fix(route): normalize /bwu/news rss output (#7295)

* fix: typo :P

* fix: reduce duplication of jQuery selectors

* fix: reformat invalid author tag

* fix: pubDate use parseDate

fix: replace depreciated url.resolve

Co-authored-by: TonyRL <TonyRL@users.noreply.github.com>
This commit is contained in:
muxinqi
2022-02-20 22:13:43 +08:00
committed by GitHub
co-authored by TonyRL
parent a1909d73be
commit 973a193a37
2 changed files with 15 additions and 18 deletions
+1 -1
View File
@@ -4,7 +4,7 @@ const util = require('./utils');
module.exports = async (ctx) => {
const title = '通知公告';
const host = 'http://news.bwu.edu.cn/';
const host = 'https://news.bwu.edu.cn/';
const path = 'tzgg.htm';
const response = await got({
+14 -17
View File
@@ -1,38 +1,35 @@
const got = require('@/utils/got');
const cheerio = require('cheerio');
const url = require('url');
const { parseDate } = require('@/utils/parse-date');
const timezone = require('@/utils/timezone');
// Processing full text
async function load(link) {
const responce = await got.get(link);
const $ = cheerio.load(responce.data);
const author = $('div[class=articleAuthor] p span').eq(1).text().slice(5);
const response = await got.get(link);
const $ = cheerio.load(response.data);
const author = $('div[class=articleAuthor] p span').eq(1).text() || 'null';
const authorName = author.indexOf(':') !== -1 ? author.split(':')[1].trim() : author;
const validAuthor = 'xcb@bwu.edu.cn (' + authorName + ')';
// Extracting text
const description = $('div[id^=vsb_content]').html();
return {
description,
author,
author: validAuthor,
};
}
const ProcessFeed = (list, cache) => {
const host = 'http://news.bwu.edu.cn/';
const host = 'https://news.bwu.edu.cn/';
return Promise.all(
list.map(async (item) => {
const $ = cheerio.load(item);
const title = $('a').text();
const link = url.resolve(host, $('a').attr('href'));
const aTag = $('a');
const title = aTag.text();
const link = new URL(aTag.attr('href'), host);
// Parsing the date
const date = new Date(
$('span[class=rightDate]')
.text()
.match(/\d{4}-\d{2}-\d{2}/)
);
const timeZone = 8;
const serverOffset = date.getTimezoneOffset() / 60;
const pubDate = new Date(date.getTime() - 60 * 60 * 1000 * (timeZone + serverOffset)).toUTCString();
const pubDate = timezone(parseDate($('span[class=rightDate]').text(), 'YYYY-MM-DD'), +8);
const single = {
title,
@@ -43,7 +40,7 @@ const ProcessFeed = (list, cache) => {
// Try caching
const other = await cache.tryGet(link, () => load(link));
// Return article information
return Promise.resolve(Object.assign({}, single, other));
return { ...single, ...other };
})
);
};