/**
 * 東高円寺 U.F.O.CLUB — https://www.ufoclub.jp/schedule/
 *
 * WordPress テーマ。fetch に Mozilla UA が必要（curl は 2 バイトを返す）。
 *
 * DOM構造:
 *   <div class="month">2026年6月</div>   ← 年月の取得元
 *   <table class="schedule-table"><tbody>
 *     <tr>
 *       <th>1</th>            ← 日
 *       <td>MON</td>          ← 曜日（スキップ）
 *       <td>
 *         <strong>【タイトル】</strong><br/>
 *         /LIVE; artist1, artist2
 *         /DJ; dj1
 *       </td>
 *       <td class="td_open">OPEN 18:30~<br/>START 19:00~</td>
 *       <td class="td_charge">ADV.¥2800(+1D)<br/>DOOR.¥3300(+1D)</td>
 *     </tr>
 *   </tbody></table>
 *
 * 月ナビ: /schedule/ (当月) / /schedule/next-month (翌月)
 */
import * as cheerio from "cheerio";
import type { Scraper, VenueMeta } from "./base";
import type { EventInput } from "~/lib/db.server";

export const venue: VenueMeta = {
  id: "ufoclub-higashikoenji",
  name: "東高円寺 U.F.O.CLUB",
  url: "https://www.ufoclub.jp",
  area: "東高円寺",
  capacity: 120,
};

const UA =
  "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/120.0.0.0 Safari/537.36";

async function scrapePage(url: string): Promise<EventInput[]> {
  const res = await fetch(url, { headers: { "User-Agent": UA } });
  if (!res.ok) throw new Error(`HTTP ${res.status} for ${url}`);
  const $ = cheerio.load(await res.text());
  const events: EventInput[] = [];

  // Parse year and month from <div class="month">2026年6月</div>
  const monthText = $("div.month").first().text().trim();
  const monthMatch = monthText.match(/(\d{4})年(\d{1,2})月/);
  if (!monthMatch) return events;
  const year = parseInt(monthMatch[1], 10);
  const month = parseInt(monthMatch[2], 10);

  $("table.schedule-table tbody tr").each((_, row) => {
    const $row = $(row);

    // Day from <th>
    const day = parseInt($row.find("th").text().trim(), 10);
    if (isNaN(day)) return;
    const date = `${year}-${String(month).padStart(2, "0")}-${String(day).padStart(2, "0")}`;

    // Event td is the 2nd <td> (index 1; 0 = day-of-week)
    const $eventTd = $row.find("td").eq(1);
    const title = $eventTd.find("strong").text().replace(/^【|】$/g, "").trim();
    if (!title) return;

    // Artists: text after </strong>, lines starting with /LIVE; /DJ; /MC;
    const $tdClone = $eventTd.clone();
    $tdClone.find("strong").remove();
    $tdClone.find("br").replaceWith("\n");
    const afterTitle = $tdClone.text();
    const artistParts: string[] = [];
    for (const line of afterTitle.split("\n")) {
      const t = line.replace(/\s+/g, " ").trim();
      const liveM = t.match(/^\/LIVE;\s*(.+)/i);
      const djM = t.match(/^\/DJ;\s*(.+)/i);
      const mcM = t.match(/^\/MC;\s*(.+)/i);
      if (liveM) artistParts.push(liveM[1].trim());
      else if (djM) artistParts.push(`DJ: ${djM[1].trim()}`);
      else if (mcM) artistParts.push(`MC: ${mcM[1].trim()}`);
    }
    const artist = artistParts.length > 0 ? artistParts.join(" / ") : null;

    // Open / Start time from td.td_open
    const $openTd = $row.find("td.td_open").clone();
    $openTd.find("br").replaceWith("\n");
    const openText = $openTd.text();
    const openMatch = openText.match(/OPEN\s+(\d{1,2}:\d{2})/i);
    const startMatch = openText.match(/START\s+(\d{1,2}:\d{2})/i);

    // Price from td.td_charge (first two lines: ADV / DOOR)
    const $chargeTd = $row.find("td.td_charge").clone();
    $chargeTd.find("br").replaceWith("\n");
    const chargeLines = $chargeTd
      .text()
      .split("\n")
      .map((l) => l.trim())
      .filter(Boolean);
    const price =
      chargeLines.length >= 2
        ? `${chargeLines[0]} / ${chargeLines[1]}`
        : (chargeLines[0] ?? null);

    events.push({
      venue_id: venue.id,
      title,
      artist,
      date,
      open_time: openMatch?.[1] ?? null,
      start_time: startMatch?.[1] ?? null,
      price,
      source_url: url,
    });
  });

  return events;
}

export const scraper: Scraper = {
  venue,
  async scrape(): Promise<EventInput[]> {
    const [thisMonth, nextMonth] = await Promise.all([
      scrapePage("https://www.ufoclub.jp/schedule/"),
      scrapePage("https://www.ufoclub.jp/schedule/next-month"),
    ]);

    const seen = new Set<string>();
    return [...thisMonth, ...nextMonth].filter((e) => {
      const key = `${e.date}|${e.title}`;
      if (seen.has(key)) return false;
      seen.add(key);
      return true;
    });
  },
};