feat(route): add 哈尔滨工业大学研究生院通知公告 (#10747)

* 添加"哈尔滨工业大学研究生院通知公告"路由

* 修改路由为v2 script

* chore(deps): bump puppeteer from 17.1.2 to 17.1.3 (#1094)

Bumps [puppeteer](https://github.com/puppeteer/puppeteer) from 17.1.2 to 17.1.3.
- [Release notes](https://github.com/puppeteer/puppeteer/releases)
- [Changelog](https://github.com/puppeteer/puppeteer/blob/main/CHANGELOG.md)
- [Commits](https://github.com/puppeteer/puppeteer/compare/v17.1.2...v17.1.3)

---
updated-dependencies:
- dependency-name: puppeteer
  dependency-type: direct:production
  update-type: version-update:semver-patch
...

Signed-off-by: dependabot[bot] <support@github.com>

Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>

* Update lib/v2/hit/hitgs.js

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

* change code style of hitgs.js

* add radar.js

* change http to https for older HIT files

* refactor: migrate to v2

* docs: fix format

* fix: code scan

Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>

docs/university.md

@@ -1233,23 +1233,33 @@ category 列表：
 
 ## 哈尔滨工业大学
 
-### 哈尔滨工业大学教务处通知公告
+::: warning 注意
+哈工大网站疑似禁止了`rsshub.app`的访问，使用路由需要自行 [部署](https://docs.rsshub.app/install)。
+:::
 
-<Route author="lty96117" example="/hit/jwc" path="/hit/jwc"/>
+### 教务处通知公告
+
+<Route author="lty96117" example="/hit/jwc" path="/hit/jwc" radar="1" rssbud="1"/>
 
 ### 今日哈工大
 
-<Route author="ranpox" example="/hit/today/10" path="/hit/today/:category" :paramsDesc="['分类编号，`10`为公告公示，`11`为新闻快讯，同时支持详细分类，使用方法见下']"/>
+<Route author="ranpox" example="/hit/today/10" path="/hit/today/:category" :paramsDesc="['分类编号，`10`为公告公示，`11`为新闻快讯，同时支持详细分类，使用方法见下']" radar="1" rssbud="1">
 
 ::: tip 提示
 今日哈工大的文章分为公告公示和新闻快讯，每个页面右侧列出了更详细的分类，其编号为每个 URL 路径的最后一个数字。
-例如会议讲座的路径为`/taxonomy/term/10/25`，则可以通过`/hit/today/25`订阅该详细类别。
+例如会议讲座的路径为`/taxonomy/term/10/25`，则可以通过 [`/hit/today/25`](https://rsshub.app/hit/today/25) 订阅该详细类别。
 :::
 
 ::: warning 注意
 部分文章需要经过统一身份认证后才能阅读全文。
 :::
 
+</Route>
+
+### 研究生院通知公告
+
+<Route author="hlmu" example="/hit/hitgs" path="/hit/hitgs" radar="1" rssbud="1"/>
+
 ## 哈尔滨工业大学（深圳）
 
 ### 哈尔滨工业大学（深圳） - 新闻中心

lib/router.js

@@ -527,9 +527,9 @@ router.get('/hhu/libNews', lazyloadRouteHandler('./routes/universities/hhu/libNe
 // 河海大学常州校区
 router.get('/hhu/libNewsc', lazyloadRouteHandler('./routes/universities/hhu/libNewsc'));
 
-// 哈尔滨工业大学
-router.get('/hit/jwc', lazyloadRouteHandler('./routes/universities/hit/jwc'));
-router.get('/hit/today/:category', lazyloadRouteHandler('./routes/universities/hit/today'));
+// 哈尔滨工业大学 migrated to v2
+// router.get('/hit/jwc', lazyloadRouteHandler('./routes/universities/hit/jwc'));
+// router.get('/hit/today/:category', lazyloadRouteHandler('./routes/universities/hit/today'));
 
 // 哈尔滨工业大学（深圳）
 router.get('/hitsz/article/:category?', lazyloadRouteHandler('./routes/universities/hitsz/article'));

lib/routes/universities/hit/jwc.js

@@ -1,53 +0,0 @@
-const got = require('../../../utils/got');
-const cheerio = require('cheerio');
-const url = require('url');
-
-const baseUrl = 'http://jwc.hit.edu.cn';
-const type = (filename) => filename.split('.').pop();
-
-module.exports = async (ctx) => {
-    const response = await got.get(`${baseUrl}/2591/list.htm`);
-
-    const { data } = response;
-    const $ = cheerio.load(data);
-    const links = $('.news_list li')
-        .slice(0, 10)
-        .map((i, el) => ({
-            pubDate: new Date($('span.fbll', el).children().first().text().replace('[', '')).toUTCString(),
-            link: url.resolve(baseUrl, $('a', el).attr('href')),
-            title: $('a', el).attr('title'),
-        }))
-        .get();
-
-    const item = await Promise.all(
-        [...links].slice(0, 10).map(async ({ pubDate, link, title }) => {
-            if (type(link) === 'htm') {
-                const { data } = await got.get(link);
-                const $ = cheerio.load(data);
-
-                const author = $('p.arti_metas>span:nth-child(3)').text().trim();
-                const description =
-                    $('div.wp_articlecontent').html() &&
-                    $('div.wp_articlecontent')
-                        .html()
-                        .replace(/src="\//g, `src="${url.resolve(baseUrl, '.')}`)
-                        .replace(/href="\//g, `href="${url.resolve(baseUrl, '.')}`)
-                        .trim();
-                // check for some bug links.
-                if (!description) {
-                    return;
-                }
-                return Promise.resolve({ pubDate, author, link, title, description });
-            } else {
-                // file to download
-                return Promise.resolve({ pubDate, link, title, description: '此链接为文件，点击以下载' });
-            }
-        })
-    );
-
-    ctx.state.data = {
-        title: '哈尔滨工业大学教务处通知公告',
-        link: `${baseUrl}/2591/list.htm`,
-        item: item.filter((x) => x),
-    };
-};

lib/routes/universities/hit/today.js

@@ -1,85 +0,0 @@
-const got = require('@/utils/got');
-const cheerio = require('cheerio');
-const url = require('url');
-
-module.exports = async (ctx) => {
-    const host = 'http://today.hit.edu.cn';
-    const category = ctx.params.category;
-
-    const response = await got({
-        method: 'get',
-        url: host + '/category/' + category,
-        headers: {
-            Referer: host,
-        },
-    });
-
-    const $ = cheerio.load(response.data);
-    const category_name = $('div.banner-title').text();
-    const list = $('.paragraph li')
-        .slice(0, 10)
-        .map((i, e) => ({
-            path: $('span span a', e).attr('href'),
-            title: $('span span a', e).text(),
-            author: $('div a', e).attr('hreflang', 'zh-hans').text(),
-        }))
-        .get();
-
-    const out = await Promise.all(
-        list.map(async (item) => {
-            const { path, title, author } = item;
-            const link = host + path;
-
-            const cache = await ctx.cache.get(link);
-            if (cache) {
-                return Promise.resolve(JSON.parse(cache));
-            }
-
-            const response = await got({
-                method: 'get',
-                url: link,
-                headers: {
-                    Referer: host,
-                },
-            });
-
-            if (response.req._headers.host === 'ids.hit.edu.cn') {
-                const single = {
-                    pubDate: new Date(link.split('/').slice(-4, -1).join('-')).toUTCString(),
-                    author,
-                    link,
-                    title,
-                    description: '请进行统一身份认证后查看全文',
-                };
-                ctx.cache.set(link, JSON.stringify(single));
-                return Promise.resolve(single);
-            }
-
-            const $ = cheerio.load(response.data);
-            const description =
-                $('.article-content').html() &&
-                $('.article-content')
-                    .html()
-                    .replace(/src="\//g, `src="${url.resolve(host, '.')}`)
-                    .replace(/href="\//g, `href="${url.resolve(host, '.')}`)
-                    .trim();
-            const single = {
-                pubDate: new Date($('.left-attr.first').text()).toUTCString(),
-                author,
-                link,
-                title,
-                description,
-            };
-
-            ctx.cache.set(link, JSON.stringify(single));
-            return Promise.resolve(single);
-        })
-    );
-
-    ctx.state.data = {
-        title: '今日哈工大-' + category_name,
-        link: host + 'category/' + category,
-        description: '今日哈工大-' + category_name,
-        item: out,
-    };
-};

lib/v2/hit/hitgs.js

@@ -0,0 +1,46 @@
+const got = require('@/utils/got');
+const cheerio = require('cheerio');
+const { parseDate } = require('@/utils/parse-date');
+
+module.exports = async (ctx) => {
+    const host = 'https://hitgs.hit.edu.cn';
+
+    const response = await got(host + '/tzgg/list.htm', {
+        headers: {
+            Referer: host,
+        },
+    });
+
+    const $ = cheerio.load(response.data);
+    const list = $('.news_list li')
+        .map((i, e) => ({
+            pubDate: parseDate($('span:nth-child(4)', e).text()),
+            title: $('span.Article_BelongCreateOrg.newsfb', e).text() + $('span a', e).attr('title'),
+            category: $('span.Article_BelongCreateOrg.newsfb', e).text().slice(1, -1),
+            link: host + $('span a', e).attr('href'),
+        }))
+        .get();
+
+    const out = await Promise.all(
+        list.map((item) =>
+            ctx.cache.tryGet(item.link, async () => {
+                const response = await got(item.link, {
+                    headers: {
+                        Referer: host,
+                    },
+                });
+                const $ = cheerio.load(response.data);
+                item.description = $('.wp_articlecontent').html();
+                item.author = $('div.infobox > div > p > span:nth-child(3)').text().slice(3);
+                return item;
+            })
+        )
+    );
+
+    ctx.state.data = {
+        title: '哈工大研究生院通知公告',
+        link: host + '/tzgg/list.htm',
+        description: '哈尔滨工业大学研究生院通知公告',
+        item: out,
+    };
+};

lib/v2/hit/jwc.js

@@ -0,0 +1,60 @@
+const got = require('@/utils/got');
+const cheerio = require('cheerio');
+const { parseDate } = require('@/utils/parse-date');
+
+const baseUrl = 'https://jwc.hit.edu.cn';
+const type = (filename) => filename.split('.').pop();
+
+module.exports = async (ctx) => {
+    const response = await got(`${baseUrl}/2591/list.htm`);
+
+    const { data } = response;
+    const $ = cheerio.load(data);
+    const links = $('.news_list li')
+        .toArray()
+        .map((el) => {
+            el = $(el);
+            return {
+                pubDate: parseDate(el.find('span.fbll').children().first().text().replace(/\[/g, '')),
+                link: new URL(el.find('a').attr('href'), baseUrl).href,
+                title: el.find('a').attr('title'),
+            };
+        });
+
+    const items = await Promise.all(
+        links.map((item) =>
+            ctx.cache.tryGet(item.link, async () => {
+                if (type(item.link) === 'htm') {
+                    try {
+                        const { data } = await got(item.link);
+                        const $ = cheerio.load(data);
+
+                        const author = $('p.arti_metas>span:nth-child(3)').text().trim();
+                        const description =
+                            $('div.wp_articlecontent').html() &&
+                            $('div.wp_articlecontent')
+                                .html()
+                                .replace(/src="\//g, `src="${new URL('.', baseUrl).href}`)
+                                .replace(/href="\//g, `href="${new URL('.', baseUrl).href}`)
+                                .trim();
+
+                        item.author = author;
+                        item.description = description;
+                    } catch (e) {
+                        // intranet
+                    }
+                } else {
+                    // file to download
+                    item.description = '此链接为文件，点击以下载';
+                }
+                return item;
+            })
+        )
+    );
+
+    ctx.state.data = {
+        title: '哈尔滨工业大学教务处通知公告',
+        link: `${baseUrl}/2591/list.htm`,
+        item: items,
+    };
+};

lib/v2/hit/maintainer.js

@@ -0,0 +1,5 @@
+module.exports = {
+    '/hitgs': ['hlmu'],
+    '/jwc': ['lty96117'],
+    '/today/:category': ['ranpox'],
+};

lib/v2/hit/radar.js

@@ -0,0 +1,29 @@
+module.exports = {
+    'hit.edu.cn': {
+        _name: '哈尔滨工业大学',
+        jwc: [
+            {
+                title: '哈尔滨工业大学教务处通知公告',
+                docs: 'https://docs.rsshub.app/university.html#ha-er-bin-gong-ye-da-xue',
+                source: '/*',
+                target: '/hit/jwc',
+            },
+        ],
+        today: [
+            {
+                title: '今日哈工大',
+                docs: 'https://docs.rsshub.app/university.html#ha-er-bin-gong-ye-da-xue',
+                source: '/category/:category',
+                target: '/hit/today/:category',
+            },
+        ],
+        hitgs: [
+            {
+                title: '哈工大研究生院通知公告',
+                docs: 'https://docs.rsshub.app/university.html#ha-er-bin-gong-ye-da-xue',
+                source: '/*',
+                target: '/hit/hitgs',
+            },
+        ],
+    },
+};

lib/v2/hit/router.js

@@ -0,0 +1,6 @@
+module.exports = function (router) {
+    // 哈尔滨工业大学
+    router.get('/hitgs', require('./hitgs'));
+    router.get('/jwc', require('./jwc'));
+    router.get('/today/:category', require('./today'));
+};

lib/v2/hit/today.js

@@ -0,0 +1,59 @@
+const got = require('@/utils/got');
+const cheerio = require('cheerio');
+const { parseDate } = require('@/utils/parse-date');
+const timezone = require('@/utils/timezone');
+
+module.exports = async (ctx) => {
+    const host = 'https://today.hit.edu.cn';
+    const category = ctx.params.category;
+
+    const response = await got(host + '/category/' + category, {
+        headers: {
+            Referer: host,
+        },
+    });
+
+    const $ = cheerio.load(response.data);
+    const list = $('.paragraph li')
+        .toArray()
+        .map((e) => ({
+            link: new URL($('span span a', e).attr('href'), host).href,
+            title: $('span span a', e).text(),
+            author: $('div a', e).attr('hreflang', 'zh-hans').text(),
+            pubDate: timezone(parseDate($('span span a', e).attr('href').split('/').slice(-4, -1).join(), 'YYYYMMDD'), 8),
+        }));
+
+    const out = await Promise.all(
+        list.map((item) =>
+            ctx.cache.tryGet(item.link, async () => {
+                try {
+                    const response = await got(item.link, {
+                        headers: {
+                            Referer: host,
+                        },
+                    });
+
+                    const $ = cheerio.load(response.data);
+                    item.pubDate = timezone(parseDate($('.left-attr.first').text().trim()), 8);
+                    item.description =
+                        $('.article-content').html() &&
+                        $('.article-content')
+                            .html()
+                            .replace(/src="\//g, `src="${new URL('.', host).href}`)
+                            .replace(/href="\//g, `href="${new URL('.', host).href}`)
+                            .trim();
+                } catch (e) {
+                    // intranet
+                    item.description = '请进行统一身份认证后查看全文';
+                }
+                return item;
+            })
+        )
+    );
+
+    ctx.state.data = {
+        title: $('head title').text().trim(),
+        link: host + '/category/' + category,
+        item: out,
+    };
+};