ScraperUtil::convertDomNodeToText() - Code Metrics - Inspection of "updated travis badge" - Ne-Lexa/google-play-scraper - Measure and Improve Code Quality continuously with Scrutinizer

Passed

Push — master ( 80ad96...45165d )

by Alexey

created 2019-05-27 19:20 UTC

ScraperUtil::convertDomNodeToText() C

↳ Parent: ScraperUtil

Complexity

Conditions	15
Paths	25

Size

Total Lines	34
Code Lines	26

Duplication

Lines	0
Ratio	0 %

Importance

Changes

Metric	Value
eloc	26
dl	0
loc	34
rs	5.9166
c	0
b	0
f	0
cc	15
nc	25
nop	1

How to fix Complexity

<?php
declare(strict_types=1);

namespace Nelexa\GPlay\Util;

class ScraperUtil
{
    /**
     * @param string $html
     * @return array
     */
    public static function extractScriptData(string $html): array
    {
        $scripts = [];
        if (preg_match_all('/>AF_initDataCallback[\s\S]*?<\/script/', $html, $matches)) {
            $scripts = array_reduce($matches[0], static function ($carry, $item) {
                if (
                    preg_match("/(ds:.*?)'/", $item, $keyMatch) &&
                    preg_match('/return ([\s\S]*?)}}\);<\//', $item, $valueMatch)
                ) {
                    $carry[$keyMatch[1]] = \GuzzleHttp\json_decode($valueMatch[1], true);
                }
                return $carry;
            }, $scripts);
        }
        return $scripts;
    }

    /**
     * @param string $html
     * @return string
     */
    public static function html2text(string $html): string
    {
        $doc = new \DOMDocument();
        $internalErrors = libxml_use_internal_errors(true);
        if (!$doc->loadHTML('<?xml encoding="utf-8" ?>' . $html)) {
            throw new \RuntimeException('error load html: ' . $html);
        }
        libxml_use_internal_errors($internalErrors);
        $text = self::convertDomNodeToText($doc);
        $text = preg_replace('/\n{3,}/', "\n\n", trim($text));
        return trim($text);
    }

    /**
     * @param \DOMNode $node
     * @return string
     */
    private static function convertDomNodeToText(\DOMNode $node): string
    {
        if ($node instanceof \DOMText) {
            $text = preg_replace('/\s+/', ' ', $node->wholeText);
        } else {
            $text = '';
            if ($node->childNodes !== null) {
                foreach ($node->childNodes as $childNode) {
                    $text .= self::convertDomNodeToText($childNode);
                }
            }

            switch ($node->nodeName) {
                case 'h1':
                case 'h2':
                case 'h3':
                case 'h4':
                case 'h5':
                case 'h6':
                case 'p':
                case 'ul':
                case 'div':
                    $text = "\n\n" . $text . "\n\n";
                    break;
                case 'li':
                    $text = '- ' . $text . "\n";
                    break;
                case 'br':
                    $text .= "\n";
                    break;
            }
        }

        return $text;
    }
}


1			<?php
2			declare(strict_types=1);
3
4			namespace Nelexa\GPlay\Util;
5
6			class ScraperUtil
7			{
8			/**
9			* @param string $html
10			* @return array
11			*/
12			public static function extractScriptData(string $html): array
13			{
14			$scripts = [];
15			if (preg_match_all('/>AF_initDataCallback[\s\S]*?<\/script/', $html, $matches)) {
16			$scripts = array_reduce($matches[0], static function ($carry, $item) {
17			if (
18			preg_match("/(ds:.*?)'/", $item, $keyMatch) &&
19			preg_match('/return ([\s\S]*?)}}\);<\//', $item, $valueMatch)
20			) {
21			$carry[$keyMatch[1]] = \GuzzleHttp\json_decode($valueMatch[1], true);
22			}
23			return $carry;
24			}, $scripts);
25			}
26			return $scripts;
27			}
28
29			/**
30			* @param string $html
31			* @return string
32			*/
33			public static function html2text(string $html): string
34			{
35			$doc = new \DOMDocument();
36			$internalErrors = libxml_use_internal_errors(true);
37			if (!$doc->loadHTML('<?xml encoding="utf-8" ?>' . $html)) {
38			throw new \RuntimeException('error load html: ' . $html);
39			}
40			libxml_use_internal_errors($internalErrors);
41			$text = self::convertDomNodeToText($doc);
42			$text = preg_replace('/\n{3,}/', "\n\n", trim($text));
43			return trim($text);
44			}
45
46			/**
47			* @param \DOMNode $node
48			* @return string
49			*/
50			private static function convertDomNodeToText(\DOMNode $node): string
51			{
52			if ($node instanceof \DOMText) {
53			$text = preg_replace('/\s+/', ' ', $node->wholeText);
54			} else {
55			$text = '';
56			if ($node->childNodes !== null) {
57			foreach ($node->childNodes as $childNode) {
58			$text .= self::convertDomNodeToText($childNode);
59			}
60			}
61
62			switch ($node->nodeName) {
63			case 'h1':
64			case 'h2':
65			case 'h3':
66			case 'h4':
67			case 'h5':
68			case 'h6':
69			case 'p':
70			case 'ul':
71			case 'div':
72			$text = "\n\n" . $text . "\n\n";
73			break;
74			case 'li':
75			$text = '- ' . $text . "\n";
76			break;
77			case 'br':
78			$text .= "\n";
79			break;
80			}
81			}
82
83			return $text;
84			}
85			}
86

Ne-Lexa / google-play-scraper

Push — master ( 80ad96...45165d )

ScraperUtil::convertDomNodeToText() C

Complexity

Size

Duplication

Importance

How to fix Complexity

Long Method

Duplication Side-by-Side

Filter issues like