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

Completed

Push — master ( 40dc42...b54772 )

by ignace nyamagana

created 2018-09-25 07:15 UTC

Parser::filterDocument() A

↳ Parent: Parser

Complexity

Conditions	4
Paths	2

Size

Total Lines

Duplication

Lines	0
Ratio	0 %

Code Coverage

Tests	7
CRAP Score	4

Importance

Changes

Metric	Value
cc	4
nc	2
nop	1
dl	0
loc	12
ccs	7
cts	7
cp	1
crap	4
rs	9.8666
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 Generator;
use SplFileObject;
use TypeError;
use function explode;
use function get_class;
use function gettype;
use function in_array;
use function is_object;
use function ltrim;
use function rtrim;
use function sprintf;
use function str_replace;
use function strlen;
use function substr;

/**
 * A Polyfill to PHP's fgetcsv behavior with the empty string as the escape parameter.
 *
 * @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 document without using the escape character
 */
final class Parser
{
    /**
     * @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|bool
     */
    private static $line;

    /**
     * Converts the document into a CSV record iterator.
     *
     * 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.
     *
     * @param SplFileObject|Stream $document
     */
    public static function parse($document, string $delimiter = ',', string $enclosure = '"'): Generator
    {
        self::$document = self::filterDocument($document);
        self::$delimiter = self::filterControl($delimiter, 'delimiter');
        self::$enclosure = self::filterControl($enclosure, 'enclosure');
        self::$trim_mask = str_replace([self::$delimiter, self::$enclosure], '', " \t\0\x0B");
        self::$document->setFlags(0);
        self::$document->rewind();
        while (self::$document->valid()) {
            $record = [];
            self::$line = self::$document->fgets();
            do {
                $method = 'extractFieldContent';
                $buffer = ltrim(self::$line, self::$trim_mask);
                if (($buffer[0] ?? '') === self::$enclosure) {
                    $method = 'extractEnclosedFieldContent';
                    self::$line = $buffer;
                }

                $record[] = self::$method();
            } while (false !== self::$line);

            yield $record;
        }
    }

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

        throw new TypeError(sprintf(
            'Expected a %s or an SplFileObject object, %s given',
            Stream::class,
            is_object($document) ? get_class($document) : gettype($document)
        ));
    }

    /**
     * Filter a control character.
     *
     * @throws Exception if the string is not a single byte character
     */
    private static function filterControl(string $value, string $name): string
    {
        if (1 === strlen($value)) {
            return $value;
        }

        throw new Exception(sprintf('Expected %s to be a single character %s given', $name, $value));
    }

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

            return null;
        }

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

        return $content;
    }

    /**
     * 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 fields content are treated as per fgetcsv behavior.
     */
    private static function extractEnclosedFieldContent(): string
    {
        if ((self::$line[0] ?? '') === self::$enclosure) {
            self::$line = substr(self::$line, 1);
        }

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

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

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

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

            return $content;
        }

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


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


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 Generator;
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 ltrim;
28		use function rtrim;
29		use function sprintf;
30		use function str_replace;
31		use function strlen;
32		use function substr;
33
34		/**
35		* A Polyfill to PHP's fgetcsv behavior with the empty string as the escape parameter.
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 parse document without using the escape character
43		*/
44		final class Parser
45		{
46		/**
47		* @internal
48		*/
49		const FIELD_BREAKS = [false, '', "\r\n", "\n", "\r"];
50
51		/**
52		* @var SplFileObject\|Stream
53		*/
54		private static $document;
55
56		/**
57		* @var string
58		*/
59		private static $delimiter;
60
61		/**
62		* @var string
63		*/
64		private static $enclosure;
65
66		/**
67		* @var string
68		*/
69		private static $trim_mask;
70
71		/**
72		* @var string\|bool
73		*/
74		private static $line;
75
76		/**
77		* Converts the document into a CSV record iterator.
78		*
79		* The returned record array is similar to the returned value of fgetcsv
80		*
81		* - If the line is empty the record will be an array with a single value equals to null
82		* - Otherwise the array contains strings.
83		*
84		* @param SplFileObject\|Stream $document
85		*/
86	42	public static function parse($document, string $delimiter = ',', string $enclosure = '"'): Generator
87		{
88	42	self::$document = self::filterDocument($document);
89	39	self::$delimiter = self::filterControl($delimiter, 'delimiter');
90	36	self::$enclosure = self::filterControl($enclosure, 'enclosure');
91	33	self::$trim_mask = str_replace([self::$delimiter, self::$enclosure], '', " \t\0\x0B");
92	33	self::$document->setFlags(0);
93	33	self::$document->rewind();
94	33	while (self::$document->valid()) {
95	33	$record = [];
96	33	self::$line = self::$document->fgets();
97		do {
98	33	$method = 'extractFieldContent';
99	33	$buffer = ltrim(self::$line, self::$trim_mask);
100	33	if (($buffer[0] ?? '') === self::$enclosure) {
101	24	$method = 'extractEnclosedFieldContent';
102	24	self::$line = $buffer;
103		}
104
105	33	$record[] = self::$method();
106	33	} while (false !== self::$line);
107
108	33	yield $record;
109		}
110	33	}
111
112		/**
113		* Filter the submitted document.
114		*
115		* @param SplFileObject\|Stream $document
116		*
117		* @return SplFileObject\|Stream
118		*/
119	9	private static function filterDocument($document)
120		{
121	9	if ($document instanceof Stream \|\| $document instanceof SplFileObject) {
122	6	return $document;
123		}
124
125	3	throw new TypeError(sprintf(
126	3	'Expected a %s or an SplFileObject object, %s given',
127	3	Stream::class,
128	3	is_object($document) ? get_class($document) : gettype($document)
129		));
130		}
131
132		/**
133		* Filter a control character.
134		*
135		* @throws Exception if the string is not a single byte character
136		*/
137	9	private static function filterControl(string $value, string $name): string
138		{
139	9	if (1 === strlen($value)) {
140	6	return $value;
141		}
142
143	6	throw new Exception(sprintf('Expected %s to be a single character %s given', $name, $value));
144		}
145
146		/**
147		* Extract field without enclosure as per RFC4180.
148		*
149		* - Leading and trailing whitespaces must be removed.
150		* - trailing line-breaks must be removed.
151		*
152		* @return null\|string
153		*/
154	33	private static function extractFieldContent()
155		{
156	33	if (in_array(self::$line, self::FIELD_BREAKS, true)) {
157	3	self::$line = false;
158
159	3	return null;
160		}
161
162	30	list($content, self::$line) = explode(self::$delimiter, self::$line, 2) + [1 => false];
163	30	if (false === self::$line) {
164	24	return rtrim($content, "\r\n");
165		}
166
167	30	return $content;
168		}
169
170		/**
171		* Extract field with enclosure as per RFC4180.
172		*
173		* - Field content can spread on multiple document lines.
174		* - Content inside enclosure must be preserved.
175		* - Double enclosure sequence must be replaced by single enclosure character.
176		* - Trailing line break must be removed if they are not part of the field content.
177		* - Invalid fields content are treated as per fgetcsv behavior.
178		*/
179	24	private static function extractEnclosedFieldContent(): string
180		{
181	24	if ((self::$line[0] ?? '') === self::$enclosure) {
182	24	self::$line = substr(self::$line, 1);
183		}
184
185	24	$content = '';
186	24	while (false !== self::$line) {
187	24	list($buffer, $remainder) = explode(self::$enclosure, self::$line, 2) + [1 => false];
188	24	$content .= $buffer;
189	24	if (false !== $remainder) {
190	21	self::$line = $remainder;
191	21	break;
192		}
193	9	self::$line = self::$document->fgets();
194		}
195
196	24	if (in_array(self::$line, self::FIELD_BREAKS, true)) {
197	12	self::$line = false;
198
199	12	return rtrim($content, "\r\n");
200		}
201
202	21	$char = self::$line[0] ?? '';
203	21	if (self::$delimiter === $char) {
204	15	self::$line = substr(self::$line, 1);
205
206	15	return $content;
207		}
208
209	18	if (self::$enclosure === $char) {
210	9	return $content.self::$enclosure.self::extractEnclosedFieldContent();
211		}
212
213
214	9	return $content.self::extractFieldContent();
215		}
216		}
217

thephpleague / csv

Push — master ( 40dc42...b54772 )

Parser::filterDocument() A

Complexity

Size

Duplication

Code Coverage

Importance

Duplication Side-by-Side

Filter issues like