From 31b56187c0f349e7232fb97befef95d9af06cb43 Mon Sep 17 00:00:00 2001
From: Andvari <31068367+dzx-dzx@users.noreply.github.com>
Date: Tue, 5 Sep 2023 23:57:42 +0800
Subject: [PATCH] feat(route): use API for reuters. (#13190)

* feat(route): use reuters api.

* Fix typo in doc.

* Refining

* Update common.js

* Update common.js

* Update common.js

* Update common.js

* Update common.js

* Refactor the code, and apologize for my stubbornness.
---
 lib/v2/reuters/common.js                 | 75 ++++++++++++++++--------
 website/docs/routes/traditional-media.md |  6 +-
 2 files changed, 52 insertions(+), 29 deletions(-)

diff --git a/lib/v2/reuters/common.js b/lib/v2/reuters/common.js
index f72250926c..2bfa07bc23 100644
--- a/lib/v2/reuters/common.js
+++ b/lib/v2/reuters/common.js
@@ -9,31 +9,54 @@ module.exports = async (ctx) => {
     const topic = ctx.params.topic ?? (category === 'authors' ? 'reuters' : '');
     const limit = ctx.query.limit ? parseInt(ctx.query.limit) : 20;
 
-    const rootUrl = 'https://www.reuters.com';
-    const currentUrl = topic ? `${rootUrl}/${category}/${topic}/` : `${rootUrl}/${category}/`;
-    const response = await got(currentUrl);
-    const $ = cheerio.load(response.data);
-
-    let items = $('.media-story-card__body__3tRWy a.media-story-card__heading__eqhp9, a.svelte-pxbp38, a.svelte-11dknnx, a.svelte-e21rsn')
-        .slice(0, limit)
-        .toArray()
-        .map((item) => {
-            item = $(item);
-            item.find('span.visually-hidden__hidden__2qXMW').remove();
+    const MUST_FETCH_BY_TOPICS = ['authors'];
+    const section_id = `/${category}/${topic ? `${topic}/` : ''}`;
+    const { title, description, rootUrl, response } = await (async () => {
+        if (!MUST_FETCH_BY_TOPICS.includes(category)) {
+            const rootUrl = 'https://www.reuters.com/pf/api/v3/content/fetch/articles-by-section-alias-or-id-v1';
+            const response = await got(rootUrl, {
+                searchParams: {
+                    query: JSON.stringify({
+                        offset: 0,
+                        size: limit,
+                        section_id,
+                        website: 'reuters',
+                    }),
+                },
+            }).json();
             return {
-                title: item.text(),
-                link: new URL(item.prop('href'), rootUrl).href,
+                title: response.result.section.title,
+                description: response.result.section.section_about,
+                rootUrl,
+                response,
             };
-        });
-    if (!items.length) {
-        const metadata = $('script#fusion-metadata').html();
-        const metadataObj = JSON.parse(metadata.match(/Fusion.globalContent=(\{[\s\S]*?});/)[1]);
-        const articles = metadataObj.arcResult?.articles ?? metadataObj.result?.articles ?? [];
-        items = articles.map((article) => ({
-            title: article.title,
-            link: rootUrl + article.canonical_url,
-        }));
-    }
+        } else {
+            const rootUrl = 'https://www.reuters.com/pf/api/v3/content/fetch/articles-by-topic-v1';
+            const response = await got(rootUrl, {
+                searchParams: {
+                    query: JSON.stringify({
+                        offset: 0,
+                        size: limit,
+                        topic_url: section_id,
+                        website: 'reuters',
+                    }),
+                },
+            }).json();
+
+            return {
+                title: `${response.result.topics[0].name} | Reuters`,
+                description: response.result.topics[0].entity_id,
+                rootUrl,
+                response,
+            };
+        }
+    })();
+
+    let items = response.result.articles.map((e) => ({
+        title: e.title,
+        link: new URL(e.canonical_url, rootUrl).href,
+    }));
+
     items = await Promise.all(
         items.map((item) =>
             ctx.cache.tryGet(item.link, async () => {
@@ -88,10 +111,10 @@ module.exports = async (ctx) => {
     );
 
     ctx.state.data = {
-        title: $('head title').text(),
-        description: $('head meta[name=description]').attr('content'),
+        title,
+        description,
         image: 'https://www.reuters.com/pf/resources/images/reuters/logo-vertical-default-512x512.png?d=116',
-        link: currentUrl,
+        link: `https://www.reuters.com${section_id}`,
         item: items,
     };
 };
diff --git a/website/docs/routes/traditional-media.md b/website/docs/routes/traditional-media.md
index be1c56501a..089a1ba580 100644
--- a/website/docs/routes/traditional-media.md
+++ b/website/docs/routes/traditional-media.md
@@ -544,13 +544,13 @@ Parameters can be obtained from the official website, for instance:
 
     | All | Aerospace & Defense | Autos & Transportation | Energy | Environment | Finance | Healthcare & Pharmaceuticals | Media & Telecom | Retail & Consumer | Sustainable Business | Charged | Future of Health | Future of Money | Take Five | Reuters Impact |
     | --- | ------------------- | ---------------------- | ------ | ----------- | ------- | ---------------------------- | --------------- | ----------------- | -------------------- | ------- | ---------------- | --------------- | --------- | -------------- |
-    |     | aerospace-defense   | autos-transportation   | energy | environment | finance | healthcare-pharmaceuticals   | media-telecom   | retail-consumer   | sustainable-business | charged | future-of-health | futrue-of-money | take-five | reuters-impact |
+    |     | aerospace-defense   | autos-transportation   | energy | environment | finance | healthcare-pharmaceuticals   | media-telecom   | retail-consumer   | sustainable-business | charged | future-of-health | future-of-money | take-five | reuters-impact |
 
 -   `legal/:topic`:
 
-    | All | Goverment | Legal Industry | Litigation | Transaction |
+    | All | Government | Legal Industry | Litigation | Transactional |
     | --- | --------- | -------------- | ---------- | ----------- |
-    |     | goverment | legalindustry  | litigation | transaction |
+    |     | government | legalindustry  | litigation | transactional |
 
 -   `authors/:topic`: