feat: add X-mol 平台 (#3279)

docs/study.md

@@ -110,3 +110,13 @@ pageClass: routes
 ### 最新
 
 <Route author="xyqfer" example="/icourse163/newest" path="/icourse163/newest" />
+
+## X-MOL 平台
+
+### 新闻
+
+<Route author="cssxsh" example="/x-mol/news/3" path="/x-mol/news/:tag?" :paramsDesc="['数字编号，可从新闻列表URL得到。为空时从新闻主页获取新闻。']" />
+
+### 期刊
+
+<Route author="cssxsh" example="/x-mol/paper/0/9" path="/x-mol/paper/:type/:magazine" :paramsDesc="['类别','机构，两个参数都可从期刊URL获取。']" />

lib/router.js

@@ -1833,6 +1833,10 @@ router.get('/kzfeed/topic/:id', require('./routes/kzfeed/topic'));
 // 腾讯新闻较真查证平台
 router.get('/factcheck', require('./routes/tencent/factcheck'));
 
+// X-MOL化学资讯平台
+router.get('/x-mol/news/:tag?', require('./routes/x-mol/news.js'));
+router.get('/x-mol/paper/:type/:magazine', require('./routes/x-mol/paper'));
+
 // 電撃Online
 router.get('/dengekionline/:type?', require('./routes/dengekionline/new'));
 

lib/routes/x-mol/news.js

@@ -0,0 +1,58 @@
+const got = require('@/utils/got');
+const cheerio = require('cheerio');
+const utils = require('./utils');
+
+module.exports = async (ctx) => {
+    const tag = ctx.params.tag;
+    const path = tag ? `/news/tag/${tag}` : '/news/index';
+    const response = await got(path, {
+        method: 'GET',
+        baseUrl: utils.host,
+    });
+    const data = response.data;
+    const $ = cheerio.load(data);
+
+    const title = $('title').text();
+    const description = $('meta[name="description"]').attr('content');
+    const newsitem = $('.newsitem');
+
+    const item = newsitem
+        .map((index, element) => {
+            const title = $(element)
+                .find('h3')
+                .find('a')
+                .text();
+            const a = $(element)
+                .find('p')
+                .find('a');
+            const link = utils.host + a.attr('href');
+            const image = $(element)
+                .find('img')
+                .attr('src');
+            const description = utils.setDesc(image, a.text());
+            const span = $(element).find('.space-right-m30');
+            const author = span
+                .text()
+                .replace('来源：', '')
+                .trim();
+            const date = utils.getDate(span.next().text());
+            const pubDate = utils.transDate(date);
+
+            const single = {
+                title: title,
+                link: link,
+                description: description,
+                author: author,
+                pubDate: pubDate,
+            };
+            return single;
+        })
+        .get();
+
+    ctx.state.data = {
+        title: title,
+        link: response.url,
+        description: description,
+        item: item,
+    };
+};

lib/routes/x-mol/paper.js

@@ -0,0 +1,76 @@
+const got = require('@/utils/got');
+const cheerio = require('cheerio');
+const utils = require('./utils');
+
+module.exports = async (ctx) => {
+    const type = ctx.params.type;
+    const magazine = ctx.params.magazine;
+    const path = `/paper/${type}/${magazine}`;
+    const response = await got(path, {
+        method: 'GET',
+        baseUrl: utils.host,
+        headers: {
+            Cookie: 'closeFloatWindow=true; journalIndexViewType=list; journalSort=publishDate',
+        },
+    });
+    const data = response.data;
+    const $ = cheerio.load(data);
+
+    const title = $('title').text();
+    const description = $('meta[name="description"]').attr('content');
+    const newsitem = $('.magazine-text');
+
+    const item = await Promise.all(
+        newsitem
+            .map(async (index, element) => {
+                const news = $(element);
+
+                const a = news.find('.magazine-text-title').find('a');
+                const title = a.text();
+                const link = utils.host + a.attr('href');
+
+                const imageId = news
+                    .find('.magazine-pic')
+                    .attr('id')
+                    .substring(9);
+                const getLink = utils.host + '/attachment/getImgUrl';
+                const noPic = utils.host + '/css/images/nothesispic.jpg';
+                const imageUrl = await ctx.cache.tryGet(getLink, async () => {
+                    const result = await got.get(getLink, {
+                        params: {
+                            attachmentId: imageId,
+                        },
+                    });
+                    return result.data;
+                });
+                const image = imageUrl || noPic;
+                const text = $(element)
+                    .find('.magazine-description')
+                    .text();
+                const description = utils.setDesc(image, text);
+
+                const span = news.find('.magazine-text-atten');
+                const arr = span.map((index, element) => $(element).text()).get();
+                const author = arr[1];
+                const date = utils.getDate(arr[0]);
+                const pubDate = utils.transDate(date);
+
+                const single = {
+                    title: title,
+                    link: link,
+                    description: description,
+                    author: author,
+                    pubDate: pubDate,
+                };
+                return Promise.resolve(single);
+            })
+            .get()
+    );
+
+    ctx.state.data = {
+        title: title,
+        link: response.url,
+        description: description,
+        item: item,
+    };
+};

lib/routes/x-mol/utils.js

@@ -0,0 +1,16 @@
+const XmolUtils = {
+    host: 'https://www.x-mol.com',
+    transDate: (date) => new Date(`${date} GMT+0800`).toUTCString(),
+    getDate: (text) => {
+        const reg = /[1-9]\d{3}-(0[1-9]|1[0-2])-(0[1-9]|[1-2][0-9]|3[0-1])/;
+        if (typeof text === 'string') {
+            const arr = text.match(reg);
+            return arr && text.match(reg)[0];
+        } else {
+            return null;
+        }
+    },
+    setDesc: (image, text) => `<p><img src="${image}"></p><p>${text}</p>`,
+};
+
+module.exports = XmolUtils;