HtmlSplitter::split() - Code Metrics - Inspection of "Apply fixes from StyleCI" - algolia/scout-extended - Measure and Improve Code Quality continuously with Scrutinizer

Completed

Pull Request — feat/html-splitter (#174)

by Nuno

created 2019-04-08 12:31 UTC

HtmlSplitter::split() A

↳ Parent: HtmlSplitter

Complexity

Conditions	2
Paths	2

Size

Total Lines	21
Code Lines	15

Duplication

Lines	0
Ratio	0 %

Code Coverage

Tests	16
CRAP Score	2

Importance

Changes

Metric	Value
cc	2
eloc	15
nc	2
nop	2
dl	0
loc	21
ccs	16
cts	16
cp	1
crap	2
rs	9.7666
c	0
b	0
f	0

<?php

declare(strict_types=1);

/**
 * This file is part of Scout Extended.
 *
 * (c) Algolia Team <[email protected]>
 *
 *  For the full copyright and license information, please view the LICENSE
 *  file that was distributed with this source code.
 */

namespace Algolia\ScoutExtended\Splitters;

use DOMXPath;
use DOMDocument;
use Algolia\ScoutExtended\Contracts\SplitterContract;

class HtmlSplitter implements SplitterContract
{
    /**
     * The list of html tags.
     *
     * @var string[]
     */
    protected $acceptedNodes = [
        'h1',
        'h2',
        'h3',
        'h4',
        'h5',
        'h6',
        'p',
    ];

    /**
     * Creates a new instance of the class.
     *
     * @param array $acceptedNodes
     *
     * @return void
     */
    public function __construct(array $acceptedNodes = null)
    {
        if ($acceptedNodes !== null) {
            $this->acceptedNodes = $acceptedNodes;
        }
    }

    /**
     * Find it's value in $acceptedNodes.
     *
     * @param array $object
     *
     * @return int
     */
    public function findValue($object): int
    {
        return array_search((key($object)), $this->acceptedNodes);

    }

    /**
     * Add object to queue.
     *
     * @param array $object
     * @param array $queue
     *
     * @return array
     */
    public function addObjectToQueue($object, $queue): array
    {
        if (count($queue) == 0) {
            $queue[] = $object;

            return $queue;
        } else {
            if ($this->findValue($object) > $this->findValue(end($queue))) {
                $queue[] = $object;

                return $queue;
            } else {
                array_pop($queue);

                return $this->addObjectToQueue($object, $queue);
            }
        }
    }

    /**
     * Importance formula.
     * Give integer from tags ranking.
     *
     * @param \DOMElement $node
     * @param array $queue
     *
     * @return int
     */
    public function importanceWeight($node, $queue): int
    {
        if ($node->nodeName == 'p') {
            if (empty(end($queue))) {
                return 0;
            }

            return (count($this->acceptedNodes) - 1) + (array_search(key(end($queue)), $this->acceptedNodes));
        }

        return array_search($node->nodeName, $this->acceptedNodes);

    }

    /**
     * Clean Records to have a correct format.
     *
     *
     * @param array $records
     *
     * @return array
     */
    public function cleanRecords($records): array
    {
        $newRecords = [];
        foreach ($records as $record) {
            foreach ($record as $r) {
                foreach ($r as $res => $values) {
                    $newRecord[$res] = $values;
                    if ($res == 'importance') {
                        $newRecords[] = $newRecord;
                        $newRecord = [];
                    }
                }
            }
        }

        return $newRecords;
    }

    /**
     * Acts a static factory.
     *
     * @param  string|array $tags
     *
     * @return static
     */
    public static function by($tags)
    {
        return new static((array) $tags);
    }

    /**
     * Splits the given value.
     *
     * @param  object $searchable
     * @param  string $value
     *
     * @return array
     */
    public function split($searchable, $value): array
    {
        $dom = new DOMDocument();
        $dom->loadHTML($value);
        $xpath = new DOMXpath($dom);
        $queue = [];
        $xpathQuery = '//'.implode(' | //', $this->acceptedNodes);
        $nodes = $xpath->query($xpathQuery);

        foreach ($nodes as $node) {
            $object = [$node->nodeName => $node->textContent];
            $importance = $this->importanceWeight($node, $queue);
            $queue = $this->addObjectToQueue($object, $queue);
            $cloneQueue = $queue;
            $cloneQueue[] = ['importance' => $importance];
            $records[] = $cloneQueue;
        }

        $records = $this->cleanRecords($records);


        return $records;
    }
}


1		<?php
2
3		declare(strict_types=1);
4
5		/**
6		* This file is part of Scout Extended.
7		*
8		* (c) Algolia Team <[email protected]>
9		*
10		* For the full copyright and license information, please view the LICENSE
11		* file that was distributed with this source code.
12		*/
13
14		namespace Algolia\ScoutExtended\Splitters;
15
16		use DOMXPath;
17		use DOMDocument;
18		use Algolia\ScoutExtended\Contracts\SplitterContract;
19
20		class HtmlSplitter implements SplitterContract
21		{
22		/**
23		* The list of html tags.
24		*
25		* @var string[]
26		*/
27		protected $acceptedNodes = [
28		'h1',
29		'h2',
30		'h3',
31		'h4',
32		'h5',
33		'h6',
34		'p',
35		];
36
37		/**
38		* Creates a new instance of the class.
39		*
40		* @param array $acceptedNodes
41		*
42		* @return void
43		*/
44	3	public function __construct(array $acceptedNodes = null)
45		{
46	3	if ($acceptedNodes !== null) {
47	2	$this->acceptedNodes = $acceptedNodes;
48		}
49	3	}
50
51		/**
52		* Find it's value in $acceptedNodes.
53		*
54		* @param array $object
55		*
56		* @return int
57		*/
58	3	public function findValue($object): int
59		{
60	3	return array_search((key($object)), $this->acceptedNodes);
		0 ignored issues – show Bug Best Practice introduced 2019-04-08 12:34 UTC by Report Bug Copy Issue Report The expression `return array_search(key(..., $this->acceptedNodes)` could return the type `false\|string` which is incompatible with the type-hinted return `integer`. Consider adding an additional type-check to rule them out. Loading history...
61		}
62
63		/**
64		* Add object to queue.
65		*
66		* @param array $object
67		* @param array $queue
68		*
69		* @return array
70		*/
71	3	public function addObjectToQueue($object, $queue): array
72		{
73	3	if (count($queue) == 0) {
74	3	$queue[] = $object;
75
76	3	return $queue;
77		} else {
78	3	if ($this->findValue($object) > $this->findValue(end($queue))) {
79	1	$queue[] = $object;
80
81	1	return $queue;
82		} else {
83	3	array_pop($queue);
84
85	3	return $this->addObjectToQueue($object, $queue);
86		}
87		}
88		}
89
90		/**
91		* Importance formula.
92		* Give integer from tags ranking.
93		*
94		* @param \DOMElement $node
95		* @param array $queue
96		*
97		* @return int
98		*/
99	3	public function importanceWeight($node, $queue): int
100		{
101	3	if ($node->nodeName == 'p') {
102	1	if (empty(end($queue))) {
103		return 0;
104		}
105
106	1	return (count($this->acceptedNodes) - 1) + (array_search(key(end($queue)), $this->acceptedNodes));
107		}
108
109	3	return array_search($node->nodeName, $this->acceptedNodes);
		0 ignored issues – show Bug Best Practice introduced 2019-04-08 12:34 UTC by Report Bug Copy Issue Report The expression `return array_search($nod..., $this->acceptedNodes)` could return the type `false\|string` which is incompatible with the type-hinted return `integer`. Consider adding an additional type-check to rule them out. Loading history...
110		}
111
112		/**
113		* Clean Records to have a correct format.
114		*
115		*
116		* @param array $records
117		*
118		* @return array
119		*/
120	3	public function cleanRecords($records): array
121		{
122	3	$newRecords = [];
123	3	foreach ($records as $record) {
124	3	foreach ($record as $r) {
125	3	foreach ($r as $res => $values) {
126	3	$newRecord[$res] = $values;
127	3	if ($res == 'importance') {
128	3	$newRecords[] = $newRecord;
129	3	$newRecord = [];
130		}
131		}
132		}
133		}
134
135	3	return $newRecords;
136		}
137
138		/**
139		* Acts a static factory.
140		*
141		* @param string\|array $tags
142		*
143		* @return static
144		*/
145	2	public static function by($tags)
146		{
147	2	return new static((array) $tags);
148		}
149
150		/**
151		* Splits the given value.
152		*
153		* @param object $searchable
154		* @param string $value
155		*
156		* @return array
157		*/
158	3	public function split($searchable, $value): array
159		{
160	3	$dom = new DOMDocument();
161	3	$dom->loadHTML($value);
162	3	$xpath = new DOMXpath($dom);
163	3	$queue = [];
164	3	$xpathQuery = '//'.implode(' \| //', $this->acceptedNodes);
165	3	$nodes = $xpath->query($xpathQuery);
166
167	3	foreach ($nodes as $node) {
168	3	$object = [$node->nodeName => $node->textContent];
169	3	$importance = $this->importanceWeight($node, $queue);
170	3	$queue = $this->addObjectToQueue($object, $queue);
171	3	$cloneQueue = $queue;
172	3	$cloneQueue[] = ['importance' => $importance];
173	3	$records[] = $cloneQueue;
174		}
175
176	3	$records = $this->cleanRecords($records);
		0 ignored issues – show Comprehensibility Best Practice introduced 2019-04-08 12:34 UTC by Report Bug Copy Issue Report The variable `$records` seems to be defined by a `foreach` iteration on line `167`. Are you sure the iterator is never empty, otherwise this variable is not defined? Loading history...
177
178	3	return $records;
179		}
180		}
181

algolia / scout-extended

Pull Request — feat/html-splitter (#174)

HtmlSplitter::split() A

Complexity

Size

Duplication

Code Coverage

Importance

Duplication Side-by-Side

Filter issues like