RFC4180Parser::extractEnclosedFieldContent() - Code Metrics - Inspection of "change parser name" - thephpleague/csv - Measure and Improve Code Quality continuously with Scrutinizer

Completed

Push — master ( 778a24...7e8982 )

by ignace nyamagana

created 2018-09-24 06:28 UTC

RFC4180Parser::extractEnclosedFieldContent() B

↳ Parent: RFC4180Parser

Complexity

Conditions	7
Paths	24

Size

Total Lines

Duplication

Lines	0
Ratio	0 %

Code Coverage

Tests	20
CRAP Score	7

Importance

Changes

Metric	Value
cc	7
nc	24
nop	1
dl	0
loc	35
ccs	20
cts	20
cp	1
crap	7
rs	8.4266
c	0
b	0
f	0

<?php

/**
 * League.Csv (https://csv.thephpleague.com).
 *
 * @author  Ignace Nyamagana Butera <[email protected]>
 * @license https://github.com/thephpleague/csv/blob/master/LICENSE (MIT License)
 * @version 9.2.0
 * @link    https://github.com/thephpleague/csv
 *
 * For the full copyright and license information, please view the LICENSE
 * file that was distributed with this source code.
 */

declare(strict_types=1);

namespace League\Csv;

use IteratorAggregate;
use SplFileObject;
use TypeError;
use function explode;
use function get_class;
use function gettype;
use function in_array;
use function is_object;
use function rtrim;
use function sprintf;
use function str_replace;
use function strlen;
use function substr;
use function trim;

/**
 * A RFC4180 Compliant Parser in Pure PHP.
 *
 * @see https://php.net/manual/en/function.fgetcsv.php
 * @see https://php.net/manual/en/function.fgets.php
 * @see https://tools.ietf.org/html/rfc4180
 * @see http://edoceo.com/utilitas/csv-file-format
 *
 * @internal used internally to produce RFC4180 compliant records
 */
final class RFC4180Parser implements IteratorAggregate
{
    /**
     * @internal
     */
    const FIELD_BREAKS = [false, '', "\r\n", "\n", "\r"];

    /**
     * @var SplFileObject|Stream
     */
    private $document;

    /**
     * @var string
     */
    private $delimiter;

    /**
     * @var string
     */
    private $enclosure;

    /**
     * @var string
     */
    private $trim_mask;

    /**
     * New instance.
     *
     * @param SplFileObject|Stream $document
     */
    public function __construct($document, string $delimiter = ',', string $enclosure = '"')
    {
        if (!$document instanceof Stream && !$document instanceof SplFileObject) {
            throw new TypeError(sprintf(
                'Expected a %s or an SplFileObject object, % given',
                Stream::class,
                is_object($document) ? get_class($document) : gettype($document)
            ));
        }

        if (1 !== strlen($delimiter)) {
            throw new Exception(sprintf('%s() expects delimiter to be a single character %s given', __METHOD__, $delimiter));
        }

        if (1 !== strlen($enclosure)) {
            throw new Exception(sprintf('%s() expects enclosure to be a single character %s given', __METHOD__, $enclosure));
        }

        $this->document = $document;
        $this->delimiter = $delimiter;
        $this->enclosure = $enclosure;
        $this->trim_mask = str_replace([$this->delimiter, $this->enclosure], '', " \t\0\x0B");
    }

    /**
     * @inheritdoc
     *
     * Converts the stream into a CSV record iterator by extracting records one by one
     *
     * The returned record array is similar to the returned value of fgetcsv
     *
     * - If the line is empty the record will be an array with a single value equals to null
     * - Otherwise the array contains strings.
     */
    public function getIterator()
    {
        $this->document->setFlags(0);
        $this->document->rewind();
        do {
            $record = [];
            $line = $this->document->fgets();
            do {
                $method = 'extractFieldContent';
                if (($line[0] ?? '') === $this->enclosure) {
                    $method = 'extractEnclosedFieldContent';
                }
                $record[] = $this->$method($line);
            } while (false !== $line);

            yield $record;
        } while ($this->document->valid());
    }

    /**
     * Extract field without enclosure as per RFC4180.
     *
     * - Leading and trailing whitespaces must be removed.
     * - trailing line-breaks must be removed.
     *
     * @param bool|string $line
     *
     * @return null|string
     */
    private function extractFieldContent(&$line)
    {
        if (in_array($line, self::FIELD_BREAKS, true)) {
            $line = false;

            return null;
        }

        list($content, $line) = explode($this->delimiter, $line, 2) + [1 => false];
        if (false === $line) {
            return trim(rtrim($content, "\r\n"), $this->trim_mask);
        }

        return trim($content, $this->trim_mask);
    }

    /**
     * Extract field with enclosure as per RFC4180.
     *
     * - Field content can spread on multiple document lines.
     * - Content inside enclosure must be preserved.
     * - Double enclosure sequence must be replaced by single enclosure character.
     * - Trailing line break must be removed if they are not part of the field content.
     * - Invalid field do not throw as per fgetcsv behavior.
     *
     * @param bool|string $line
     */
    private function extractEnclosedFieldContent(&$line): string
    {
        if (($line[0] ?? '') === $this->enclosure) {
            $line = substr($line, 1);
        }

        $content = '';
        while (false !== $line) {
            list($buffer, $line) = explode($this->enclosure, $line, 2) + [1 => false];
            $content .= $buffer;
            if (false !== $line) {
                break;
            }
            $line = $this->document->fgets();
        }

        if (in_array($line, self::FIELD_BREAKS, true)) {
            $line = false;

            return rtrim($content, "\r\n");
        }

        $char = $line[0] ?? '';
        if ($char === $this->delimiter) {
            $line = substr($line, 1);

            return $content;
        }

        if ($char === $this->enclosure) {
            return $content.$this->enclosure.$this->extractEnclosedFieldContent($line);
        }

        return $content.$this->extractFieldContent($line);
    }
}


1		<?php
2
3		/**
4		* League.Csv (https://csv.thephpleague.com).
5		*
6		* @author Ignace Nyamagana Butera <[email protected]>
7		* @license https://github.com/thephpleague/csv/blob/master/LICENSE (MIT License)
8		* @version 9.2.0
9		* @link https://github.com/thephpleague/csv
10		*
11		* For the full copyright and license information, please view the LICENSE
12		* file that was distributed with this source code.
13		*/
14
15		declare(strict_types=1);
16
17		namespace League\Csv;
18
19		use IteratorAggregate;
20		use SplFileObject;
21		use TypeError;
22		use function explode;
23		use function get_class;
24		use function gettype;
25		use function in_array;
26		use function is_object;
27		use function rtrim;
28		use function sprintf;
29		use function str_replace;
30		use function strlen;
31		use function substr;
32		use function trim;
33
34		/**
35		* A RFC4180 Compliant Parser in Pure PHP.
36		*
37		* @see https://php.net/manual/en/function.fgetcsv.php
38		* @see https://php.net/manual/en/function.fgets.php
39		* @see https://tools.ietf.org/html/rfc4180
40		* @see http://edoceo.com/utilitas/csv-file-format
41		*
42		* @internal used internally to produce RFC4180 compliant records
43		*/
44		final class RFC4180Parser implements IteratorAggregate
45		{
46		/**
47		* @internal
48		*/
49		const FIELD_BREAKS = [false, '', "\r\n", "\n", "\r"];
50
51		/**
52		* @var SplFileObject\|Stream
53		*/
54		private $document;
55
56		/**
57		* @var string
58		*/
59		private $delimiter;
60
61		/**
62		* @var string
63		*/
64		private $enclosure;
65
66		/**
67		* @var string
68		*/
69		private $trim_mask;
70
71		/**
72		* New instance.
73		*
74		* @param SplFileObject\|Stream $document
75		*/
76	12	public function __construct($document, string $delimiter = ',', string $enclosure = '"')
77		{
78	12	if (!$document instanceof Stream && !$document instanceof SplFileObject) {
79	3	throw new TypeError(sprintf(
80	3	'Expected a %s or an SplFileObject object, % given',
81	3	Stream::class,
82	3	is_object($document) ? get_class($document) : gettype($document)
83		));
84		}
85
86	9	if (1 !== strlen($delimiter)) {
87	3	throw new Exception(sprintf('%s() expects delimiter to be a single character %s given', __METHOD__, $delimiter));
88		}
89
90	6	if (1 !== strlen($enclosure)) {
91	3	throw new Exception(sprintf('%s() expects enclosure to be a single character %s given', __METHOD__, $enclosure));
92		}
93
94	3	$this->document = $document;
95	3	$this->delimiter = $delimiter;
96	3	$this->enclosure = $enclosure;
97	3	$this->trim_mask = str_replace([$this->delimiter, $this->enclosure], '', " \t\0\x0B");
98	3	}
99
100		/**
101		* @inheritdoc
102		*
103		* Converts the stream into a CSV record iterator by extracting records one by one
104		*
105		* The returned record array is similar to the returned value of fgetcsv
106		*
107		* - If the line is empty the record will be an array with a single value equals to null
108		* - Otherwise the array contains strings.
109		*/
110	30	public function getIterator()
111		{
112	30	$this->document->setFlags(0);
113	30	$this->document->rewind();
114		do {
115	30	$record = [];
116	30	$line = $this->document->fgets();
117		do {
118	30	$method = 'extractFieldContent';
119	30	if (($line[0] ?? '') === $this->enclosure) {
120	24	$method = 'extractEnclosedFieldContent';
121		}
122	30	$record[] = $this->$method($line);
123	30	} while (false !== $line);
124
125	30	yield $record;
126	30	} while ($this->document->valid());
127	30	}
128
129		/**
130		* Extract field without enclosure as per RFC4180.
131		*
132		* - Leading and trailing whitespaces must be removed.
133		* - trailing line-breaks must be removed.
134		*
135		* @param bool\|string $line
136		*
137		* @return null\|string
138		*/
139	30	private function extractFieldContent(&$line)
140		{
141	30	if (in_array($line, self::FIELD_BREAKS, true)) {
142	3	$line = false;
143
144	3	return null;
145		}
146
147	27	list($content, $line) = explode($this->delimiter, $line, 2) + [1 => false];
148	27	if (false === $line) {
149	24	return trim(rtrim($content, "\r\n"), $this->trim_mask);
150		}
151
152	27	return trim($content, $this->trim_mask);
153		}
154
155		/**
156		* Extract field with enclosure as per RFC4180.
157		*
158		* - Field content can spread on multiple document lines.
159		* - Content inside enclosure must be preserved.
160		* - Double enclosure sequence must be replaced by single enclosure character.
161		* - Trailing line break must be removed if they are not part of the field content.
162		* - Invalid field do not throw as per fgetcsv behavior.
163		*
164		* @param bool\|string $line
165		*/
166	24	private function extractEnclosedFieldContent(&$line): string
167		{
168	24	if (($line[0] ?? '') === $this->enclosure) {
169	24	$line = substr($line, 1);
170		}
171
172	24	$content = '';
173	24	while (false !== $line) {
174	24	list($buffer, $line) = explode($this->enclosure, $line, 2) + [1 => false];
175	24	$content .= $buffer;
176	24	if (false !== $line) {
177	21	break;
178		}
179	9	$line = $this->document->fgets();
180		}
181
182	24	if (in_array($line, self::FIELD_BREAKS, true)) {
183	9	$line = false;
184
185	9	return rtrim($content, "\r\n");
186		}
187
188	21	$char = $line[0] ?? '';
189	21	if ($char === $this->delimiter) {
190	15	$line = substr($line, 1);
191
192	15	return $content;
193		}
194
195	15	if ($char === $this->enclosure) {
196	9	return $content.$this->enclosure.$this->extractEnclosedFieldContent($line);
197		}
198
199	6	return $content.$this->extractFieldContent($line);
200		}
201		}
202

thephpleague / csv

Push — master ( 778a24...7e8982 )

RFC4180Parser::extractEnclosedFieldContent() B

Complexity

Size

Duplication

Code Coverage

Importance

Duplication Side-by-Side

Filter issues like