xiaoxue_admin
/
RSSHub
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Retrive latest posts from nga (#1424) 

Currently the NGA Thread API `/nga/post/:tid` retrieves only the first post in the thread. This PR retrieves all the posts in the last page of the given thread.  

This closes #1423 .
This commit is contained in:
Keren Zhu 2019-01-20 05:45:14 -08:00 committed by DIYgod
parent 7e43bd2f2c
commit a2ba04daf6
1 changed files with 57 additions and 33 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

90
lib/routes/nga/post.js
View File

@ -3,45 +3,69 @@ const cheerio = require('cheerio');
const iconv = require('iconv-lite');
module.exports = async (ctx) => {
const { tid } = ctx.params;
const link = `https://nga.178.com/read.php?tid=${tid}&rand=${Math.random() * 1000}`;
const timestamp = Math.floor(Date.now() / 1000);
const response = await axios({
method: 'get',
url: link,
responseType: 'arraybuffer',
headers: {
Cookie: `guestJs=${timestamp};`,
},
});
const getPageUrl = (tid, page = 1, hash = '') => `https://nga.178.com/read.php?tid=${tid}&page=${page}&rand=${Math.random() * 1000}#${hash}`;
const getPage = async (tid, pageId = 1) => {
const link = getPageUrl(tid, pageId);
const timestamp = Math.floor(Date.now() / 1000);
const response = await axios({
method: 'get',
url: link,
responseType: 'arraybuffer',
headers: {
Cookie: `guestJs=${timestamp};`,
},
});
const htmlString = iconv.decode(response.data, 'gbk');
const $ = cheerio.load(htmlString);
const htmlString = iconv.decode(response.data, 'gbk');
return cheerio.load(htmlString);
};
const getLastPageId = async (tid) => {
const $ = await getPage(tid);
const nav = $('#pagebtop');
const match = nav.html().match(/\{0:'\/read.php\?tid=(\d+)',1:(\d+).*?\}/);
return match ? match[2] : 1;
};
const eraseTag = (str) => str.replace(/\[(\w+)\].+?\[\/\1\]/g, '');
const eraseQuote = (str) => str.replace(/\[quote\](.+?)\[\/quote\]/, (m, b) => b).replace(/\[b\].*?\[uid=\d+\](.+?)\[\/uid\].+?\[\/b\]/, (m, user) => `@${user} `);
const pipeImg = (str) => str.replace(/\[img\](.+?)\[\/img\]/g, (m, src) => `<img src='${src[0] === '.' ? 'https://img.nga.178.com/attachments' + src.substr(1) : src}'></img>`);
const tid = ctx.params.tid;
const pageId = await getLastPageId(tid);
const $ = await getPage(tid, pageId);
const title = $('title').text() || '';
const description = $('#postcontentandsubject0').html() || '';
let pubDate = new Date($('#postdate0').text()).toUTCString();
if ($('#alertc0').length > 0) {
const matches = htmlString.match(/<span id='alertc0'><\/span><script>commonui\.loadAlertInfo\('\[E(\d+)[\s\S]+?\t|\n\]','alertc0'\)<\/script>/);
const items = $('#m_posts_c')
.children()
.filter('table')
.map((ind, post_) => {
const post = $(post_);
const content = post.find('.postcontent').first();
const title = eraseTag(content.text());
const description = eraseTag(pipeImg(eraseQuote(content.html() || '')));
const postId = content.attr('id');
const link = getPageUrl(tid, pageId, postId);
const pubDate = new Date(
post
.find('.postInfo > span')
.first()
.text()
).toUTCString();
if (matches && matches[1]) {
pubDate = new Date(parseInt(matches[1]) * 1000).toUTCString();
}
}
const items = [
{
title,
link,
description,
guid: pubDate,
pubDate,
},
];
return {
title,
link,
description,
pubDate,
guid: postId,
};
});
ctx.state.data = {
title: `nga-${title}`,
link,
item: items,
link: getPageUrl(tid, pageId),
item: items.get(),
};
};