From e30cbac17fa65fc84ad889498ba99680cb1b84da Mon Sep 17 00:00:00 2001
From: Bin Wang <bin.wang@mail.binwang.me>
Date: Sun, 17 Sep 2023 16:01:48 -0400
Subject: [PATCH] fix(route/cbc): Fix article content html tag (#13325)

* fix(route/cbc): Fix article content html tag

* refactor: migrate to v2

* fix: parse first jsonld only

---------
---
 lib/router.js                            |  2 +-
 lib/v2/cbc/maintainer.js                 |  3 ++
 lib/v2/cbc/radar.js                      | 13 +++++++++
 lib/v2/cbc/router.js                     |  3 ++
 lib/{routes => v2}/cbc/topics.js         | 37 +++++++-----------------
 website/docs/routes/traditional-media.md |  6 ++--
 6 files changed, 34 insertions(+), 30 deletions(-)
 create mode 100644 lib/v2/cbc/maintainer.js
 create mode 100644 lib/v2/cbc/radar.js
 create mode 100644 lib/v2/cbc/router.js
 rename lib/{routes => v2}/cbc/topics.js (58%)

diff --git a/lib/router.js b/lib/router.js
index 17bc34476..50ee32839 100644
--- a/lib/router.js
+++ b/lib/router.js
@@ -1443,7 +1443,7 @@ router.get('/xinquji/today/internal', lazyloadRouteHandler('./routes/xinquji/int
 router.get('/gbcc/trust', lazyloadRouteHandler('./routes/gbcc/trust'));
 
 // CBC
-router.get('/cbc/topics/:topic?', lazyloadRouteHandler('./routes/cbc/topics'));
+// router.get('/cbc/topics/:topic?', lazyloadRouteHandler('./routes/cbc/topics'));
 
 // discuz
 // router.get('/discuz/:ver([7|x])/:cid([0-9]{2})/:link(.*)', lazyloadRouteHandler('./routes/discuz/discuz'));
diff --git a/lib/v2/cbc/maintainer.js b/lib/v2/cbc/maintainer.js
new file mode 100644
index 000000000..363672ee3
--- /dev/null
+++ b/lib/v2/cbc/maintainer.js
@@ -0,0 +1,3 @@
+module.exports = {
+    '/topics/:topic?': ['wb14123'],
+};
diff --git a/lib/v2/cbc/radar.js b/lib/v2/cbc/radar.js
new file mode 100644
index 000000000..34d9253a7
--- /dev/null
+++ b/lib/v2/cbc/radar.js
@@ -0,0 +1,13 @@
+module.exports = {
+    'cbc.ca': {
+        _name: 'Canadian Broadcasting Corporation',
+        '.': [
+            {
+                title: 'News',
+                docs: 'https://docs.rsshub.app/routes/traditional-media#canadian-broadcasting-corporation',
+                source: ['/news'],
+                target: '/cbc/topics',
+            },
+        ],
+    },
+};
diff --git a/lib/v2/cbc/router.js b/lib/v2/cbc/router.js
new file mode 100644
index 000000000..6d2fc6901
--- /dev/null
+++ b/lib/v2/cbc/router.js
@@ -0,0 +1,3 @@
+module.exports = (router) => {
+    router.get('/topics/:topic?', require('./topics'));
+};
diff --git a/lib/routes/cbc/topics.js b/lib/v2/cbc/topics.js
similarity index 58%
rename from lib/routes/cbc/topics.js
rename to lib/v2/cbc/topics.js
index 85701bdb0..df5bbf449 100644
--- a/lib/routes/cbc/topics.js
+++ b/lib/v2/cbc/topics.js
@@ -4,12 +4,9 @@ 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 url = `${baseUrl}/news${topic ? `/${topic.replace('-', '/')}` : ''}`;
 
-    const response = await got({
-        method: 'get',
-        url,
-    });
+    const response = await got(url);
 
     const data = response.data;
 
@@ -27,13 +24,13 @@ module.exports = async (ctx) => {
     $('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);
+        links.map((link) =>
+            ctx.cache.tryGet(link, async () => {
+                const result = await got(link);
 
                 const $ = cheerio.load(result.data);
 
-                const head = JSON.parse($('script[type="application/ld+json"]').html());
+                const head = JSON.parse($('script[type="application/ld+json"]').first().text());
                 if (!head) {
                     return [];
                 }
@@ -44,28 +41,16 @@ module.exports = async (ctx) => {
                     author = head.author.map((author) => author.name).join(' & ');
                 }
                 const pubDate = head.datePublished;
-                const description = $('div.storyWrapper').html();
+                const description = $('div[data-cy=storyWrapper]').html();
 
-                return [title, author, pubDate, description];
-            });
-
-            if (!title) {
-                return Promise.resolve(undefined);
-            }
-
-            return {
-                title,
-                description,
-                pubDate,
-                link,
-                author,
-            };
-        })
+                return { title, author, pubDate, description, link };
+            })
+        )
     );
 
     ctx.state.data = {
         title: $('title').text(),
         link: url,
-        item: out.filter((x) => x),
+        item: out.filter((x) => x.title),
     };
 };
diff --git a/website/docs/routes/traditional-media.md b/website/docs/routes/traditional-media.md
index 493207517..85996ff8a 100644
--- a/website/docs/routes/traditional-media.md
+++ b/website/docs/routes/traditional-media.md
@@ -163,11 +163,11 @@ Refer to [Boston.com's feed page](https://www.boston.com/rss-feeds) for tags. Fo
 
 </Route>
 
-## CBC {#cbc}
+## Canadian Broadcasting Corporation {#canadian-broadcasting-corporation}
 
-通过提取文章全文，以提供比官方源更佳的阅读体验。
+### News {#canadian-broadcasting-corporation-news}
 
-<Route author="wb14123" 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 `/`。']}/>
+<Route author="wb14123" example="/cbc/topics" path="/cbc/topics/:topic?" paramsDesc={['Channel,`Top Stories` by default. For secondary channel like `canada/toronto`, use `-` to replace `/`']} radar="1"/>
 
 ## Chicago Tribune {#chicago-tribune}