Conditions | 6 |
Paths | 6 |
Total Lines | 30 |
Code Lines | 18 |
Lines | 0 |
Ratio | 0 % |
Changes | 2 | ||
Bugs | 1 | Features | 0 |
1 | <?php declare(strict_types=1); |
||
24 | private function isHighLinkDensity(Element $node, float $limit = 1.0): bool { |
||
25 | $links = $node->find('a, [onclick]'); |
||
26 | |||
27 | if ($links->count() == 0) { |
||
28 | return false; |
||
29 | } |
||
30 | |||
31 | $words = preg_split('@[\s]+@iu', $node->text(), -1, PREG_SPLIT_NO_EMPTY); |
||
32 | |||
33 | if (!is_array($words) || empty($words)) { |
||
|
|||
34 | return false; |
||
35 | } |
||
36 | |||
37 | $sb = []; |
||
38 | foreach ($links as $link) { |
||
39 | $sb[] = Helper::textNormalise($link->text()); |
||
40 | } |
||
41 | |||
42 | $linkText = implode('', $sb); |
||
43 | $linkWords = explode(' ', $linkText); |
||
44 | $numberOfLinkWords = count($linkWords); |
||
45 | $numberOfLinks = $links->count(); |
||
46 | $linkDivisor = $numberOfLinkWords / count($words); |
||
47 | $score = $linkDivisor * $numberOfLinks; |
||
48 | |||
49 | if ($score >= $limit) { |
||
50 | return true; |
||
51 | } |
||
52 | |||
53 | return false; |
||
54 | } |
||
56 |