From b1a030c5f0fd6890abc2cfad5185c6d37244f010 Mon Sep 17 00:00:00 2001
From: "xxx.Yan" <yansiyu.old.dog@hotmail.com>
Date: Sun, 13 Oct 2024 03:33:39 +0800
Subject: [PATCH] feat(route): add route UK Parliament Lords Library research
 by topic (#16933)

* feat(route): add route UK Parliament Lords Library Research by Topic

* chore(route): rename parameter from tag to topic.

* fix: close browser after puppeteer usage

---------
---
 lib/routes/parliament.uk/lordslibrary.ts | 55 ++++++++++++++++++++++++
 lib/routes/parliament.uk/namespace.ts    |  6 +++
 2 files changed, 61 insertions(+)
 create mode 100644 lib/routes/parliament.uk/lordslibrary.ts
 create mode 100644 lib/routes/parliament.uk/namespace.ts

diff --git a/lib/routes/parliament.uk/lordslibrary.ts b/lib/routes/parliament.uk/lordslibrary.ts
new file mode 100644
index 00000000000000..e2d07f1cadac00
--- /dev/null
+++ b/lib/routes/parliament.uk/lordslibrary.ts
@@ -0,0 +1,55 @@
+import { load } from 'cheerio';
+import { Route } from '@/types';
+import puppeteer from '@/utils/puppeteer';
+import timezone from '@/utils/timezone';
+
+export const route: Route = {
+    path: '/lordslibrary/type/:topic?',
+    categories: ['government'],
+    example: '/parliament.uk/lordslibrary/type/research-briefing',
+    parameters: { topic: 'research by topic, string, example: [research-briefing|buisness|economy]' },
+    features: {
+        requireConfig: false,
+        requirePuppeteer: true,
+        antiCrawler: false,
+        supportBT: false,
+        supportPodcast: false,
+        supportScihub: false,
+    },
+    name: 'House of Lords Library',
+    maintainers: ['AntiKnot'],
+    handler,
+};
+
+async function handler(ctx) {
+    const { topic } = ctx.req.param();
+    const baseUrl = 'https://lordslibrary.parliament.uk/type';
+    const url = `${baseUrl}/${topic}/`;
+    const browser = await puppeteer();
+    const page = await browser.newPage();
+    await page.setRequestInterception(true);
+    page.on('request', (request) => {
+        request.resourceType() === 'document' ? request.continue() : request.abort();
+    });
+    await page.goto(url, {
+        waitUntil: 'domcontentloaded',
+    });
+
+    const html = await page.evaluate(() => document.documentElement.innerHTML);
+    await page.close();
+    const $ = load(html);
+    const items = $('article.card--horizontal')
+        .map((_, article) => ({
+            title: $(article).find('.card__text a').text().trim(),
+            link: $(article).find('.card__text a').attr('href'),
+            description: $(article).find('p').last().text().trim(),
+            pubDate: timezone($(article).find('.card__date time').attr('datetime')),
+        }))
+        .toArray();
+    browser.close();
+    return {
+        title: `parliament - lordslibrary - ${topic}`,
+        link: baseUrl,
+        item: items,
+    };
+}
diff --git a/lib/routes/parliament.uk/namespace.ts b/lib/routes/parliament.uk/namespace.ts
new file mode 100644
index 00000000000000..0480446f06aa4a
--- /dev/null
+++ b/lib/routes/parliament.uk/namespace.ts
@@ -0,0 +1,6 @@
+import type { Namespace } from '@/types';
+
+export const namespace: Namespace = {
+    name: 'UK Parliament',
+    url: 'parliament.uk',
+};