Created
June 29, 2014 10:53
-
-
Save huzhifeng/c8a37d34c4e23cd2707d to your computer and use it in GitHub Desktop.
Node.js crawler for http://www.youqu5.com
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
var request = require('request'); | |
var cheerio = require('cheerio'); | |
//var Iconv = require('iconv').Iconv; | |
//var iconv = new Iconv('GBK', 'UTF-8//TRANSLIT//IGNORE'); | |
var iconv = require('iconv-lite'); | |
var util = require('util'); | |
var baseUrl = 'http://www.youqu5.com'; | |
var navItems = [ | |
{ | |
tid: 'qutu', | |
tname: '雷人趣图' | |
}, | |
{ | |
tid: 'gaoxiaotupian', | |
tname: '爆笑图片' | |
}, | |
{ | |
tid: 'neihantu', | |
tname: '邪恶内涵图' | |
}, | |
{ | |
tid: 'gaoxiaogif', | |
tname: '搞笑图片动态' | |
}, | |
{ | |
tid: 'chuanbangjingtou', | |
tname: '穿帮镜头' | |
}, | |
{ | |
tid: 'PSegaoqutu', | |
tname: '恶搞图片' | |
}, | |
{ | |
tid: 'shijiezhizui', | |
tname: '世界之最' | |
}, | |
{ | |
tid: 'yangyantupian', | |
tname: '养眼图片' | |
}, | |
{ | |
tid: 'gaoxiaomeinv', | |
tname: '搞笑美女' | |
}, | |
]; | |
var options = { | |
'url': '', | |
'method': 'GET', | |
'headers': { | |
'User-Agent': 'Mozilla/5.0 (Windows NT 5.1) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/31.0.1650.63 Safari/537.36', | |
'Connection': 'Keep-Alive', | |
'Host': 'www.youqu5.com' | |
}, | |
'encoding': null, | |
//'proxy': 'http://127.0.0.1:7788' | |
}; | |
function gbkDecode(data) { | |
return iconv.decode(data, 'gbk');//iconv.convert(body).toString(); | |
} | |
function isFullUrl(url) { | |
return /^https?:\/\//.test(url); | |
} | |
function formatUrl(url) { | |
if (!url) { | |
return ''; | |
} | |
if (isFullUrl(url)) { | |
return url; | |
} else { | |
if (url.charAt(0) === '/') { | |
return util.format('http://www.youqu5.com%s', url); | |
} else { | |
return util.format('http://www.youqu5.com/%s', url); | |
} | |
} | |
} | |
function crawlDetail(url) { | |
options.url = url; | |
request(options, function(err, res, body) { | |
console.log(util.format('Current detail: %s', url)); | |
if (err || res.statusCode != 200) { | |
console.log(util.inspect({ | |
'err': err, | |
'res': res | |
}, {depth: null})); | |
return; | |
} | |
var $ = cheerio.load(gbkDecode(body)); | |
var eImgs = $('div.tcontent img'); | |
var eDate = $('div.info p.time').first(); | |
var eTitle = $('div.info h1').first(); | |
if ((eImgs.length < 1) || | |
(eDate.length !== 1) || | |
(eTitle.length !== 1) || | |
!eDate.text().trim() || | |
!eTitle.text().trim()) { | |
console.log('Invalid eImgs, eDate or eTitle'); | |
console.log(util.inspect({ | |
'eImgs': eImgs, | |
'eDate': eDate, | |
'eTitle': eTitle | |
}, {depth: null})); | |
return; | |
} | |
var title = eTitle.text().trim() || $('title').text().trim(); | |
var date = eDate.text().trim(); | |
var match = date.match(/(\d{4}-\d{2}-\d{2})/); | |
if (match && (match.length === 2)) { | |
date = match[1]; | |
} else { | |
date = ''; | |
} | |
if (!title || !date) { | |
console.log('Invalid title or date'); | |
console.dir({ | |
'title': title, | |
'date': date | |
}); | |
return; | |
} | |
var imgs = []; | |
$(eImgs).each(function() { | |
var src = formatUrl($(this).attr('src')); | |
var alt = $(this).attr('title'); | |
if (!isFullUrl(src)) { | |
console.log('Invalid src:' + src); | |
console.dir($(this)); | |
return; | |
} | |
var img = { | |
'src': src | |
}; | |
if (alt) { | |
img.alt = alt; | |
} | |
imgs.push(img); | |
}); | |
if (imgs.length === 0) { | |
console.log('Invalid imgs'); | |
return; | |
} | |
var obj = { | |
'title': title, | |
'date': date, | |
'created': new Date(), | |
'tags': ['gif'], | |
'imgs': imgs, | |
'link': url | |
}; | |
console.dir(obj); | |
}); | |
} | |
function crawlTag(entry) { | |
options.url = entry.url; | |
request(options, function(err, res, body) { | |
console.log(util.format('Current page: %s', entry.url)); | |
if (err || res.statusCode != 200) { | |
console.log(util.inspect({ | |
'err': err, | |
'res': res | |
}, {depth: null})); | |
return; | |
} | |
var $ = cheerio.load(gbkDecode(body)); | |
$('div.box-l-m ul.pic-m li a.preview').each(function() { | |
var link = $(this).attr('href'); | |
if (isFullUrl(link)) { | |
setTimeout(crawlDetail, 1000, link); | |
} | |
}); | |
var nextUrl = ''; | |
$('div.pagelist a').each(function() { | |
if ($(this).text() === '下一页') { | |
nextUrl = $(this).attr('href'); | |
} | |
}); | |
if (isFullUrl(nextUrl)) { | |
entry.url = nextUrl; | |
setTimeout(crawlTag, 1000, entry); | |
} else { | |
console.log(util.format('Last page of %s[%s]', entry.tname, entry.tid)); | |
} | |
}); | |
} | |
function main() { | |
navItems.forEach(function(entry, index, list) { | |
entry.url = util.format('%s/%s/', baseUrl, entry.tid); | |
setTimeout(crawlTag, index * 30 * 1000, entry); | |
}); | |
} | |
main(); |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment