feat: add CBC (#5179)

docs/en/traditional-media.md

@@ -37,6 +37,12 @@ Refer to [Boston.com's feed page](https://www.boston.com/rss-feeds) for tags. Fo
 
 </RouteEn>
 
+## CBC
+
+Provide full article RSS for CBC topics.
+
+<RouteEn author="wangbin" example="/cbc/topics" path="/cbc/topics/:topic?" :paramsDesc="['CBC topics. Default to Top Stories. For second level topics like canada/toronto, need to replace `/` by `-`.']"/>
+
 ## Chicago Tribune
 
 ### News

docs/traditional-media.md

@@ -43,6 +43,12 @@ pageClass: routes
 
 </Route>
 
+## CBC
+
+通过提取文章全文，以提供比官方源更佳的阅读体验。
+
+<Route author="wangbin" example="/cbc/topics" path="/cbc/topics/:topic?" :paramsDesc="['CBC 频道。默认为 Top Stories。二级话题如 canada/toronto，需要用 `-` 替换掉 `/`。']"/>
+
 ## Chicago Tribune
 
 ### 新闻

lib/router.js

@@ -2363,6 +2363,9 @@ router.get('/gbcc/trust', require('./routes/gbcc/trust'));
 // Associated Press
 router.get('/apnews/topics/:topic', require('./routes/apnews/topics'));
 
+// CBC
+router.get('/cbc/topics/:topic?', require('./routes/cbc/topics'));
+
 // discuz
 router.get('/discuz/:ver([7|x])/:cid([0-9]{2})/:link(.*)', require('./routes/discuz/discuz'));
 router.get('/discuz/:ver([7|x])/:link(.*)', require('./routes/discuz/discuz'));

lib/routes/cbc/topics.js

@@ -0,0 +1,72 @@
+const got = require('@/utils/got');
+const cheerio = require('cheerio');
+
+module.exports = async (ctx) => {
+    const baseUrl = 'https://www.cbc.ca';
+    const topic = ctx.params.topic || '';
+    const url = `${baseUrl}/news/${topic.replace('-', '/')}`;
+
+    const response = await got({
+        method: 'get',
+        url: url,
+    });
+
+    const data = response.data;
+
+    const $ = cheerio.load(data);
+    const links = [];
+
+    function pushLinks(index, item) {
+        const link = item.attribs.href;
+        if (link.startsWith('/')) {
+            links.push(baseUrl + link);
+        }
+    }
+
+    $('a.contentWrapper').each(pushLinks);
+    $('a.card').each(pushLinks);
+
+    const out = await Promise.all(
+        links.map(async (link) => {
+            const [title, author, pubDate, description] = await ctx.cache.tryGet(link, async () => {
+                const result = await got.get(link);
+
+                const $ = cheerio.load(result.data);
+
+                const head = JSON.parse($('script[type="application/ld+json"]').html());
+                if (!head) {
+                    return [];
+                }
+
+                const title = head.headline;
+                let author = '';
+                if (head.author) {
+                    author = head.author.map((author) => author.name).join(' & ');
+                }
+                const pubDate = head.datePublished;
+                const description = $('div.storyWrapper').html();
+
+                return [title, author, pubDate, description];
+            });
+
+            if (!title) {
+                return Promise.resolve(undefined);
+            }
+
+            const item = {
+                title: title,
+                description: description,
+                pubDate: pubDate,
+                link: link,
+                author: author,
+            };
+            return Promise.resolve(item);
+        })
+    );
+
+    ctx.state.data = {
+        title: $('title').text(),
+        link: url,
+        item: out.filter((x) => x),
+    };
+};