fix(route): mingpao more fancybox (#17209)

DIYgod · Oct 21, 2024 · 5b0ec7d · 5b0ec7d
1 parent 195327b
commit 5b0ec7d
Show file tree

Hide file tree

Showing 2 changed files with 98 additions and 37 deletions.
diff --git a/lib/routes/mingpao/index.ts b/lib/routes/mingpao/index.ts
@@ -3,13 +3,12 @@ import { getCurrentPath } from '@/utils/helpers';
 const __dirname = getCurrentPath(import.meta.url);
 
 import cache from '@/utils/cache';
-import got from '@/utils/got';
+import ofetch from '@/utils/ofetch';
 import parser from '@/utils/rss-parser';
-import { load } from 'cheerio';
+import * as cheerio from 'cheerio';
 import { parseDate } from '@/utils/parse-date';
 import { art } from '@/utils/render';
 import path from 'node:path';
-import logger from '@/utils/logger';
 
 const renderFanBox = (media) =>
     art(path.join(__dirname, 'templates/fancybox.art'), {
@@ -22,11 +21,79 @@ const renderDesc = (media, desc) =>
         desc,
     });
 
+const fixFancybox = (element, $) => {
+    const $e = $(element);
+    const url = new URL($e.attr('href'));
+    let video;
+    if (url.hostname === 'videop.mingpao.com') {
+        video = new URL(url.searchParams.get('file'));
+        video.hostname = 'cfrvideo.mingpao.com'; // use cloudflare cdn
+        video = video.href;
+    }
+    return {
+        href: url.href,
+        title: $e.attr('title'),
+        video,
+    };
+};
+
 export const route: Route = {
     path: '/:type?/:category?',
-    name: 'Unknown',
+    name: '新聞',
+    example: '/mingpao/ins/all',
+    parameters: {
+        type: {
+            description: '新聞類型',
+            default: 'ins',
+            options: [
+                { value: 'ins', label: '即時新聞' },
+                { value: 'pns', label: '每日明報' },
+            ],
+        },
+        category: '頻道，見下表',
+    },
+    radar: [
+        {
+            title: '即時新聞',
+            source: ['news.mingpao.com/ins/:categoryName/section/:date/:category'],
+            target: '/mingpao/ins/:category',
+        },
+        {
+            title: '每日明報',
+            source: ['news.mingpao.com/pns/:categoryName/section/:date/:category'],
+            target: '/mingpao/pns/:category',
+        },
+    ],
     maintainers: ['TonyRL'],
     handler,
+    description: `| category | 即時新聞頻道 |
+| -------- | ------------ |
+| all      | 總目錄       |
+| s00001   | 港聞         |
+| s00002   | 經濟         |
+| s00003   | 地產         |
+| s00004   | 兩岸         |
+| s00005   | 國際         |
+| s00006   | 體育         |
+| s00007   | 娛樂         |
+| s00022   | 文摘         |
+| s00024   | 熱點         |
+
+| category | 每日明報頻道 |
+| -------- | ------------ |
+| s00001   | 要聞         |
+| s00002   | 港聞         |
+| s00003   | 社評         |
+| s00004   | 經濟         |
+| s00005   | 副刊         |
+| s00011   | 教育         |
+| s00012   | 觀點         |
+| s00013   | 中國         |
+| s00014   | 國際         |
+| s00015   | 體育         |
+| s00016   | 娛樂         |
+| s00017   | English      |
+| s00018   | 作家專欄     |`,
 };
 
 async function handler(ctx) {
@@ -35,26 +102,16 @@ async function handler(ctx) {
     const link = `https://news.mingpao.com/rss/${type}/${category}.xml`;
 
     const feed = await parser.parseURL(link);
-
     const items = await Promise.all(
         feed.items.map((item) =>
             cache.tryGet(item.link, async () => {
-                let response;
-                try {
-                    response = await got(item.link, {
-                        headers: {
-                            Referer: 'https://news.mingpao.com/',
-                        },
-                    });
-                } catch (error) {
-                    if (error instanceof got.MaxRedirectsError) {
-                        logger.error(`MaxRedirectsError when requesting ${decodeURIComponent(item.link)}`);
-                        return item;
-                    }
-                    throw error;
-                }
+                const response = await ofetch(item.link, {
+                    headers: {
+                        Referer: 'https://news.mingpao.com/',
+                    },
+                });
 
-                const $ = load(response.data);
+                const $ = cheerio.load(response);
                 const fancyboxImg = $('a.fancybox').length ? $('a.fancybox') : $('a.fancybox-buttons');
 
                 // remove unwanted elements
@@ -68,21 +125,25 @@ async function handler(ctx) {
                 item.category = item.categories;
 
                 // fix fancybox image
-                const fancybox = fancyboxImg.toArray().map((e) => {
-                    e = $(e);
-                    const href = new URL(e.attr('href'));
-                    let video;
-                    if (href.hostname === 'videop.mingpao.com') {
-                        video = new URL(href.searchParams.get('file'));
-                        video.hostname = 'cfrvideo.mingpao.com'; // use cloudflare cdn
-                        video = video.href;
-                    }
-                    return {
-                        href: href.href,
-                        title: e.attr('title'),
-                        video,
-                    };
-                });
+                let fancybox = fancyboxImg.toArray().map((e) => fixFancybox(e, $));
+                const script = $('script')
+                    .toArray()
+                    .find((e) => $(e).text()?.includes("$('#lower').prepend('"));
+                const lowerContent = script
+                    ? $(script)
+                          .text()
+                          ?.match(/\$\('#lower'\)\.prepend\('(.*)'\);/)?.[1]
+                          ?.replaceAll(/\\"/g, '"')
+                    : '';
+                if (lowerContent) {
+                    const $ = cheerio.load(lowerContent, null, false);
+                    fancybox = [
+                        ...fancybox,
+                        ...$('a.fancybox')
+                            .toArray()
+                            .map((e) => fixFancybox(e, $)),
+                    ];
+                }
 
                 // remove unwanted key value
                 delete item.categories;
@@ -91,7 +152,7 @@ async function handler(ctx) {
                 delete item.creator;
                 delete item.isoDate;
 
-                item.description = renderDesc(fancybox, $('.txt4').html() ?? $('.article_content.line_1_5em').html());
+                item.description = renderDesc(fancybox, $('.txt4').html() ?? $('.article_content.line_1_5em').html() ?? $('.txt3').html());
                 item.pubDate = parseDate(item.pubDate);
                 item.guid = item.link.includes('?') ? item.link : item.link.substring(0, item.link.lastIndexOf('/'));
 

diff --git a/lib/routes/mingpao/namespace.ts b/lib/routes/mingpao/namespace.ts
@@ -1,6 +1,6 @@
 import type { Namespace } from '@/types';
 
 export const namespace: Namespace = {
-    name: 'Ming Pao 明报',
+    name: '明報',
     url: 'mingpao.com',
 };