Merge pull request #2051 from DIYgod/master

[pull] master from diygod:master

lib/routes/ce/district.ts

@@ -0,0 +1,91 @@
+import { Route, ViewType } from '@/types';
+import cache from '@/utils/cache';
+import { parseDate } from '@/utils/parse-date';
+import timezone from '@/utils/timezone';
+import { load } from 'cheerio';
+import { ofetch } from 'ofetch';
+
+export const route: Route = {
+    path: '/district/:category?',
+    name: '地方经济',
+    url: 'district.ce.cn',
+    maintainers: ['cscnk52'],
+    handler,
+    example: '/ce/district',
+    parameters: { category: '栏目标识，默认为 roll（即时新闻）' },
+    description: `| 即时新闻 | 经济动态 | 独家视角 | 专题 | 数说地方 | 地方播报 | 专稿 | 港澳台 |
+|----------|----------|----------|------|----------|----------|------|--------|
+| roll     | jjdt     | poll     | ch   | ssdf     | dfbb     | zg   | gat    |`,
+    categories: ['traditional-media'],
+    features: {
+        requireConfig: false,
+        requirePuppeteer: false,
+        antiCrawler: false,
+        supportBT: false,
+        supportPodcast: false,
+        supportScihub: false,
+        supportRadar: true,
+    },
+    radar: [
+        {
+            source: ['district.ce.cn/newarea/:category/index.shtml'],
+            target: '/district/:category?',
+        },
+        {
+            source: ['district.ce.cn/newarea/:category'],
+            target: '/district/:category?',
+        },
+        {
+            source: ['district.ce.cn'],
+            target: '/district',
+        },
+    ],
+    view: ViewType.Articles,
+};
+
+async function handler(ctx) {
+    const rootUrl = 'http://district.ce.cn/';
+    const { category = 'roll' } = ctx.req.param();
+    const url = `${rootUrl}newarea/${category}/index.shtml`;
+    const GB2312Response = await ofetch(url, { responseType: 'arrayBuffer' });
+
+    // originally site use gb2312 encoding
+    const response = new TextDecoder('gb2312').decode(new Uint8Array(GB2312Response));
+    const $ = load(response);
+
+    const bigTitle = $('div.channl a').eq(1).attr('title');
+
+    const list = $('div.sec_left li')
+        .toArray()
+        .map((e) => {
+            const element = $(e);
+            const title = element.find('a').text().trim();
+            const link = new URL(element.find('a').attr('href'), url).href;
+            return {
+                title,
+                link,
+            };
+        });
+
+    const items = await Promise.all(
+        list.map((item) =>
+            cache.tryGet(item.link, async () => {
+                const GB2312Response = await ofetch(item.link, { responseType: 'arrayBuffer' });
+                const response = new TextDecoder('gb2312').decode(new Uint8Array(GB2312Response));
+                const $ = load(response);
+
+                const pubDateText = $('span#articleTime').text().trim();
+                item.pubDate = timezone(parseDate(pubDateText, 'YYYY年MM月DD日 HH:mm'), +8);
+
+                item.description = $('div.TRS_Editor').html();
+                return item;
+            })
+        )
+    );
+
+    return {
+        title: `中国经济网地方经济 - ${bigTitle}`,
+        link: url,
+        item: items,
+    };
+}

lib/routes/ce/namespace.ts

@@ -0,0 +1,8 @@
+import type { Namespace } from '@/types';
+
+export const namespace: Namespace = {
+    name: '中国经济网',
+    url: 'www.ce.cn',
+    categories: ['traditional-media'],
+    lang: 'zh-CN',
+};

lib/routes/whitehouse/news.ts

@@ -0,0 +1,82 @@
+import { Route } from '@/types';
+import cache from '@/utils/cache';
+import got from '@/utils/got';
+import { load } from 'cheerio';
+import { parseDate } from '@/utils/parse-date';
+
+export const route: Route = {
+    path: '/news/:category?',
+    categories: ['government'],
+    example: '/whitehouse/news',
+    parameters: { category: 'Category, see below, all by default' },
+    features: {
+        requireConfig: false,
+        requirePuppeteer: false,
+        antiCrawler: false,
+        supportBT: false,
+        supportPodcast: false,
+        supportScihub: false,
+    },
+    radar: [
+        {
+            source: ['whitehouse.gov/:category', 'whitehouse.gov/'],
+            target: '/news/:category',
+        },
+    ],
+    name: 'News',
+    maintainers: ['nczitzk', 'hkamran80'],
+    handler,
+    description: `| All | Articles | Briefings and Statements | Presidential Actions | Remarks |
+  | --- | -------- | ------------------------ | -------------------- | ------- |
+  |     | articles | briefings-statements     | presidential-actions | remarks |`,
+};
+
+async function handler(ctx) {
+    const category = ctx.req.param('category') ?? 'news';
+
+    const rootUrl = 'https://www.whitehouse.gov';
+    const currentUrl = `${rootUrl}/${category}/`;
+    const response = await got({
+        method: 'get',
+        url: currentUrl,
+    });
+
+    const $ = load(response.data);
+
+    const list = $('.post')
+        .toArray()
+        .map((item) => {
+            item = $(item);
+
+            const a = item.find('a').first();
+            return {
+                title: a.text(),
+                link: a.attr('href'),
+                pubDate: parseDate(item.find('time').attr('datetime')),
+                category: [item.find('a[rel^=tag]').first().text()],
+            };
+        });
+
+    const items = await Promise.all(
+        list.map((item) =>
+            cache.tryGet(item.link, async () => {
+                const response = await got({
+                    method: 'get',
+                    url: item.link,
+                });
+                const $ = load(response.data);
+
+                $('.wp-block-whitehouse-topper').remove();
+                item.description = $('.entry-content').html();
+
+                return item;
+            })
+        )
+    );
+
+    return {
+        title: $('title').text(),
+        link: currentUrl,
+        item: items,
+    };
+}