1
|
|
|
<?php |
2
|
|
|
namespace SEOstats\Services\Google; |
3
|
|
|
|
4
|
|
|
/** |
5
|
|
|
* SEOstats extension for Google data. |
6
|
|
|
* |
7
|
|
|
* @package SEOstats |
8
|
|
|
* @author Stephan Schmitz <[email protected]> |
9
|
|
|
* @copyright Copyright (c) 2010 - present Stephan Schmitz |
10
|
|
|
* @license http://eyecatchup.mit-license.org/ MIT License |
11
|
|
|
* @updated 2013/12/17 |
12
|
|
|
*/ |
13
|
|
|
|
14
|
|
|
use SEOstats\Common\SEOstatsException as E; |
15
|
|
|
use SEOstats\SEOstats as SEOstats; |
16
|
|
|
use SEOstats\Config as Config; |
17
|
|
|
use SEOstats\Helper as Helper; |
18
|
|
|
|
19
|
|
|
class Search extends SEOstats |
20
|
|
|
{ |
21
|
|
|
|
22
|
|
|
/** |
23
|
|
|
* Returns array, containing detailed results for any Google search. |
24
|
|
|
* |
25
|
|
|
* @param string $query String, containing the search query. |
26
|
|
|
* @param string $tld String, containing the desired Google top level domain. |
|
|
|
|
27
|
|
|
* @return array Returns array, containing the keys 'URL', 'Title' and 'Description'. |
28
|
|
|
*/ |
29
|
12 |
|
public static function getSerps($query, $maxResults=100, $domain=false) |
30
|
|
|
{ |
31
|
12 |
|
$q = rawurlencode($query); |
32
|
12 |
|
$maxPage = ceil(($maxResults/10)-1); |
33
|
12 |
|
$result = new Helper\ArrayHandle (); |
34
|
12 |
|
$pages = 1; |
35
|
12 |
|
$delay = 0; |
36
|
|
|
|
37
|
12 |
|
$domainRexExp = static::getDomainFilter($domain); |
38
|
|
|
|
39
|
12 |
|
for ($start=0; $start<$pages; $start++) { |
40
|
|
|
|
41
|
12 |
|
$haveNextPage = static::makeRequest ($start, $q, $result, $domainRexExp); |
42
|
12 |
|
if (!$haveNextPage) { |
43
|
9 |
|
$pages -= 1; |
44
|
9 |
|
} else { |
45
|
5 |
|
$pages += 1; |
46
|
5 |
|
$delay += 200000; |
47
|
5 |
|
usleep($delay); |
48
|
|
|
} |
49
|
|
|
|
50
|
12 |
|
if ($start == $maxPage) { |
51
|
9 |
|
$pages -= 1; |
52
|
9 |
|
} |
53
|
12 |
|
} // for ($start=0; $start<$pages; $start++) |
54
|
|
|
|
55
|
12 |
|
return $result->toArray(); |
56
|
|
|
} |
57
|
|
|
|
58
|
12 |
|
protected static function makeRequest ($start, $query, $result, $domainRexExp) |
59
|
|
|
{ |
60
|
12 |
|
$ref = static::getReference($start, $query); |
61
|
12 |
|
$nextSerp = static::getNextSerp($start, $query); |
62
|
|
|
|
63
|
12 |
|
$curledSerp = utf8_decode( static::gCurl($nextSerp, $ref) ); |
64
|
|
|
|
65
|
12 |
|
static::guardNoCaptcha($curledSerp); |
66
|
|
|
|
67
|
12 |
|
$matches = array(); |
68
|
12 |
|
preg_match_all('#<h3 class="?r"?>(.*?)</h3>#', $curledSerp, $matches); |
69
|
|
|
|
70
|
12 |
|
if (empty($matches[1])) { |
71
|
|
|
// No [@id="rso"]/li/h3 on currect page |
72
|
8 |
|
return false; |
73
|
|
|
} |
74
|
|
|
|
75
|
6 |
|
static::parseResults($matches, $domainRexExp, $start * 10, $result); |
76
|
|
|
|
77
|
6 |
|
if ( preg_match('#id="?pnnext"?#', $curledSerp) ) { |
78
|
|
|
// Found 'Next'-link on currect page |
79
|
5 |
|
return true; |
80
|
|
|
} |
81
|
|
|
|
82
|
|
|
// No 'Next'-link on currect page |
83
|
1 |
|
return false; |
84
|
|
|
} |
85
|
|
|
|
86
|
12 |
|
protected static function getReference ($start, $query) |
87
|
|
|
{ |
88
|
|
|
return 0 == $start |
89
|
12 |
|
? 'ncr' |
90
|
12 |
|
: sprintf('search?q=%s&hl=en&prmd=imvns&start=%s0&sa=N', $query, $start); |
91
|
|
|
} |
92
|
|
|
|
93
|
12 |
|
protected static function getDomainFilter ($domain) |
94
|
|
|
{ |
95
|
|
|
return $domain |
96
|
12 |
|
? "#^(https?://)?[^/]*{$domain}#i" |
97
|
12 |
|
: false; |
98
|
|
|
} |
99
|
|
|
|
100
|
12 |
|
protected static function getNextSerp ($start, $query) |
101
|
|
|
{ |
102
|
|
|
return 0 == $start |
103
|
12 |
|
? sprintf('search?q=%s&filter=0', $query) |
104
|
12 |
|
: sprintf('search?q=%s&filter=0&start=%s0', $query, $start); |
105
|
|
|
} |
106
|
|
|
|
107
|
12 |
|
protected static function guardNoCaptcha ($response) |
108
|
|
|
{ |
109
|
12 |
|
if (preg_match("#answer[=|/]86640#i", $response)) { |
110
|
|
|
print('Please read: https://support.google.com/websearch/answer/86640'); |
111
|
|
|
exit(); |
112
|
|
|
} |
113
|
12 |
|
} |
114
|
|
|
|
115
|
6 |
|
protected static function parseResults ($matches, $domainRexExp, $start, $result) |
116
|
|
|
{ |
117
|
6 |
|
$c = 0; |
118
|
|
|
|
119
|
6 |
|
foreach ($matches[1] as $link) { |
120
|
6 |
|
$match = static::parseLink($link); |
121
|
|
|
|
122
|
6 |
|
$c++; |
123
|
6 |
|
$resCnt = $start + $c; |
124
|
6 |
|
if (! $domainRexExp) { |
125
|
3 |
|
$result->setElement($resCnt, array( |
126
|
3 |
|
'url' => $match[1], |
127
|
3 |
|
'headline' => trim(strip_tags($match[2])) |
128
|
3 |
|
)); |
129
|
6 |
|
} elseif (preg_match($domainRexExp, $match[1])) { |
130
|
3 |
|
$result->push(array( |
131
|
3 |
|
'position' => $resCnt, |
132
|
3 |
|
'url' => $match[1], |
133
|
3 |
|
'headline' => trim(strip_tags($match[2])) |
134
|
3 |
|
)); |
135
|
3 |
|
} |
136
|
6 |
|
} // foreach ($matches[1] as $link) |
137
|
6 |
|
} |
138
|
|
|
|
139
|
6 |
|
protected static function parseLink($link) |
140
|
|
|
{ |
141
|
6 |
|
$isValidLink = preg_match('#<a\s+[^>]*href=[\'"]?([^\'" ]+)[\'"]?[^>]*>(.*?)</a>#', $link, $match); |
142
|
|
|
|
143
|
|
|
// is valid and not webmaster link |
144
|
6 |
|
return ( !$isValidLink || self::isAGoogleWebmasterLink($match[1]) ) |
145
|
6 |
|
? false |
146
|
6 |
|
: $match; |
147
|
|
|
} |
148
|
|
|
|
149
|
6 |
|
protected static function isAGoogleWebmasterLink($url) |
150
|
|
|
{ |
151
|
6 |
|
return preg_match('#^https?://www.google.com/(?:intl/.+/)?webmasters#', $url); |
152
|
|
|
} |
153
|
|
|
|
154
|
2 |
|
protected static function gCurl($path, $ref, $useCookie = Config\DefaultSettings::ALLOW_GOOGLE_COOKIES) |
155
|
|
|
{ |
156
|
2 |
|
$url = sprintf('https://www.google.%s/', Config\DefaultSettings::GOOGLE_TLD); |
157
|
2 |
|
$referer = $ref == '' ? $url : $ref; |
158
|
2 |
|
$url .= $path; |
159
|
|
|
|
160
|
2 |
|
$ua = "Mozilla/5.0 (Macintosh; Intel Mac OS X 10_9_3) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/35.0.1916.153 Safari/537.36"; |
161
|
2 |
|
if (isset($_SERVER["HTTP_USER_AGENT"]) && 0 < strlen($_SERVER["HTTP_USER_AGENT"])) { |
162
|
|
|
$ua = $_SERVER["HTTP_USER_AGENT"]; |
163
|
|
|
} |
164
|
|
|
|
165
|
|
|
$header = array( |
166
|
2 |
|
'Host: www.google.' . Config\DefaultSettings::GOOGLE_TLD, |
167
|
2 |
|
'Connection: keep-alive', |
168
|
2 |
|
'Cache-Control: max-age=0', |
169
|
2 |
|
'User-Agent: ' . $ua, |
170
|
2 |
|
'Accept: text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8', |
171
|
2 |
|
'Referer: ' . $referer, |
172
|
2 |
|
'Accept-Language: ' . Config\DefaultSettings::HTTP_HEADER_ACCEPT_LANGUAGE, |
173
|
|
|
'Accept-Charset: ISO-8859-1,utf-8;q=0.7,*;q=0.7' |
174
|
2 |
|
); |
175
|
|
|
|
176
|
2 |
|
$ch = curl_init($url); |
177
|
2 |
|
curl_setopt($ch, CURLOPT_SSL_VERIFYPEER, 0); |
178
|
2 |
|
curl_setopt($ch, CURLOPT_FOLLOWLOCATION, 1); |
179
|
2 |
|
curl_setopt($ch, CURLOPT_RETURNTRANSFER, 1); |
180
|
2 |
|
curl_setopt($ch, CURLOPT_HTTPHEADER, $header); |
181
|
2 |
|
curl_setopt($ch, CURLOPT_USERAGENT, $ua); |
182
|
2 |
|
if ($useCookie == 1) { |
183
|
|
|
curl_setopt($ch, CURLOPT_COOKIEJAR, __DIR__ . '/cookie.txt'); |
184
|
|
|
curl_setopt($ch, CURLOPT_COOKIEFILE, __DIR__ . '/cookie.txt'); |
185
|
|
|
} |
186
|
2 |
|
curl_setopt($ch, CURLOPT_CONNECTTIMEOUT, 30); |
187
|
2 |
|
$result = curl_exec($ch); |
188
|
2 |
|
$info = curl_getinfo($ch); |
189
|
2 |
|
curl_close($ch); |
190
|
2 |
|
return ($info['http_code']!=200) ? false : $result; |
191
|
|
|
} |
192
|
|
|
} |
193
|
|
|
|
This check looks for PHPDoc comments describing methods or function parameters that do not exist on the corresponding method or function.
Consider the following example. The parameter
$italy
is not defined by the methodfinale(...)
.The most likely cause is that the parameter was removed, but the annotation was not.