Result.php 6.31 KB
Newer Older
1
2
<?php

3
namespace App\Models;
4

5
6


7
8
9
class Result
{
	
Dominik Hebeler's avatar
Dominik Hebeler committed
10
	function __construct ( \SimpleXMLElement $provider, $titel, $link, $anzeigeLink , $descr, $gefVon, $sourceRank, $partnershop = false, $image = "" )
11
	{
12
		$this->titel = strip_tags(trim($titel));
13
14
		$this->link = trim($link);
		$this->anzeigeLink = trim($anzeigeLink);
15
16
		$this->descr = strip_tags(trim($descr));
		$this->descr = preg_replace("/\n+/si", " ", $this->descr);
17
18
19
20
21
22
		if( strlen($this->descr) > 250 )
		{
			$this->descr = wordwrap($this->descr, 250);
			$this->descr = substr($this->descr, 0, strpos($this->descr, "\n"));

		}
23
		$this->gefVon = trim($gefVon);
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
		$this->proxyLink = $this->generateProxyLink($this->link);
		$this->sourceRank = $sourceRank;
		if($this->sourceRank <= 0 || $this->sourceRank > 20)
			$this->sourceRank = 20;
		$this->sourceRank = 20 - $this->sourceRank;
		if(isset($provider["engineBoost"]))
		{
			$this->engineBoost = $provider["engineBoost"];
		}else
		{
			$this->engineBoost = 1;
		}

		$this->valid = true;
		$this->host = @parse_url($link, PHP_URL_HOST);
		$this->strippedHost = $this->getStrippedHost($this->anzeigeLink);
		$this->strippedDomain = $this->getStrippedDomain($this->strippedHost);
		$this->strippedLink = $this->getStrippedLink($this->anzeigeLink);
		$this->rank = 0;
		$this->partnershop = $partnershop;
Dominik Hebeler's avatar
Dominik Hebeler committed
44
		$this->image = $image;
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107

		#die($this->anzeigeLink . "\r\n" . $this->strippedHost);
	}

	public function rank (\App\MetaGer $metager)
	{

		$rank = 0;
		$rank += ($this->sourceRank * 0.02);

		#URL-Boost
		$link = $this->anzeigeLink;
		if(strpos($link, "http") !== 0)
		{
			$link = "http://" . $link;
		}
		$link = @parse_url($link, PHP_URL_HOST) . @parse_url($link, PHP_URL_PATH);
		$tmpLi = $link;
		$tmpEingabe = $metager->getQ();
		$count = 0;
		$tmpLink = "";

		$regex = [
			"/\s+/si",
			"/http:/si",
			"/https:/si",
			"/www\./si",
			"/\//si",
			"/\./si",
			"/-/si"
			];
		foreach($regex as $reg)
		{
			$link = preg_replace($regex, "", $link);
			$tmpEingabe = preg_replace($regex, "", $tmpEingabe);
		}
		#die($tmpLi . "<br>" . $link . "<br>" . $tmpEingabe . "<br><br>");
		foreach(str_split($tmpEingabe) as $char)
		{
			if(strpos(strtolower($tmpLink), strtolower($char)) >= 0)
			{
				$count++;
				$tmpLink = str_replace(urlencode($char), "", $tmpLink);
			}
			if(strlen($this->descr) > 80 && strlen($link) > 0)
			{
				$rank += $count /((strlen($link)) * 60);
			}
		}

		# Boost für Vorkommen der Suchwörter:
		$maxRank = 0.1;
		$tmpTitle = $this->titel;
		$tmpDescription = $this->descr;
		$isWithin = false;
		$tmpRank = 0;
		$tmpEingabe = $metager->getQ();
		$tmpEingabe = preg_replace("/\b\w{1,3}\b/si", "", $tmpEingabe);
		$tmpEingabe = preg_replace("/\s+/si", " ", $tmpEingabe);
		#die($tmpEingabe);
		foreach(explode(" ", trim($tmpEingabe)) as $el)
		{
			$el = preg_quote($el, "/");
108
			if(strlen($tmpTitle) > 0)
109
			{
110
111
112
113
114
115
				if(preg_match("/\b$el\b/si", $tmpTitle))
				{
					$tmpRank += .7 * .6 * $maxRank;
				}elseif (strpos($tmpTitle, $el) !== false) {
					$tmpRank += .3 * .6 * $maxRank;
				}
116
			}
117
			if( strlen($tmpDescription) > 0 )
118
			{
119
120
121
122
123
124
				if(preg_match("/\b$el\b/si", $tmpDescription))
				{
					$tmpRank += .7 * .4 * $maxRank;
				}elseif (strpos($tmpDescription, $el) !== false) {
					$tmpRank += .3 * .4 * $maxRank;
				}
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
			}
		}
		$tmpRank /= sizeof(explode(" ", trim($tmpEingabe))) * 10;
		$rank += $tmpRank;

		if($this->engineBoost > 0)
		{
			$rank *= floatval($this->engineBoost);
		}

		$this->rank = $rank;
	}

	public function getRank ()
	{
		return $this->rank;
	}

	public function isValid (\App\MetaGer $metager)
	{
		# Zunächst die persönlich ( über URL-Parameter ) definierten Blacklists:
		if(in_array($this->strippedHost, $metager->getUserHostBlacklist())
			|| in_array($this->strippedDomain, $metager->getUserDomainBlacklist()))
			return false;
		
		# Jetzt unsere URL und Domain Blacklist
		if($this->strippedHost !== "" && (in_array($this->strippedHost, $metager->getDomainBlacklist()) || in_array($this->strippedLink, $metager->getUrlBlacklist())))
		{
			return false;
		}

		$text = $this->titel . " " . $this->descr;

		if($metager->getLang() !== "all")
		{
			$result = $metager->getLanguageDetect()->detect($text, 1);
			$lang = "";
			foreach($result as $key => $value)
			{
				$lang = $key;
			}

			if($lang !== "" && $lang !== $metager->getLang())
				return false;
		}

		# Wir wenden die Stoppwortsuche an und schmeißen entsprechende Ergebnisse raus:
		foreach($metager->getStopWords() as $stopWord)
		{
			if(stripos($text, $stopWord) !== false)
			{
				return false;
			}
		}

		# Abschließend noch 2 Überprüfungen. Einmal den Host filter, der Sicherstellt, dass von jedem Host maximal 3 Links angezeigt werden
		# und dann noch den Dublettefilter, der sicher stellt, dass wir nach Möglichkeit keinen Link doppelt in der Ergebnisliste haben
		# Diese Überprüfung führen wir unter bestimmten Bedingungen nicht durch:
		if($metager->getSite() === "" &&
			strpos($this->strippedHost, "ncbi.nlm.nih.gov") === false &&
			strpos($this->strippedHost, "twitter.com") === false &&
			strpos($this->strippedHost, "www.ladenpreis.net") === false &&
			strpos($this->strippedHost, "ncbi.nlm.nih.gov") === false &&
			strpos($this->strippedHost, "www.onenewspage.com") === false)
		{
			$count = $metager->getHostCount($this->strippedHost);
			if($count >= 3)
			{
				return false;
			}
		}

		# Unabhängig davon unser Dublettenfilter:
		if($metager->addLink($this->strippedLink))
		{
			$metager->addHostCount($this->strippedHost);
			return true;
		}else
		{
			return false;
		}
	}

208
	private function getStrippedHost ($link)
209
210
211
212
213
214
215
	{
		if(strpos($link, "http") !== 0)
			$link = "http://" . $link;
		$link = @parse_url($link, PHP_URL_HOST);
		$link = preg_replace("/^www\./si", "", $link);
		return $link;
	}
216
	private function getStrippedLink ($link)
217
218
219
220
221
222
223
224
	{
		if(strpos($link, "http") !== 0)
			$link = "http://" . $link;
		$host = $this->strippedHost;
		$path = @parse_url($link , PHP_URL_PATH);
		return $host . $path;
	}

225
	private function getStrippedDomain ($link)
226
227
228
229
230
231
232
233
234
235
	{
		if(preg_match("/([^\.]*\.[^\.]*)$/si", $link, $match))
		{
			return $match[1];
		}else
		{
			return $link;
		}		
	}

236
	private function generateProxyLink ($link)
237
238
239
240
241
242
243
244
	{
		if(!$link)
			return "";
		$tmp = $link;
		$tmp = preg_replace("/\r?\n$/s", "", $tmp);
		$tmp = preg_replace("#^([\w+.-]+)://#s", "$1/", $tmp);
		return "https://proxy.suma-ev.de/cgi-bin/nph-proxy.cgi/en/I0/" . $tmp;
		
245
246
	}
}