AdCrawler::setUrlParser() - Code Metrics - Inspection of "Merge pull request #16 from rfussien/develop" - rfussien/leboncoin-crawler - Measure and Improve Code Quality continuously with Scrutinizer

Completed

Push — master ( 00c88c...56f5c3 )

by Rémi

created 2017-04-06 19:54 UTC

AdCrawler::setUrlParser() A

↳ Parent: AdCrawler

Complexity

Conditions	1
Paths	1

Size

Total Lines	4
Code Lines	2

Duplication

Lines	0
Ratio	0 %

Code Coverage

Tests	3
CRAP Score	1

Importance

Changes

Metric	Value
c	0
b	0
f	0
dl	0
loc	4
ccs	3
cts	3
cp	1
rs	10
cc	1
eloc	2
nc	1
nop	1
crap	1

<?php

namespace Lbc\Crawler;

use Lbc\Filter\CitySanitizer;
use Lbc\Filter\CpSanitizer;
use Lbc\Filter\DefaultSanitizer;
use Lbc\Filter\KeySanitizer;
use Lbc\Filter\PriceSanitizer;
use Lbc\Parser\AdUrlParser;
use League\Uri\Schemes\Http;
use Symfony\Component\DomCrawler\Crawler;

/**
 * Class AdCrawler
 * @package Lbc\Crawler
 */
class AdCrawler extends CrawlerAbstract
{
    /**
     * @param $url
     * @return AdUrlParser
     */
    protected function setUrlParser($url)
    {
        $this->url = new AdUrlParser($url);
    }

    /**
     * Return a full ad information
     *
     * @return array
     */
    public function getAll()
    {
        return array_merge(
            [
                'id'       => $this->getUrlParser()->getId(),
class A
{
    public function foo() { }
}

class B extends A
{
    public function bar() { }
}

/**
 * @param A|B $x
 */
function someFunction($x)
{
    $x->foo(); // This call is fine as the method exists in A and B.
    $x->bar(); // This method only exists in B and might cause an error.
}
                'category' => $this->getUrlParser()->getCategory(),
            ],
            $this->getPictures(),
            $this->getProperties(),
            $this->getDescription()
        );
    }

    /**
     * Return an array with the Thumbs pictures url
     *
     * @param Crawler $node
     * @return array
     */
    public function getPictures(Crawler $node = null)
    {
        $node = $node ?: $this->node;

        $images = [];
        $images_thumbs = [];

        $node
            ->filter('.adview_main script')
            ->each(function (Crawler $crawler) use (&$images, &$images_thumbs) {
                preg_match_all(
                    '#//img.+.leboncoin.fr/.*\.jpg#',
                    $crawler->html(),
                    $matches
                );

                if (count($matches[0]) > 0) {
                    foreach ($matches[0] as $image) {
                        if (preg_match('/thumb/', $image)) {
                            array_push(
                                $images_thumbs,
                                (string)Http::createFromString($image)
                                    ->withScheme($this->sheme)
                            );
                        } else {
                            array_push(
                                $images,
                                (string)Http::createFromString($image)
                                    ->withScheme($this->sheme)
                            );
                        }
                    }
                }
            });

        return [
            'images'        => $images,
            'images_thumbs' => $images_thumbs,
        ];
    }

    /**
     * Return the common information (price, cp, city)
     *
     * @param Crawler $node
     *
     * @return array
     */
    public function getProperties(Crawler $node = null)
    {
        $node = $node ?: $this->node;

        $properties = [];

        $properties['title'] = DefaultSanitizer::clean(
            $this->node->filter('h1')->text()
        );

        $node->filter('h2')
            ->each(function (Crawler $crawler) use (&$properties) {
                $properties = array_merge(
                    $properties,
                    $this->sanitize(
                        $crawler->filter('.property')->text(),
                        $crawler->filter('.value')->text()
                    )
                );
            });

        return ['properties' => $properties];
    }

    /**
     * Return the description
     *
     * @param Crawler $node
     * @return string
     */
    public function getDescription(Crawler $node = null)

    {
        return ['description' => $this->node->filter("p#description")->text()];
    }

    /**
     * Transform the properties name into a snake_case string
     *
     * @param string $value
     * @return string
     */
    private function sanitize($key, $value)
    {
        $key = KeySanitizer::clean($key);

        switch ($key) {
            case 'prix':
                return ['price' => PriceSanitizer::clean($value)];
                break;
switch ($x) {
    case 1:
        return 'foo';
        break; // This break is not necessary and can be left off.
}
            case 'ville':
                return [
                    'city' => CitySanitizer::clean($value),
                    'cp'   => CpSanitizer::clean($value),
                ];
                break;
switch ($x) {
    case 1:
        return 'foo';
        break; // This break is not necessary and can be left off.
}
            default:
                return [$key => DefaultSanitizer::clean($value)];
        }
    }
}


1		<?php
2
3		namespace Lbc\Crawler;
4
5		use Lbc\Filter\CitySanitizer;
6		use Lbc\Filter\CpSanitizer;
7		use Lbc\Filter\DefaultSanitizer;
8		use Lbc\Filter\KeySanitizer;
9		use Lbc\Filter\PriceSanitizer;
10		use Lbc\Parser\AdUrlParser;
11		use League\Uri\Schemes\Http;
12		use Symfony\Component\DomCrawler\Crawler;
13
14		/**
15		* Class AdCrawler
16		* @package Lbc\Crawler
17		*/
18		class AdCrawler extends CrawlerAbstract
19		{
20		/**
21		* @param $url
22		* @return AdUrlParser
23		*/
24	14	protected function setUrlParser($url)
25		{
26	14	$this->url = new AdUrlParser($url);
27	14	}
28
29		/**
30		* Return a full ad information
31		*
32		* @return array
33		*/
34	14	public function getAll()
35		{
36	14	return array_merge(
37		[
38	6	'id' => $this->getUrlParser()->getId(),
		0 ignored issues – show Bug introduced 2017-04-06 19:53 UTC by Report Bug Copy Issue Report The method `getId` does only exist in `Lbc\Parser\AdUrlParser`, but not in `Lbc\Parser\SearchResultUrlParser`. It seems like the method you are trying to call exists only in some of the possible types. Let’s take a look at an example: class A { public function foo() { } } class B extends A { public function bar() { } } /** * @param A\|B $x / function someFunction($x) { $x->foo(); // This call is fine as the method exists in A and B. $x->bar(); // This method only exists in B and might cause an error. } Available Fixes Add an additional type-check: /* * @param A\|B $x / function someFunction($x) { $x->foo(); if ($x instanceof B) { $x->bar(); } } Only allow a single type to be passed if the variable comes from a parameter: function someFunction(B $x) { /* ... */ } Loading history...
39	6	'category' => $this->getUrlParser()->getCategory(),
40	6	],
41	6	$this->getPictures(),
42	6	$this->getProperties(),
43	6	$this->getDescription()
44	6	);
45		}
46
47		/**
48		* Return an array with the Thumbs pictures url
49		*
50		* @param Crawler $node
51		* @return array
52		*/
53	8	public function getPictures(Crawler $node = null)
54		{
55	8	$node = $node ?: $this->node;
56
57	8	$images = [];
58	8	$images_thumbs = [];
59
60		$node
61	8	->filter('.adview_main script')
62		->each(function (Crawler $crawler) use (&$images, &$images_thumbs) {
63	8	preg_match_all(
64	8	'#//img.+.leboncoin.fr/.*\.jpg#',
65	8	$crawler->html(),
66		$matches
67	8	);
68
69	8	if (count($matches[0]) > 0) {
70	8	foreach ($matches[0] as $image) {
71	8	if (preg_match('/thumb/', $image)) {
72	8	array_push(
73	8	$images_thumbs,
74	8	(string)Http::createFromString($image)
75	8	->withScheme($this->sheme)
76	8	);
77	8	} else {
78	8	array_push(
79	8	$images,
80	8	(string)Http::createFromString($image)
81	8	->withScheme($this->sheme)
82	8	);
83		}
84	8	}
85	8	}
86	8	});
87
88		return [
89	8	'images' => $images,
90	8	'images_thumbs' => $images_thumbs,
91	8	];
92		}
93
94		/**
95		* Return the common information (price, cp, city)
96		*
97		* @param Crawler $node
98		*
99		* @return array
100		*/
101	8	public function getProperties(Crawler $node = null)
102		{
103	8	$node = $node ?: $this->node;
104
105	8	$properties = [];
106
107	8	$properties['title'] = DefaultSanitizer::clean(
108	8	$this->node->filter('h1')->text()
109	8	);
110
111	8	$node->filter('h2')
112	8	->each(function (Crawler $crawler) use (&$properties) {
113	8	$properties = array_merge(
114	8	$properties,
115	8	$this->sanitize(
116	8	$crawler->filter('.property')->text(),
117	8	$crawler->filter('.value')->text()
118	8	)
119	8	);
120	8	});
121
122	8	return ['properties' => $properties];
123		}
124
125		/**
126		* Return the description
127		*
128		* @param Crawler $node
129		* @return string
130		*/
131	8	public function getDescription(Crawler $node = null)
		0 ignored issues – show Unused Code introduced 2017-04-06 19:53 UTC by Report Bug Copy Issue Report The parameter `$node` is not used and could be removed. This check looks from parameters that have been defined for a function or method, but which are not used in the method body. Loading history...
132		{
133	8	return ['description' => $this->node->filter("p#description")->text()];
134		}
135
136		/**
137		* Transform the properties name into a snake_case string
138		*
139		* @param string $value
140		* @return string
141		*/
142	8	private function sanitize($key, $value)
143		{
144	8	$key = KeySanitizer::clean($key);
145
146		switch ($key) {
147	8	case 'prix':
148	8	return ['price' => PriceSanitizer::clean($value)];
149		break;
		0 ignored issues – show Unused Code introduced 2017-04-06 19:53 UTC by Report Bug Copy Issue Report `break` is not strictly necessary here and could be removed. The `break` statement is not necessary if it is preceded for example by a `return` statement: switch ($x) { case 1: return 'foo'; break; // This break is not necessary and can be left off. } If you would like to keep this construct to be consistent with other `case` statements, you can safely mark this issue as a false-positive. Loading history...
150	8	case 'ville':
151		return [
152	8	'city' => CitySanitizer::clean($value),
153	8	'cp' => CpSanitizer::clean($value),
154	8	];
155		break;
		0 ignored issues – show Unused Code introduced 2017-04-06 19:53 UTC by Report Bug Copy Issue Report `break` is not strictly necessary here and could be removed. The `break` statement is not necessary if it is preceded for example by a `return` statement: switch ($x) { case 1: return 'foo'; break; // This break is not necessary and can be left off. } If you would like to keep this construct to be consistent with other `case` statements, you can safely mark this issue as a false-positive. Loading history...
156	8	default:
157	8	return [$key => DefaultSanitizer::clean($value)];
158	8	}
159		}
160		}
161

rfussien / leboncoin-crawler

Push — master ( 00c88c...56f5c3 )

AdCrawler::setUrlParser() A

Complexity

Size

Duplication

Code Coverage

Importance

Available Fixes

Duplication Side-by-Side

Filter issues like