|
1
|
|
|
<?php |
|
2
|
|
|
/** |
|
3
|
|
|
* Created by PhpStorm. |
|
4
|
|
|
* User: lenon |
|
5
|
|
|
* Date: 23/04/16 |
|
6
|
|
|
* Time: 01:43. |
|
7
|
|
|
*/ |
|
8
|
|
|
|
|
9
|
|
|
namespace Aszone\SearchHacking\Engines; |
|
10
|
|
|
|
|
11
|
|
|
use Aszone\SearchHacking\Utils; |
|
12
|
|
|
|
|
13
|
|
|
class DuckDuckGo extends Engine |
|
14
|
|
|
{ |
|
15
|
|
|
public function run() |
|
16
|
|
|
{ |
|
17
|
|
|
$exit = false; |
|
18
|
|
|
$count = 0; |
|
19
|
|
|
$numPaginator = 0; |
|
|
|
|
|
|
20
|
|
|
$countProxyVirgin = rand(0, count($this->listOfVirginProxies) - 1); |
|
21
|
|
|
$resultFinal = array(); |
|
22
|
|
|
$countError = 0; |
|
23
|
|
|
|
|
24
|
|
|
$numberForUrl = $this->getNumberForUrl(); |
|
25
|
|
|
|
|
26
|
|
|
while ($exit == false) { |
|
|
|
|
|
|
27
|
|
|
switch ($count) { |
|
28
|
|
|
case 0: |
|
29
|
|
|
$numPaginator = 0; |
|
30
|
|
|
break; |
|
31
|
|
|
case 1: |
|
32
|
|
|
$numPaginator = 30; |
|
33
|
|
|
break; |
|
34
|
|
|
case 2: |
|
35
|
|
|
$numPaginator = 80; |
|
36
|
|
|
break; |
|
37
|
|
|
case 3: |
|
38
|
|
|
$numPaginator = 130; |
|
39
|
|
|
break; |
|
40
|
|
|
case 4: |
|
41
|
|
|
$numPaginator = 180; |
|
42
|
|
|
break; |
|
43
|
|
|
default: |
|
44
|
|
|
$numPaginator = 230; |
|
45
|
|
|
break; |
|
46
|
|
|
|
|
47
|
|
|
} |
|
48
|
|
|
|
|
49
|
|
|
$urlOfSearch = 'https://duckduckgo.com/d.js?q='.urlencode($this->commandData['dork']).'&ct=BR&ss_mkt=us&sp=1&l=wt-wt&vqd='.$numberForUrl.'&p=1&s='.$numPaginator; |
|
50
|
|
|
|
|
51
|
|
|
$this->output('Page '.$count."\n"); |
|
52
|
|
|
|
|
53
|
|
|
if ($this->commandData['virginProxies']) { |
|
54
|
|
|
$body = Utils::getBodyByVirginProxies($urlOfSearch, $this->listOfVirginProxies[$countProxyVirgin], $this->proxy); |
|
55
|
|
|
|
|
56
|
|
|
$arrLinks = $this->getLinks($body); |
|
57
|
|
|
|
|
58
|
|
|
if ($this->checkReturnError($body)) { |
|
59
|
|
|
$this->output("You has a problem with proxy, probaly you stress the engenier ...\n"); |
|
60
|
|
|
--$count; |
|
61
|
|
|
++$countError; |
|
62
|
|
|
if ($countError == 4) { |
|
63
|
|
|
$exit = true; |
|
64
|
|
|
} |
|
65
|
|
|
} else { |
|
66
|
|
|
$countError = 0; |
|
67
|
|
|
} |
|
68
|
|
|
|
|
69
|
|
|
//Check if next virgin proxy or repeat of 0 |
|
70
|
|
|
if ($countProxyVirgin == count($this->listOfVirginProxies) - 1) { |
|
71
|
|
|
$countProxyVirgin = 0; |
|
72
|
|
|
} else { |
|
73
|
|
|
++$countProxyVirgin; |
|
74
|
|
|
} |
|
75
|
|
|
} else { |
|
76
|
|
|
$body = Utils::getBody($urlOfSearch, $this->proxy); |
|
77
|
|
|
|
|
78
|
|
|
$arrLinks = $this->getLinks($body); |
|
79
|
|
|
} |
|
80
|
|
|
|
|
81
|
|
|
$this->output("\n".$urlOfSearch."\n"); |
|
82
|
|
|
|
|
83
|
|
|
$results = $this->sanitazeLinks($arrLinks); |
|
84
|
|
|
|
|
85
|
|
|
if ((count($results) == 0 and $body != 'repeat')) { |
|
|
|
|
|
|
86
|
|
|
$exit = true; |
|
87
|
|
|
} |
|
88
|
|
|
|
|
89
|
|
|
$resultFinal = array_merge($resultFinal, $results); |
|
90
|
|
|
|
|
91
|
|
|
++$count; |
|
92
|
|
|
} |
|
93
|
|
|
|
|
94
|
|
|
return $resultFinal; |
|
95
|
|
|
} |
|
96
|
|
|
|
|
97
|
|
|
private function getNumberForUrl() |
|
98
|
|
|
{ |
|
99
|
|
|
$firstUrlOfSearch = 'https://duckduckgo.com/?q='.urlencode($this->commandData['dork']).'&search_plus_one=form&ia=web'; |
|
100
|
|
|
$body = Utils::getBody($firstUrlOfSearch, $this->proxy); |
|
101
|
|
|
|
|
102
|
|
|
$validXmlrpc = preg_match("/','.*&vqd=(.*?)&/", $body, $matches, PREG_OFFSET_CAPTURE); |
|
|
|
|
|
|
103
|
|
|
|
|
104
|
|
|
if (isset($matches[1][0])) { |
|
105
|
|
|
return $matches[1][0]; |
|
106
|
|
|
} |
|
107
|
|
|
|
|
108
|
|
|
return false; |
|
109
|
|
|
} |
|
110
|
|
|
|
|
111
|
|
View Code Duplication |
public function sanitazeLinks($links = array()) |
|
|
|
|
|
|
112
|
|
|
{ |
|
113
|
|
|
$hrefs = array(); |
|
114
|
|
|
|
|
115
|
|
|
if (!empty($links)) { |
|
116
|
|
|
foreach ($links as $keyLink => $valueLink) { |
|
117
|
|
|
$validResultOfBlackList = Utils::checkBlacklist($valueLink); |
|
118
|
|
|
if (!$validResultOfBlackList and $valueLink) { |
|
|
|
|
|
|
119
|
|
|
$hrefs[] = $valueLink; |
|
120
|
|
|
} |
|
121
|
|
|
} |
|
122
|
|
|
|
|
123
|
|
|
$hrefs = array_unique($hrefs); |
|
124
|
|
|
} |
|
125
|
|
|
|
|
126
|
|
|
return $hrefs; |
|
127
|
|
|
} |
|
128
|
|
|
|
|
129
|
|
|
public function getLinks($body) |
|
130
|
|
|
{ |
|
131
|
|
|
$result = []; |
|
132
|
|
|
$validXmlrpc = preg_match("/DDG\.Data\.languages\.resultLanguages', (.*?)\);if/", $body, $matches, PREG_OFFSET_CAPTURE); |
|
|
|
|
|
|
133
|
|
|
|
|
134
|
|
|
if (isset($matches[1][0])) { |
|
135
|
|
|
$resultJson = json_decode($matches[1][0]); |
|
136
|
|
|
foreach ($resultJson as $resultsByLanguage) { |
|
137
|
|
|
foreach ($resultsByLanguage as $resultByLanguage) { |
|
138
|
|
|
$result[] = $resultByLanguage; |
|
139
|
|
|
} |
|
140
|
|
|
} |
|
141
|
|
|
} |
|
142
|
|
|
|
|
143
|
|
|
return $result; |
|
144
|
|
|
} |
|
145
|
|
|
|
|
146
|
|
|
public function checkReturnError($body) |
|
147
|
|
|
{ |
|
148
|
|
|
$valid = preg_match("/Dvar q=window\.location\.href\.indexOf/", $body, $matches, PREG_OFFSET_CAPTURE); |
|
149
|
|
|
|
|
150
|
|
|
if ($valid) { |
|
151
|
|
|
return true; |
|
152
|
|
|
} |
|
153
|
|
|
|
|
154
|
|
|
return false; |
|
155
|
|
|
} |
|
156
|
|
|
} |
|
157
|
|
|
|
aszone /
search-hacking
Loading history...
This check looks for variable assignements that are either overwritten by other assignments or where the variable is not used subsequently.
Both the
$myVarassignment in line 1 and the$higherassignment in line 2 are dead. The first because$myVaris never used and the second because$higheris always overwritten for every possible time line.