EmptyEscapeParser::filterDocument() - Code Metrics - Inspection of "Feature/drop php71" - thephpleague/csv - Measure and Improve Code Quality continuously with Scrutinizer

Completed

Pull Request — master (#376)

by ignace nyamagana

created 2020-02-18 20:40 UTC

EmptyEscapeParser::filterDocument() A

↳ Parent: EmptyEscapeParser

Complexity

Conditions	3
Paths	2

Size

Total Lines

Duplication

Lines	0
Ratio	0 %

Code Coverage

Tests	3
CRAP Score	5.1971

Importance

Changes

Metric	Value
cc	3
nc	2
nop	1
dl	0
loc	13
ccs	3
cts	8
cp	0.375
crap	5.1971
rs	9.8333
c	0
b	0
f	0

<?php

/**
 * League.Csv (https://csv.thephpleague.com)
 *
 * (c) Ignace Nyamagana Butera <[email protected]>
 *
 * 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\Polyfill;

use League\Csv\Stream;
use function explode;
use function get_class;
use function in_array;
use function ltrim;
use function rtrim;
use function sprintf;
use function str_replace;
use function substr;

/**
 * A Polyfill to PHP's \SplFileObject to enable parsing the CSV document
 * without taking into account the escape character.
 *
 * @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 parse a CSV document without using the escape character
 */
final class EmptyEscapeParser
{
    /**
     * @internal
     */
    const FIELD_BREAKS = [false, '', "\r\n", "\n", "\r"];

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

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

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

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

    /**
     * @var string|false
     */
    private static $line;

    /**
     * @codeCoverageIgnore
     */
    private function __construct()
    {
    }

    /**
     * Converts the document into a CSV record iterator.
     *
     * In PHP7.4+ you'll be able to do
     *
     * <code>
     * $file = new \SplFileObject('/path/to/file.csv', 'r');
     * $file->setFlags(\SplFileObject::READ_CSV | \SplFileObject::READ_AHEAD | \SplFileObject::SKIP_EMPTY);
     * $file->setCsvControl($delimiter, $enclosure, '');
     * foreach ($file as $record) {
     *    //$record escape mechanism is blocked by the empty string
     * }
     * </code>
     *
     * In PHP7.3- you can do
     *
     * <code>
     * $file = new \SplFileObject('/path/to/file.csv', 'r');
     * $it = EmptyEscapeParser::parse($file); //parsing will be done while ignoring the escape character value.
     * foreach ($it as $record) {
     *    //fgetcsv is not directly use hence the escape char is not taken into account
     * }
     * </code>
     *
     * Each record array contains strings elements.
     *
     * @param \SplFileObject|Stream $document
     *
     * @return \Generator|array[]
     */
    public static function parse($document): \Generator
    {
        self::$document = self::filterDocument($document);
function acceptsInteger($int) { }

$x = '123'; // string "123"

// Instead of
acceptsInteger($x);

// we recommend to use
acceptsInteger((integer) $x);
        list(self::$delimiter, self::$enclosure, ) = self::$document->getCsvControl();
        self::$trim_mask = str_replace([self::$delimiter, self::$enclosure], '', " \t\0\x0B");
        self::$document->setFlags(0);
        self::$document->rewind();
        while (self::$document->valid()) {
            $record = self::extractRecord();
            if ([null] === $record || !in_array(null, $record, true)) {
                yield $record;
            }
        }
    }

    /**
     * Filters the submitted document.
     *
     * @return \SplFileObject|Stream
     */
    private static function filterDocument(object $document)
    {
        if ($document instanceof Stream || $document instanceof \SplFileObject) {
            return $document;
        }

        throw new \TypeError(sprintf(
            '%s::parse expects parameter 1 to be a %s or a \SplFileObject object, %s given',
            self::class,
            Stream::class,
            get_class($document)
        ));
    }

    /**
     * Extracts a record form the CSV document.
     */
    private static function extractRecord(): array
    {
        $record = [];
        self::$line = self::$document->fgets();
        do {
            $is_field_enclosed = false;
            $buffer = '';
            if (false !== self::$line) {
                $buffer = ltrim(self::$line, self::$trim_mask);
            }

            if (($buffer[0] ?? '') === self::$enclosure) {
                $is_field_enclosed = true;
                self::$line = $buffer;
            }

            $record[] = $is_field_enclosed ? self::extractEnclosedFieldContent() : self::extractFieldContent();
        } while (false !== self::$line);

        return $record;
    }

    /**
     * Extracts the content from a field without enclosure.
     *
     * - Field content can not spread on multiple document lines.
     * - Content must be preserved.
     * - Trailing line-breaks must be removed.
     *
     * @return string|null
     */
    private static function extractFieldContent()
    {
        if (in_array(self::$line, self::FIELD_BREAKS, true)) {
            self::$line = false;

            return null;
        }

        /** @var array<string> $result */
        $result = explode(self::$delimiter, self::$line, 2);
        /** @var string $content */
        [$content, $remainder] = $result + [1 => false];


        /* @var string|false line */
        self::$line = $remainder;
        if (false === self::$line) {
            return rtrim($content, "\r\n");
        }

        return $content;
    }

    /**
     * Extracts the content from a field with enclosure.
     *
     * - Field content can spread on multiple document lines.
     * - Content between consecutive enclosure characters 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 content is treated as per fgetcsv behavior.
     *
     * @return string|null
     */
    private static function extractEnclosedFieldContent()
    {
        if (false !== self::$line && self::$line[0] === self::$enclosure) {
            self::$line = substr(self::$line, 1);
        }

        $content = '';
        while (false !== self::$line) {
            /** @var array $result */
            $result = explode(self::$enclosure, self::$line, 2);
            [$buffer, $remainder] = $result + [1 => false];

            $content .= $buffer;
            self::$line = $remainder;
            if (false !== self::$line) {
                break;
            }

            if (self::$document->valid()) {
                self::$line = self::$document->fgets();
                continue;
            }

            if ($buffer === rtrim($content, "\r\n")) {
                return null;
            }
        }

        if (in_array(self::$line, self::FIELD_BREAKS, true)) {
            self::$line = false;
            if (!self::$document->valid()) {
                return $content;
            }

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

        $char = self::$line[0] ?? '';
        if ($char === self::$delimiter) {
            self::$line = substr(self::$line, 1);

            return $content;
        }

        if ($char === self::$enclosure) {
            return $content.self::$enclosure.self::extractEnclosedFieldContent();
        }

        return $content.self::extractFieldContent();
    }
}


1		<?php
2
3		/**
4		* League.Csv (https://csv.thephpleague.com)
5		*
6		* (c) Ignace Nyamagana Butera <[email protected]>
7		*
8		* For the full copyright and license information, please view the LICENSE
9		* file that was distributed with this source code.
10		*/
11
12		declare(strict_types=1);
13
14		namespace League\Csv\Polyfill;
15
16		use League\Csv\Stream;
17		use function explode;
18		use function get_class;
19		use function in_array;
20		use function ltrim;
21		use function rtrim;
22		use function sprintf;
23		use function str_replace;
24		use function substr;
25
26		/**
27		* A Polyfill to PHP's \SplFileObject to enable parsing the CSV document
28		* without taking into account the escape character.
29		*
30		* @see https://php.net/manual/en/function.fgetcsv.php
31		* @see https://php.net/manual/en/function.fgets.php
32		* @see https://tools.ietf.org/html/rfc4180
33		* @see http://edoceo.com/utilitas/csv-file-format
34		*
35		* @internal used internally to parse a CSV document without using the escape character
36		*/
37		final class EmptyEscapeParser
38		{
39		/**
40		* @internal
41		*/
42		const FIELD_BREAKS = [false, '', "\r\n", "\n", "\r"];
43
44		/**
45		* @var \SplFileObject\|Stream
46		*/
47		private static $document;
48
49		/**
50		* @var string
51		*/
52		private static $delimiter;
53
54		/**
55		* @var string
56		*/
57		private static $enclosure;
58
59		/**
60		* @var string
61		*/
62		private static $trim_mask;
63
64		/**
65		* @var string\|false
66		*/
67		private static $line;
68
69		/**
70		* @codeCoverageIgnore
71		*/
72		private function __construct()
73		{
74		}
75
76		/**
77		* Converts the document into a CSV record iterator.
78		*
79		* In PHP7.4+ you'll be able to do
80		*
81		* <code>
82		* $file = new \SplFileObject('/path/to/file.csv', 'r');
83		* $file->setFlags(\SplFileObject::READ_CSV \| \SplFileObject::READ_AHEAD \| \SplFileObject::SKIP_EMPTY);
84		* $file->setCsvControl($delimiter, $enclosure, '');
85		* foreach ($file as $record) {
86		* //$record escape mechanism is blocked by the empty string
87		* }
88		* </code>
89		*
90		* In PHP7.3- you can do
91		*
92		* <code>
93		* $file = new \SplFileObject('/path/to/file.csv', 'r');
94		* $it = EmptyEscapeParser::parse($file); //parsing will be done while ignoring the escape character value.
95		* foreach ($it as $record) {
96		* //fgetcsv is not directly use hence the escape char is not taken into account
97		* }
98		* </code>
99		*
100		* Each record array contains strings elements.
101		*
102		* @param \SplFileObject\|Stream $document
103		*
104		* @return \Generator\|array[]
105		*/
106	45	public static function parse($document): \Generator
107		{
108	45	self::$document = self::filterDocument($document);
		0 ignored issues – show Documentation introduced 2020-02-18 20:29 UTC by Report Bug Copy Issue Report `$document` is of type `object<SplFileObject>\|object<League\Csv\Stream>`, but the function expects a `object<League\Csv\Polyfill\object>`. It seems like the type of the argument is not accepted by the function/method which you are calling. In some cases, in particular if PHP’s automatic type-juggling kicks in this might be fine. In other cases, however this might be a bug. We suggest to add an explicit type cast like in the following example: function acceptsInteger($int) { } $x = '123'; // string "123" // Instead of acceptsInteger($x); // we recommend to use acceptsInteger((integer) $x); Loading history...
109	42	list(self::$delimiter, self::$enclosure, ) = self::$document->getCsvControl();
110	42	self::$trim_mask = str_replace([self::$delimiter, self::$enclosure], '', " \t\0\x0B");
111	42	self::$document->setFlags(0);
112	42	self::$document->rewind();
113	42	while (self::$document->valid()) {
114	42	$record = self::extractRecord();
115	42	if ([null] === $record \|\| !in_array(null, $record, true)) {
116	39	yield $record;
117		}
118		}
119	42	}
120
121		/**
122		* Filters the submitted document.
123		*
124		* @return \SplFileObject\|Stream
125		*/
126	3	private static function filterDocument(object $document)
127		{
128	3	if ($document instanceof Stream \|\| $document instanceof \SplFileObject) {
129	3	return $document;
130		}
131
132		throw new \TypeError(sprintf(
133		'%s::parse expects parameter 1 to be a %s or a \SplFileObject object, %s given',
134		self::class,
135		Stream::class,
136		get_class($document)
137		));
138		}
139
140		/**
141		* Extracts a record form the CSV document.
142		*/
143	42	private static function extractRecord(): array
144		{
145	42	$record = [];
146	42	self::$line = self::$document->fgets();
147		do {
148	42	$is_field_enclosed = false;
149	42	$buffer = '';
150	42	if (false !== self::$line) {
151	42	$buffer = ltrim(self::$line, self::$trim_mask);
152		}
153
154	42	if (($buffer[0] ?? '') === self::$enclosure) {
155	30	$is_field_enclosed = true;
156	30	self::$line = $buffer;
157		}
158
159	42	$record[] = $is_field_enclosed ? self::extractEnclosedFieldContent() : self::extractFieldContent();
160	42	} while (false !== self::$line);
161
162	42	return $record;
163		}
164
165		/**
166		* Extracts the content from a field without enclosure.
167		*
168		* - Field content can not spread on multiple document lines.
169		* - Content must be preserved.
170		* - Trailing line-breaks must be removed.
171		*
172		* @return string\|null
173		*/
174	42	private static function extractFieldContent()
175		{
176	42	if (in_array(self::$line, self::FIELD_BREAKS, true)) {
177	6	self::$line = false;
178
179	6	return null;
180		}
181
182		/** @var array<string> $result */
183	39	$result = explode(self::$delimiter, self::$line, 2);
184		/** @var string $content */
185	39	[$content, $remainder] = $result + [1 => false];
		0 ignored issues – show Bug introduced 2020-02-18 20:29 UTC by Report Bug Copy Issue Report The variable `$content` does not exist. Did you forget to declare it? This check marks access to variables or properties that have not been declared yet. While PHP has no explicit notion of declaring a variable, accessing it before a value is assigned to it is most likely a bug. Loading history... Bug introduced 2020-02-18 20:29 UTC by Report Bug Copy Issue Report The variable `$remainder` does not exist. Did you forget to declare it? This check marks access to variables or properties that have not been declared yet. While PHP has no explicit notion of declaring a variable, accessing it before a value is assigned to it is most likely a bug. Loading history...
186
187		/* @var string\|false line */
188	39	self::$line = $remainder;
189	39	if (false === self::$line) {
190	27	return rtrim($content, "\r\n");
191		}
192
193	39	return $content;
194		}
195
196		/**
197		* Extracts the content from a field with enclosure.
198		*
199		* - Field content can spread on multiple document lines.
200		* - Content between consecutive enclosure characters must be preserved.
201		* - Double enclosure sequence must be replaced by single enclosure character.
202		* - Trailing line break must be removed if they are not part of the field content.
203		* - Invalid field content is treated as per fgetcsv behavior.
204		*
205		* @return string\|null
206		*/
207	30	private static function extractEnclosedFieldContent()
208		{
209	30	if (false !== self::$line && self::$line[0] === self::$enclosure) {
210	30	self::$line = substr(self::$line, 1);
211		}
212
213	30	$content = '';
214	30	while (false !== self::$line) {
215		/** @var array $result */
216	30	$result = explode(self::$enclosure, self::$line, 2);
217	30	[$buffer, $remainder] = $result + [1 => false];
		0 ignored issues – show Bug introduced 2019-06-04 20:03 UTC by Report Bug Copy Issue Report The variable `$buffer` does not exist. Did you forget to declare it? This check marks access to variables or properties that have not been declared yet. While PHP has no explicit notion of declaring a variable, accessing it before a value is assigned to it is most likely a bug. Loading history... Bug introduced 2019-06-04 20:03 UTC by Report Bug Copy Issue Report The variable `$remainder` does not exist. Did you forget to declare it? This check marks access to variables or properties that have not been declared yet. While PHP has no explicit notion of declaring a variable, accessing it before a value is assigned to it is most likely a bug. Loading history...
218	30	$content .= $buffer;
219	30	self::$line = $remainder;
220	30	if (false !== self::$line) {
221	21	break;
222		}
223
224	15	if (self::$document->valid()) {
225	9	self::$line = self::$document->fgets();
226	9	continue;
227		}
228
229	9	if ($buffer === rtrim($content, "\r\n")) {
230	3	return null;
231		}
232		}
233
234	27	if (in_array(self::$line, self::FIELD_BREAKS, true)) {
235	15	self::$line = false;
236	15	if (!self::$document->valid()) {
237	12	return $content;
238		}
239
240	9	return rtrim($content, "\r\n");
241		}
242
243	21	$char = self::$line[0] ?? '';
244	21	if ($char === self::$delimiter) {
245	15	self::$line = substr(self::$line, 1);
246
247	15	return $content;
248		}
249
250	18	if ($char === self::$enclosure) {
251	9	return $content.self::$enclosure.self::extractEnclosedFieldContent();
252		}
253
254	9	return $content.self::extractFieldContent();
255		}
256		}
257

thephpleague / csv

Pull Request — master (#376)

EmptyEscapeParser::filterDocument() A

Complexity

Size

Duplication

Code Coverage

Importance

Duplication Side-by-Side

Filter issues like