feat(indonesian): add Grensia source plugin

plugins/indonesian/grensia.ts

@@ -0,0 +1,277 @@
+import { load as parseHTML } from 'cheerio';
+import { fetchApi } from '@libs/fetch';
+import { Plugin } from '@/types/plugin';
+import { defaultCover } from '@libs/defaultCover';
+
+const BASE = 'https://grensia.blogspot.com';
+const SEARCH_PAGE_SIZE = 150;
+const LIST_LABEL = `${BASE}/search/label/novel?max-results=${SEARCH_PAGE_SIZE}`;
+const MAX_PAGES = 30;
+const LIST_PAGES = 3;
+
+type Post = {
+  title: string;
+  description?: string;
+  url: string;
+  published: string;
+  cover?: string;
+};
+
+type Landing = {
+  name: string;
+  url: string;
+  cover?: string;
+};
+
+const normalize = (text: string) =>
+  text.toLowerCase().replace(/[^a-z0-9]+/g, '');
+
+// Blogger chapter titles look like "Series V4 Chapter 12 Part 3", so the series
+// name is everything before the volume, chapter, part, or epilogue marker.
+const seriesName = (title: string) =>
+  title
+    .replace(/^Novel\s+/i, '')
+    .replace(/\s*\((LN|WN)\)/gi, '')
+    .replace(/\s+(Bahasa\s+)?Indonesia\b.*$/i, '')
+    .replace(
+      /\s+(V\d+|Vol\.?\s*\d+|Volume\s*\d+|Chapter\b|Ch\.?\s*\d|Epilog|Epilogue|Prolog|Prologue|Kata Penutup|Part\s*\d|Side Story|Extra|Bonus|\d+(\.\d+)?$|-\s*\d).*$/i,
+      '',
+    )
+    .replace(/[\s\-–:]+$/, '')
+    .trim();
+
+// The "List Novel" index posts are titled "Novel <name> (LN) Indonesia" etc.
+const landingName = (title: string) =>
+  title
+    .replace(/^Novel\s+/i, '')
+    .replace(/\s*\((LN|WN)\)/gi, '')
+    .replace(/\s+(Bahasa\s+)?Indonesia\b.*$/i, '')
+    .replace(/\s+Bahasa.*$/i, '')
+    .trim();
+
+// Some series headlines use a romanized alias ("Isekai wa Heiwa deshita Chapter 1"),
+// but the structured description repeats the catalog title ("Novel <name> Indonesia").
+const descriptionSeries = (description?: string) => {
+  const match = /^Novel\s+(.+?)\s+Indonesia\b/i.exec(description || '');
+  return match ? match[1].trim() : undefined;
+};
+
+// A post belongs to every series its headline or description names.
+const seriesNames = (post: Post) => {
+  const names = [seriesName(post.title)];
+  const fromDescription = descriptionSeries(post.description);
+  if (fromDescription) names.push(fromDescription);
+  return names;
+};
+
+// Some headlines drop words from the catalog title, but the description begins with it.
+const belongsTo = (post: Post, key: string) => {
+  if (seriesNames(post).some(name => normalize(name) === key)) return true;
+  const description = normalize(decodeEntities(post.description || ''));
+  return description.replace(/^novel/, '').indexOf(key) === 0;
+};
+
+// Some posts point at Blogger's generic "Shoebox" logo (base64 "U2hvZWJveA")
+// instead of artwork, so those are treated as having no cover.
+const usableCover = (url?: string) =>
+  url && url.indexOf('U2hvZWJveA') < 0 ? url : undefined;
+
+// Headlines in the structured data keep HTML entities such as &#9733; for ★.
+const decodeEntities = (text: string) => parseHTML(`<p>${text}</p>`).text();
+
+const chapterNumberOf = (title: string) => {
+  const match = /Chapter\s*(\d+(?:\.\d+)?)/i.exec(title);
+  return match ? Number(match[1]) : undefined;
+};
+
+/**
+ * Reads the post cards on a Blogger listing or search page. Each card carries a
+ * JSON-LD block with the headline, publish date, canonical URL and image.
+ * The "Older Posts" link is followed by the caller to reach every page.
+ */
+async function readPage(url: string) {
+  const body = await (await fetchApi(url)).text();
+  const $ = parseHTML(body);
+  const posts: Post[] = [];
+
+  $("script[type='application/ld+json']").each((i, el) => {
+    try {
+      const data = JSON.parse($(el).html() || '{}');
+      const link = data.mainEntityOfPage && data.mainEntityOfPage['@id'];
+      if (!data.headline || !link) return;
+      posts.push({
+        title: decodeEntities(String(data.headline)).trim(),
+        description: data.description && String(data.description),
+        url: link,
+        published: data.datePublished,
+        cover: usableCover(data.image && data.image.url),
+      });
+    } catch (error) {
+      // Skip malformed structured data rather than failing the whole page.
+    }
+  });
+
+  const older = $('a.blog-pager-older-link').attr('href');
+  return { posts, next: older };
+}
+
+async function readAll(startUrl: string, maxPages: number) {
+  const posts: Post[] = [];
+  let url: string | undefined = startUrl;
+  for (let i = 0; url && i < maxPages; i++) {
+    const page = await readPage(url);
+    posts.push(...page.posts);
+    url = page.next;
+  }
+  return posts;
+}
+
+class GrensiaPlugin implements Plugin.PluginBase {
+  id = 'grensia';
+  name = 'Grensia';
+  icon = 'src/id/grensia/icon.png';
+  site = `${BASE}/`;
+  version = '1.0.0';
+  filters = undefined;
+
+  private landingCache?: Promise<Landing[]>;
+
+  // The site's own "List Novel" index: one post per series.
+  private getLandings() {
+    if (!this.landingCache) {
+      this.landingCache = readAll(LIST_LABEL, LIST_PAGES)
+        .then(posts => {
+          const landings: Landing[] = [];
+          posts.forEach(post => {
+            if (post.url.indexOf('novel-translate-indonesia') >= 0) return;
+            const name = landingName(post.title);
+            if (name) landings.push({ name, url: post.url, cover: post.cover });
+          });
+          return landings;
+        })
+        .catch(error => {
+          // Let the next call retry instead of caching a failed request.
+          this.landingCache = undefined;
+          throw error;
+        });
+    }
+    return this.landingCache;
+  }
+
+  async popularNovels(pageNo: number): Promise<Plugin.NovelItem[]> {
+    if (pageNo > 1) return [];
+
+    const landings = await this.getLandings();
+    const seen: Record<string, boolean> = {};
+    const novels: Plugin.NovelItem[] = [];
+    landings.forEach(landing => {
+      const key = normalize(landing.name);
+      if (seen[key]) return;
+      seen[key] = true;
+      novels.push({
+        name: landing.name,
+        path: encodeURIComponent(landing.name),
+        cover: landing.cover || defaultCover,
+      });
+    });
+    return novels;
+  }
+
+  async parseNovel(novelPath: string): Promise<Plugin.SourceNovel> {
+    const name = decodeURIComponent(novelPath);
+    const landings = await this.getLandings();
+    const landing = landings.find(
+      item => normalize(item.name) === normalize(name),
+    );
+
+    const novel: Plugin.SourceNovel = {
+      path: novelPath,
+      name,
+    };
+
+    // Chapters are every post whose series name matches this novel's name.
+    // The reader's search is used because it returns the full history, unlike
+    // the Atom feed, which truncates long label runs.
+    const posts = await readAll(
+      `${BASE}/search?q=${encodeURIComponent(name)}&max-results=${SEARCH_PAGE_SIZE}`,
+      MAX_PAGES,
+    );
+    const key = normalize(name);
+    // The List Novel posts share the series name but are index pages, not chapters.
+    const indexUrls = landings.map(item => item.url);
+    const matching = posts.filter(
+      post => indexUrls.indexOf(post.url) < 0 && belongsTo(post, key),
+    );
+    // Oldest first; posts published in the same minute keep chapter order.
+    matching.sort(
+      (a, b) =>
+        Date.parse(a.published) - Date.parse(b.published) ||
+        (chapterNumberOf(a.title) || 0) - (chapterNumberOf(b.title) || 0),
+    );
+
+    const withCover = matching.find(post => !!post.cover);
+    novel.cover =
+      (landing && landing.cover) ||
+      (withCover && withCover.cover) ||
+      defaultCover;
+
+    novel.chapters = matching.map(post => ({
+      name: post.title,
+      path: post.url.replace(BASE, ''),
+      releaseTime: new Date(post.published).toISOString(),
+      chapterNumber: chapterNumberOf(post.title),
+    }));
+
+    return novel;
+  }
+
+  async parseChapter(chapterPath: string): Promise<string> {
+    const body = await (await fetchApi(BASE + chapterPath)).text();
+    const $ = parseHTML(body);
+    $(
+      '.post-body script, .post-body style, .post-body ins, .post-body iframe, .post-body a[name="more"]',
+    ).remove();
+    return $('.post-body').first().html() || '';
+  }
+
+  async searchNovels(
+    searchTerm: string,
+    pageNo: number,
+  ): Promise<Plugin.NovelItem[]> {
+    if (pageNo > 1) return [];
+
+    const posts = await readAll(
+      `${BASE}/search?q=${encodeURIComponent(searchTerm)}&max-results=${SEARCH_PAGE_SIZE}`,
+      1,
+    );
+    const landings = await this.getLandings();
+    const term = normalize(searchTerm);
+    const seen: Record<string, boolean> = {};
+    const novels: Plugin.NovelItem[] = [];
+
+    const indexUrls = landings.map(item => item.url);
+    posts.forEach(post => {
+      if (indexUrls.indexOf(post.url) >= 0) return;
+      seriesNames(post).forEach(name => {
+        const key = normalize(name);
+        if (!key || seen[key] || key.indexOf(term) < 0) return;
+        seen[key] = true;
+
+        // Prefer the series' own index cover, then the first chapter's image.
+        const landing = landings.find(item => normalize(item.name) === key);
+        novels.push({
+          name,
+          path: encodeURIComponent(name),
+          cover: (landing && landing.cover) || post.cover || defaultCover,
+        });
+      });
+    });
+
+    return novels;
+  }
+
+  resolveUrl = (path: string, isNovel?: boolean) =>
+    isNovel ? `${BASE}/search?q=${path}` : BASE + path;
+}
+
+export default new GrensiaPlugin();