From 532ef65992c445ad25a68293af4da8b2000a5b86 Mon Sep 17 00:00:00 2001
From: minty_frankie <77310871+mintyfrankie@users.noreply.github.com>
Date: Sun, 2 Jun 2024 09:59:23 +0200
Subject: [PATCH] feat(route): add route SustainabilityMagazine articles
 (#15666)

* feat(susmag): add namespace

* fix(sustamag): update namespace

* feat(sustamag): add articles route

* fix(sustamag): fix bugs

- fix items' title parsing
- fix items' link parsing

* feat(sustamag): add full-text support

* fix(doc): set requirePuppeteer

* refactor(route): remove puppeteer requirement

* refactor(route): change article list fectching method to GraphQL

* fix: apply suggestions from code review

Co-authored-by: Tony <TonyRL@users.noreply.github.com>

* refactor: change article items fetching selectors

* fix: article description

---------
---
 lib/routes/sustainabilitymag/articles.ts  | 144 ++++++++++++++++++++++
 lib/routes/sustainabilitymag/namespace.ts |   6 +
 2 files changed, 150 insertions(+)
 create mode 100644 lib/routes/sustainabilitymag/articles.ts
 create mode 100644 lib/routes/sustainabilitymag/namespace.ts

diff --git a/lib/routes/sustainabilitymag/articles.ts b/lib/routes/sustainabilitymag/articles.ts
new file mode 100644
index 000000000..8be6ed653
--- /dev/null
+++ b/lib/routes/sustainabilitymag/articles.ts
@@ -0,0 +1,144 @@
+import { Route } from '@/types';
+import cache from '@/utils/cache';
+import oftech from '@/utils/ofetch';
+import { parseDate } from '@/utils/parse-date';
+import { load } from 'cheerio';
+
+export const route: Route = {
+    path: '/articles',
+    name: 'Articles',
+    url: 'sustainabilitymag.com/articles',
+    maintainers: ['mintyfrankie'],
+    categories: ['other'],
+    example: '/sustainabilitymag/articles',
+    radar: [
+        {
+            source: ['https://sustainabilitymag.com/articles'],
+            target: '/sustainabilitymag/articles',
+        },
+    ],
+    features: {
+        requireConfig: false,
+        requirePuppeteer: false,
+        antiCrawler: false,
+        supportBT: false,
+        supportPodcast: false,
+        supportScihub: false,
+    },
+    handler,
+};
+
+const findLargestImgKey = (images) =>
+    Object.keys(images)
+        .filter((key) => key.startsWith('inline_free_') || key.startsWith('hero_landscape_'))
+        .sort((a, b) => Number.parseInt(b.split('_')[2]) - Number.parseInt(a.split('_')[2]))[0];
+
+const renderFigure = (url, caption) => `<figure><img src="${url}" alt="${caption}" /><figcaption>${caption}</figcaption></figure>`;
+
+const render = (widgets) =>
+    widgets
+        .map((w) => {
+            switch (w.type) {
+                case 'text':
+                    return w.html;
+                case 'keyFacts':
+                    return `<div><ul>${w.keyFacts.map((k) => `<li>${k.text}</li>`).join('')}</ul></div>`;
+                case 'inlineVideo':
+                    return w.provider === 'youtube'
+                        ? `<iframe id="ytplayer" type="text/html" width="640" height="360" src="https://www.youtube-nocookie.com/embed/${w.videoId}" frameborder="0" allowfullscreen></iframe>`
+                        : new Error(`Unhandled inlineVideo provider: ${w.provider}`);
+                case 'inlineImage':
+                    return w.inlineImageImages
+                        .map((image) => {
+                            const i = image.images[findLargestImgKey(image.images)][0];
+                            return renderFigure(i.url, i.caption);
+                        })
+                        .join('');
+                default:
+                    throw new Error(`Unhandled widget type: ${w.type}`);
+            }
+        })
+        .join('');
+
+async function handler() {
+    const baseURL = `https://sustainabilitymag.com`;
+    const feedURL = `${baseURL}/articles`;
+    const feedLang = 'en';
+    const feedDescription = 'Sustainability Magazine Articles';
+
+    const requestEndpoint = `${baseURL}/graphql`;
+    const requestBody = JSON.stringify({
+        query: `query PaginatedQuery($url: String!, $page: Int = 1, $widgetType: String!) {
+          paginatedWidget(url: $url, widgetType: $widgetType) {
+            ... on SimpleArticleGridWidget {
+              articles(page: $page) {
+                results {
+                  _id
+                  headline
+                  fullUrlPath
+                  featured
+                  category
+                  contentType
+                  tags {
+                    tag
+                  }
+                  attribution
+                  subAttribution
+                  sell
+                  images {
+                    thumbnail_widescreen_553 {
+                      url
+                    }
+                  }
+                }
+              }
+            }
+          }
+        }`,
+        operationName: 'PaginatedQuery',
+        variables: {
+            widgetType: 'simpleArticleGrid',
+            page: 1,
+            url: 'https://sustainabilitymag.com/articles',
+        },
+    });
+
+    const results = await oftech(requestEndpoint, {
+        method: 'POST',
+        headers: { 'Content-Type': 'application/json' },
+        body: requestBody,
+    });
+
+    const list = results.data.paginatedWidget.articles.results.map((item) => ({
+        title: item.headline,
+        link: `${baseURL}${item.fullUrlPath}`,
+        image: item.images?.thumbnail_widescreen_553?.url,
+        category: item.category,
+    }));
+
+    const items = await Promise.all(
+        list.map((item) =>
+            cache.tryGet(item.link, async () => {
+                const response = await oftech(item.link);
+                const $ = load(response);
+
+                const nextData = JSON.parse($('script#__NEXT_DATA__').text());
+                const article = nextData.props.pageProps.article;
+                item.pubDate = parseDate(article.displayDate);
+                item.author = article.author.name;
+                const heroImage = article.images[findLargestImgKey(article.images)][0];
+                item.description = renderFigure(heroImage.url, heroImage.caption) + render(article.body.widgets);
+
+                return item;
+            })
+        )
+    );
+
+    return {
+        title: 'Sustainability Magazine Articles',
+        language: feedLang,
+        description: feedDescription,
+        link: `https://${feedURL}`,
+        item: items,
+    };
+}
diff --git a/lib/routes/sustainabilitymag/namespace.ts b/lib/routes/sustainabilitymag/namespace.ts
new file mode 100644
index 000000000..b43a42250
--- /dev/null
+++ b/lib/routes/sustainabilitymag/namespace.ts
@@ -0,0 +1,6 @@
+import type { Namespace } from '@/types';
+
+export const namespace: Namespace = {
+    name: 'Sustainability Magazine',
+    url: 'https://sustainabilitymag.com',
+};