当前位置:首页 >> 网络编程

nodejs爬虫遇到的乱码问题汇总

上一篇文章中使用nodejs程序解析了网页编码为gbk,gb2312,以及utf-8的情况,这里面有三种特殊的乱码情况需要单独的说明一下.

1,网页编码为utf-8,但是解析为乱码,代表网站为www.guoguo-app.com。

这个问题真是个逗逼问题,查看网页源码中给出的编码方式为utf8,如下:

<meta charset="UTF-8">
<title>查快递</title>

由于解析出来的始终是乱码,我就抓包看了下,报文中的编码方式为gbk,果然我使用gbk的方式之后,得到的不再是乱码了。淘宝为了反爬虫也是操碎了新,但是我也很好奇这种方式是怎么实现的,知道的告诉我。

GET / HTTP/1.1
Host: www.guoguo-app.com
Connection: close

HTTP/1.1 200 OK
Date: Thu, 06 Apr 2017 01:56:23 GMT
Content-Type: text/html;charset=GBK
Transfer-Encoding: chunked
Connection: close
Vary: Accept-Encoding
Vary: Accept-Encoding
Content-Language: zh-CN
Server: Tengine/Aserver


1,网页编码为utf-8,解析为乱码情况二,代表网站http//andersonjiang.blog.sohu.com/

单纯的查看网页源码看不出任何毛病,于是我又抓了一次包,得到如下情形:

GET / HTTP/1.1
Host: andersonjiang.blog.sohu.com
Connection: close

HTTP/1.1 200 OK
Content-Type: text/html; charset=GBK
Transfer-Encoding: chunked
Connection: close
Server: nginx
Date: Thu, 06 Apr 2017 02:10:33 GMT
Vary: Accept-Encoding
Expires: Thu, 01 Jan 1970 00:00:00 GMT
RHOST: 192.168.110.68@11177
Pragma: No-cache
Cache-Control: no-cache
Content-Language: en-US
Content-Encoding: gzip
FSS-Cache: MISS from 13539701.18454911.21477824
FSS-Proxy: Powered by 9935166.11245896.17873234

andersonjiang.blog.sohu.com这个网站同时采用了Transfer-Encoding: chunked传输编码和Content-Encoding: gzip内容编码功能,由于nodejs爬虫没有gzip解包功能,因此该网站提取不到任何字段,即title和charset等。要想提取此类网站则要添加gzip解包功能。

下面两个网站www.cr173.com以及www.csdn.net是正常的抓包情况。

GET / HTTP/1.1
Host: www.cr173.com
Connection: close

HTTP/1.1 200 OK
Expires: Thu, 06 Apr 2017 02:42:20 GMT
Date: Thu, 06 Apr 2017 02:12:20 GMT
Content-Type: text/html
Last-Modified: Thu, 06 Apr 2017 00:52:42 GMT
ETag: "96a4141970aed21:0"
Cache-Control: max-age=1800
Accept-Ranges: bytes
Content-Length: 158902
Accept-Ranges: bytes
X-Varnish: 1075189606
Via: 1.1 varnish
X-Via: 1.1 dxxz46:4 (Cdn Cache Server V2.0), 1.1 oudxin15:1 (Cdn Cache Server V2.0)
Connection: close
GET / HTTP/1.1
Host: www.csdn.net
Connection: close

HTTP/1.1 200 OK
Server: openresty
Date: Thu, 06 Apr 2017 02:18:59 GMT
Content-Type: text/html; charset=utf-8
Content-Length: 99363
Connection: close
Vary: Accept-Encoding
Last-Modified: Thu, 06 Apr 2017 02:10:02 GMT
Vary: Accept-Encoding
ETag: "58e5a37a-18423"
Accept-Ranges: bytes

3,网页编码为其他形式的编码,解析为乱码,例如:

(1)编码为Big5,代表网站为 www.ruten.com.tw, www.ctgoodjobs.hk

(2)编码为Shift_JIS,代表网站为www.vector.co.jp,www.smbc.co.jp

(3)编码为windows-12,代表网站为www.tff.org,www.pravda.com.ua

(4)编码为EUC-JP,代表网站为www.showtime.jp

(5)编码为EUC-KR ,代表网站为www.incruit.com,www.samsunghospital.com,

由于iconv-lite的说明中支持如下的编码方式:

Currently only a small part of encodings supported:

 All node.js native encodings: 'utf8', 'ucs2', 'ascii', 'binary', 'base64'.
 Base encodings: 'latin1'
 Cyrillic encodings: 'windows-1251', 'koi8-r', 'iso 8859-5'.
 Simplified chinese: 'gbk', 'gb2313'.

Other encodings are easy to add, see the source. Please, participate
因此对于上述出现的网页编码,只有自己添加解码方式加以解决了。

总之要写一个通用的爬虫程序还有很长的路要走。