feat(route): add 国家药品监督管理局医疗器械标准管理中心 (#10206)

* feat(route): add 国家药品监督管理局医疗器械标准管理中心

* fix: shorten the route

* fix: bad urls

* fix: only 1 puppeteer request
This commit is contained in:
Ethan Shen
2022-07-16 20:47:53 +08:00
committed by GitHub
parent 0e563cde52
commit 54aea1e4e7
5 changed files with 115 additions and 0 deletions
+16
View File
@@ -231,6 +231,22 @@ pageClass: routes
</Route>
## 国家药品监督管理局医疗器械标准管理中心
### 通用
<Route author="nczitzk" example="/gov/nifdc/bshff/ylqxbzhgl/qxggtzh" path="/gov/nifdc/:path+" :paramsDesc="['路径,默认为公告通告']" radar="1" rssbud="1" puppeteer="1">
::: tip 提示
路径处填写对应页面 URL 中 `https://www.nifdc.gov.cn/nifdc/` 与 `/index.html` 之间的字段,下面是一个例子。
若订阅 [公告通告](https://www.nifdc.org.cn/nifdc/bshff/ylqxbzhgl/qxggtzh/index.html) 则将对应页面 URL <https://www.nifdc.org.cn/nifdc/bshff/ylqxbzhgl/qxggtzh/index.html> 中 `https://www.nifdc.org.cn/nifdc/` 和 `/index.html` 之间的字段 `bshff/ylqxbzhgl/qxggtzh` 作为路径填入。此时路由为 [`/gov/nifdc/bshff/ylqxbzhgl/qxggtzh`](https://rsshub.app/gov/nifdc/bshff/ylqxbzhgl/qxggtzh)
:::
</Route>
## 国家药品审评网站
### 首页
+1
View File
@@ -12,6 +12,7 @@ module.exports = {
'/miit/zcjd': ['Yoge-Code'],
'/moe/:type': ['Crawler995'],
'/moj/aac/news/:type?': ['TonyRL'],
'/nifdc/:path+': ['nczitzk'],
'/nmpa/:path+': ['TonyRL'],
'/nrta/news/:category?': ['yuxinliu-alex'],
'/nsfc/news/:type?': ['Derekmini', 'nczitzk'],
+86
View File
@@ -0,0 +1,86 @@
const got = require('@/utils/got');
const cheerio = require('cheerio');
const timezone = require('@/utils/timezone');
const { parseDate } = require('@/utils/parse-date');
module.exports = async (ctx) => {
const params = ctx.path === '/nifdc' ? '/nifdc/bshff/ylqxbzhgl/qxggtzh' : ctx.path;
const rootUrl = 'https://www.nifdc.org.cn';
const currentUrl = `${rootUrl}${params}/index.html`;
const response = await got({
method: 'get',
url: currentUrl,
});
const $ = cheerio.load(response.data);
let items = $('.list ul li a')
.slice(0, ctx.query.limit ? parseInt(ctx.query.limit) : 20)
.toArray()
.map((item) => {
item = $(item);
const link = item.attr('href');
return {
title: item.text(),
link: /^http/.test(link) ? link : new URL(link, currentUrl).href,
pubDate: parseDate(
item
.next()
.text()
.match(/\((.*)\)/)
),
};
});
const browser = await require('@/utils/puppeteer')();
items = await Promise.all(
items.map((item) =>
ctx.cache.tryGet(item.link, async () => {
if (/^https:\/\/www\.nmpa\.gov\.cn\//.test(item.link)) {
const page = await browser.newPage();
await page.setRequestInterception(true);
page.on('request', (request) => {
request.resourceType() === 'document' || request.resourceType() === 'script' ? request.continue() : request.abort();
});
await page.goto(item.link, {
waitUntil: 'domcontentloaded',
});
await page.waitForSelector('.text');
const html = await page.evaluate(() => document.documentElement.innerHTML);
await page.close();
const content = cheerio.load(html);
item.description = content('.text').html();
} else {
const detailResponse = await got({
method: 'get',
url: item.link,
});
const content = cheerio.load(detailResponse.data);
item.description = content('.text').html();
item.pubDate = timezone(parseDate(content('meta[name="PubDate"]').attr('content')), +8);
}
return item;
})
)
);
await browser.close();
ctx.state.data = {
title: $('title').text().replace(/----/, ' - '),
link: currentUrl,
item: items,
};
};
+11
View File
@@ -461,6 +461,17 @@ module.exports = {
},
],
},
'nifdc.gov.cn': {
_name: '国家药品监督管理局医疗器械标准管理中心',
'.': [
{
title: '通用',
docs: 'https://docs.rsshub.app/government.html#guo-jia-yao-pin-jian-du-guan-li-ju-yi-liao-qi-xie-biao-zhun-guan-li-zhong-xin',
source: ['/*path'],
target: (params) => `/gov/nifdc/${params.path.replace('/index.html', '')}`,
},
],
},
'nmpa.gov.cn': {
_name: '国家药品监督管理局',
'.': [
+1
View File
@@ -12,6 +12,7 @@ module.exports = function (router) {
router.get('/miit/zcjd', require('./miit/zcjd'));
router.get('/moe/:type', require('./moe/moe'));
router.get('/moj/aac/news/:type?', require('./moj/aac/news'));
router.get(/nifdc\/([\w/-]+)?/, require('./nifdc'));
router.get(/\/nmpa\/([\w][\w/]+)?/, require('./nmpa/generic'));
router.get('/nrta/news/:category?', require('./nrta/news'));
router.get('/nsfc/news/:type?', require('./nsfc'));