        artist: hall,
+        date,
+        open_time: openMatch?.[1] ?? null,
+        start_time: startMatch?.[1] ?? null,
+        ticket_url:
+          $el.find("a[href*='eplus'], a[href*='lawson'], a[href*='ticket']").first().attr("href") ?? null,
+        image_url: $el.find("img").first().attr("src")
+          ? absoluteUrl($el.find("img").first().attr("src")!, venue.url)
+          : null,
+        source_url: detailHref ? absoluteUrl(detailHref, venue.url) : null,
+      });
+    });
+
+    return events;
+  },
+};
+
+function parseJapaneseDate(raw: string): string | null {
+  const m =
+    raw.match(/(\d{4})[./年](\d{1,2})[./月](\d{1,2})/) ||
+    raw.match(/(\d{1,2})[./月](\d{1,2})/);
+  if (!m) return null;
+  if (m.length === 4) {
+    return `${m[1]}-${m[2].padStart(2, "0")}-${m[3].padStart(2, "0")}`;
+  }
+  const year = new Date().getFullYear();
+  return `${year}-${m[1].padStart(2, "0")}-${m[2].padStart(2, "0")}`;
+}
+
+function absoluteUrl(url: string, base: string): string {
+  if (url.startsWith("http")) return url;
+  return url.startsWith("/") ? base + url : `${base}/${url}`;
+}