feat: Add pacilution.com (普世社会科学研究所) router (#18036)

* Update package.json * Update package.json * feat: add RSS subscription for (NWMU) Northwestern Minzu University. * fix: fix array generation forms for dom-parser. * fix: fix url replacing behavior. * fix: fix cover images, typos and maintainer. * fix: remove router from deprecated routers. * fix: fix router name and example. * fix: fix items limit. * fix: fix date parser. * fix: use cache for the RSS data. * fix: typo * feat: add router for pacilution.com * fix: fix the url concat Co-authored-by: Tony <TonyRL@users.noreply.github.com> * fix: remove unused protocol Co-authored-by: Tony <TonyRL@users.noreply.github.com> * fix: remove unnecessary promise wrapping. * fix: add radar --------- Co-authored-by: pull[bot] <39814207+pull[bot]@users.noreply.github.com>
DIYgod · Jan 4, 2025 · ad0876c · ad0876c
1 parent 0091fc8
commit ad0876c
Show file tree

Hide file tree

Showing 2 changed files with 108 additions and 0 deletions.
diff --git a/lib/routes/pacilution/latest.ts b/lib/routes/pacilution/latest.ts
@@ -0,0 +1,102 @@
+import { DataItem, Route } from '@/types';
+import cache from '@/utils/cache';
+import got from '@/utils/got';
+import { parseDate } from '@/utils/parse-date';
+import { load } from 'cheerio';
+import iconv from 'iconv-lite';
+
+const BASE_URL = 'http://www.pacilution.com/';
+
+const handler: Route['handler'] = async () => {
+    // Fetch the target page
+    const response = await got({
+        method: 'get',
+        url: BASE_URL,
+        responseType: 'buffer',
+    });
+    const $ = load(iconv.decode(response.data, 'gb2312'));
+
+    // Select all list items containing target information
+    const ITEM_SELECTOR = 'ul[class*="ullbxwnew"] > li';
+    const listItems = $(ITEM_SELECTOR);
+
+    // Map through each list item to extract details
+    const contentLinkList = listItems.toArray().map((element) => {
+        const title = $(element).find('a').text();
+        const relativeHref = $(element).find('a').attr('href') || '';
+        const link = `${BASE_URL}${relativeHref}`;
+
+        return {
+            title,
+            link,
+        };
+    });
+
+    return {
+        title: '普世社会科学研究网最新文章',
+        description: '普世社会科学研究网首页上不同板块的最新文章汇总集合',
+        link: BASE_URL,
+        image: 'http://www.pacilution.com/img/top_banner.jpg',
+        item: (
+            await Promise.all(
+                contentLinkList.map((item) =>
+                    cache.tryGet(item.link, async () => {
+                        try {
+                            const CONTENT_SELECTOR = '#MyContent';
+                            const DATE_SELECTOR = 'td[class*="con_info"] > span';
+                            const response = await got({
+                                method: 'get',
+                                url: item.link,
+                                responseType: 'buffer',
+                            });
+                            const targetPage = load(iconv.decode(response.data, 'gb2312'));
+                            const content = targetPage(CONTENT_SELECTOR).html() || '';
+                            const date = parseDate(targetPage(DATE_SELECTOR).text().trim().replaceAll('日', '')).toISOString();
+                            return {
+                                title: item.title,
+                                pubDate: date,
+                                link: item.link,
+                                description: content,
+                                category: ['journal'],
+                                guid: item.link,
+                                id: item.link,
+                                image: 'http://www.pacilution.com/img/top_banner.jpg',
+                                content,
+                                updated: date,
+                                language: 'zh-cn',
+                            };
+                        } catch {
+                            return null as unknown as DataItem;
+                        }
+                    })
+                )
+            )
+        ).filter((item) => item !== null) as DataItem[],
+        allowEmpty: true,
+        language: 'zh-cn',
+        feedLink: 'https://rsshub.app/pacilution/latest',
+        id: 'https://rsshub.app/pacilution/latest',
+    };
+};
+
+export const route: Route = {
+    path: '/latest',
+    name: '最新文章',
+    maintainers: ['PrinOrange'],
+    handler,
+    categories: ['journal'],
+    features: {
+        requireConfig: false,
+        requirePuppeteer: false,
+        antiCrawler: false,
+        supportBT: false,
+        supportPodcast: false,
+        supportScihub: false,
+    },
+    example: '/pacilution/latest',
+    radar: [
+        {
+            source: ['www.pacilution.com'],
+        },
+    ],
+};
diff --git a/lib/routes/pacilution/namespace.ts b/lib/routes/pacilution/namespace.ts
@@ -0,0 +1,6 @@
+import type { Namespace } from '@/types';
+
+export const namespace: Namespace = {
+    name: '普世社会科学研究所',
+    url: 'www.pacilution.com',
+};