如何抓取开了gzip的网页
有时候用 file_get_contents() 函数抓取网页会发生乱码现象。有两个原因会导致乱码，一个是编码问题，一个是目标页面开了Gzip。
编码问题好办，把抓取到的内容转下编码即可($content=iconv("GBK", "UTF-8//IGNORE", $content);)，我们这里讨论的是如何抓取开了Gzip的页面。怎么判断呢？获取的头部当中有Content-Encoding: gzip说明内容是GZIP压缩的。用FireBug看一下就知道页面开了gzip没有。下面是用firebug查看我的博客的头信息，Gzip是开了的。
1 |
请求头信息原始头信息 |
2 |
Accept text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8 |
3 |
Accept-Encoding gzip, deflate |
4 |
Accept-Language zh-cn,zh;q=0.8,en-us;q=0.5,en;q=0.3 |
5 |
Connection keep-alive |
6 |
Cookie __utma=225240837.787252530.1317310581.1335406161.1335411401.1537; __utmz=225240837.1326850415.887.3.utmcsr=google|utmccn=(organic)|utmcmd=organic|utmctr=%E4%BB%BB%E4%BD%95%E9%A1%B9%E7%9B%AE%E9%83%BD%E4%B8%8D%E4%BC%9A%E9%82%A3%E4%B9%88%E7%AE%80%E5%8D%95%20site%3Awww.nowamagic.net; PHPSESSID=888mj4425p8s0m7s0frre3ovc7; __utmc=225240837; __utmb=225240837.1.10.1335411401 |
7 |
Host www.nowamagic.net |
8 |
User-Agent Mozilla/5.0 (Windows NT 5.1; rv:12.0) Gecko/20100101 Firefox/12.0 |
下面介绍一些解决方案：
1. 使用自带的zlib库
如果服务器已经装了zlib库，用下面的代码可以轻易解决乱码问题。
1 |
$data = file_get_contents("compress.zlib://".$url); |
2. 使用CURL代替file_get_contents
1 |
function curl_get($url, $gzip=false){ |
2 |
$curl = curl_init($url); |
3 |
curl_setopt($curl, CURLOPT_RETURNTRANSFER, 1); |
4 |
curl_setopt($curl, CURLOPT_CONNECTTIMEOUT, 10); |
5 |
if($gzip) curl_setopt($curl, CURLOPT_ENCODING, "gzip"); // 关键在这里 |
6 |
$content = curl_exec($curl); |
7 |
curl_close($curl); |
8 |
return $content; |
9 |
} |
3. 使用gzip解压函数
001 |
function gzdecode($data) { |
002 |
$len = strlen($data); |
003 |
if ($len < 18 || strcmp(substr($data,0,2),"\x1f\x8b")) { |
004 |
return null; // Not GZIP format (See RFC 1952) |
005 |
} |
006 |
$method = ord(substr($data,2,1)); // Compression method |
007 |
$flags = ord(substr($data,3,1)); // Flags |
008 |
if ($flags & 31 != $flags) { |
009 |
// Reserved bits are set -- NOT ALLOWED by RFC 1952 |
010 |
return null; |
011 |
} |
012 |
// NOTE: $mtime may be negative (PHP integer limitations) |
013 |
$mtime = unpack("V", substr($data,4,4)); |
014 |
$mtime = $mtime[1]; |
015 |
$xfl = substr($data,8,1); |
016 |
$os = substr($data,8,1); |
017 |
$headerlen = 10; |
018 |
$extralen = 0; |
019 |
$extra = ""; |
020 |
if ($flags & 4) { |
021 |
// 2-byte length prefixed EXTRA data in header |
022 |
if ($len - $headerlen - 2 < 8) { |
023 |
return false; // Invalid format |
024 |
} |
025 |
$extralen = unpack("v",substr($data,8,2)); |
026 |
$extralen = $extralen[1]; |
027 |
if ($len - $headerlen - 2 - $extralen < 8) { |
028 |
return false; // Invalid format |
029 |
} |
030 |
$extra = substr($data,10,$extralen); |
031 |
$headerlen += 2 + $extralen; |
032 |
} |
033 |
034 |
$filenamelen = 0; |
035 |
$filename = ""; |
036 |
if ($flags & 8) { |
037 |
// C-style string file NAME data in header |
038 |
if ($len - $headerlen - 1 < 8) { |
039 |
return false; // Invalid format |
040 |
} |
041 |
$filenamelen = strpos(substr($data,8+$extralen),chr(0)); |
042 |
if ($filenamelen === false || $len - $headerlen - $filenamelen - 1 < 8) { |
043 |
return false; // Invalid format |
044 |
} |
045 |
$filename = substr($data,$headerlen,$filenamelen); |
046 |
$headerlen += $filenamelen + 1; |
047 |
} |
048 |
049 |
$commentlen = 0; |
050 |
$comment = ""; |
051 |
if ($flags & 16) { |
052 |
// C-style string COMMENT data in header |
053 |
if ($len - $headerlen - 1 < 8) { |
054 |
return false; // Invalid format |
055 |
} |
056 |
$commentlen = strpos(substr($data,8+$extralen+$filenamelen),chr(0)); |
057 |
if ($commentlen === false || $len - $headerlen - $commentlen - 1 < 8) { |
058 |
return false; // Invalid header format |
059 |
} |
060 |
$comment = substr($data,$headerlen,$commentlen); |
061 |
$headerlen += $commentlen + 1; |
062 |
} |
063 |
064 |
$headercrc = ""; |
065 |
if ($flags & 1) { |
066 |
// 2-bytes (lowest order) of CRC32 on header present |
067 |
if ($len - $headerlen - 2 < 8) { |
068 |
return false; // Invalid format |
069 |
} |
070 |
$calccrc = crc32(substr($data,0,$headerlen)) & 0xffff; |
071 |
$headercrc = unpack("v", substr($data,$headerlen,2)); |
072 |
$headercrc = $headercrc[1]; |
073 |
if ($headercrc != $calccrc) { |
074 |
return false; // Bad header CRC |
075 |
} |
076 |
$headerlen += 2; |
077 |
} |
078 |
079 |
// GZIP FOOTER - These be negative due to PHP's limitations |
080 |
$datacrc = unpack("V",substr($data,-8,4)); |
081 |
$datacrc = $datacrc[1]; |
082 |
$isize = unpack("V",substr($data,-4)); |
083 |
$isize = $isize[1]; |
084 |
085 |
// Perform the decompression: |
086 |
$bodylen = $len-$headerlen-8; |
087 |
if ($bodylen < 1) { |
088 |
// This should never happen - IMPLEMENTATION BUG! |
089 |
return null; |
090 |
} |
091 |
$body = substr($data,$headerlen,$bodylen); |
092 |
$data = ""; |
093 |
if ($bodylen > 0) { |
094 |
switch ($method) { |
095 |
case 8: |
096 |
// Currently the only supported compression method: |
097 |
$data = gzinflate($body); |
098 |
break; |
099 |
default: |
100 |
// Unknown compression method |
101 |
return false; |
102 |
} |
103 |
} else { |
104 |
// I'm not sure if zero-byte body content is allowed. |
105 |
// Allow it for now... Do nothing... |
106 |
} |
107 |
108 |
// Verifiy decompressed size and CRC32: |
109 |
// NOTE: This may fail with large data sizes depending on how |
110 |
// PHP's integer limitations affect strlen() since $isize |
111 |
// may be negative for large sizes. |
112 |
if ($isize != strlen($data) || crc32($data) != $datacrc) { |
113 |
// Bad format! Length or CRC doesn't match! |
114 |
return false; |
115 |
} |
116 |
return $data; |
117 |
} |
使用：
1 |
$html=file_get_contents('http://www.nowamagic.net/librarys/veda/'); |
2 |
$html=gzdecode($html); |
就介绍这三个方法，应该能解决大部分gzip引起的抓取乱码问题了。
浙公网安备 33010602011771号