Strings::normalize() - Code Metrics - Inspection of "Merge pull request #194 from robmachado/master" - nfephp-org/sped-common - Measure and Improve Code Quality continuously with Scrutinizer

Passed

Push — master ( 051020...3a085f )

by Roberto

created 2018-09-15 13:47 UTC

Strings::normalize() A

↳ Parent: Strings

Complexity

Conditions	1
Paths	1

Size

Total Lines

Duplication

Lines	0
Ratio	0 %

Code Coverage

Tests	10
CRAP Score	1

Importance

Changes

Metric	Value
dl	0
loc	22
ccs	10
cts	10
cp	1
rs	9.568
c	0
b	0
f	0
cc	1
nc	1
nop	1
crap	1

<?php

namespace NFePHP\Common;

/**
 * Classe auxiliar para o tratamento de strings
 * @category   NFePHP
 * @package    NFePHP\Common\Strings
 * @copyright  Copyright (c) 2008-2017
 * @license    http://www.gnu.org/licenses/lesser.html LGPL v3
 * @author     Roberto L. Machado <linux dot rlm at gmail dot com>
 * @link       http://github.com/nfephp-org/nfephp for the canonical source repository
 */

use ForceUTF8\Encoding;

class Strings
{
    /**
     * Replace all specials characters from string and retuns only 128 basics
     * NOTE: only for UTF-8
     * @param string $string
     * @return  string
     */
    public static function replaceSpecialsChars($string)
    {
        $string = self::squashCharacters($string);
        $string = str_replace('&', 'e', $string);
        $string = preg_replace("/[^a-zA-Z0-9 @#,-_.;:$%\/]/", "", $string);
        return preg_replace("/[<>]/", "", $string);
    }
    
    /**
     * Clear inputs for build in XML
     * Only UTF-8 characters is acceptable
     * & isolated, less than, greater than, quotation marks and apostrophes
     * should be replaced by their html equivalent
     * Carriage Return, Tab and Line Feed is not acceptable in strings
     * Multiple spaces is not acceptable in strings
     * And no other control character is acceptable either
     * @param string|null $input
     * @return string|null
     */
    public static function replaceUnacceptableCharacters($input)
    {
        if (empty($input)) {
            return $input;
        }
        //& isolated, less than, greater than, quotation marks and apostrophes
        //should be replaced by their html equivalent
        $input = str_replace(
            ['& ','<','>','"',"'"],
            ['&amp; ','&lt;','&gt;','&quot;','&#39;'],
            $input
        );
        $input = self::normalize($input);
        return trim($input);
    }
    
    /**
     * Converts all UTF-8 remains in ASCII
     * @param string $input
     * @return string
     */
    public static function toASCII($input)
    {
        $input = self::normalize($input);
        $input = self::squashCharacters($input);
        return mb_convert_encoding($input, 'ascii');
    }
    
    /**
     * Replaces all accented characters of their ASCII equivalents
     * @param string $input
     * @return string
     */
    public static function squashCharacters($input)
    {
        $input = trim($input);
        $aFind = ['á','à','ã','â','é','ê','í','ó','ô','õ','ú','ü',
            'ç','Á','À','Ã','Â','É','Ê','Í','Ó','Ô','Õ','Ú','Ü','Ç'];
        $aSubs = ['a','a','a','a','e','e','i','o','o','o','u','u',
            'c','A','A','A','A','E','E','I','O','O','O','U','U','C'];
        return str_replace($aFind, $aSubs, $input);
    }
    
    /**
     * Replace all non-UTF-8 chars to UTF-8
     * Remove all control chars
     * Remove all multiple spaces
     * @param string $input
     * @return string
     */
    public static function normalize($input)
    {
        //Carriage Return, Tab and Line Feed is not acceptable in strings
        $input = str_replace(["\r","\t","\n"], "", $input);
        //Multiple spaces is not acceptable in strings
        $input = preg_replace('/(?:\s\s+)/', ' ', $input);
        //Only UTF-8 characters is acceptable
        $input = Encoding::fixUTF8($input);
        $input = preg_replace(
            '/[\x00-\x08\x10\x0B\x0C\x0E-\x19\x7F]'.
            '|[\x00-\x7F][\x80-\xBF]+'.
            '|([\xC0\xC1]|[\xF0-\xFF])[\x80-\xBF]*'.
            '|[\xC2-\xDF]((?![\x80-\xBF])|[\x80-\xBF]{2,})'.
            '|[\xE0-\xEF](([\x80-\xBF](?![\x80-\xBF]))|(?![\x80-\xBF]{2})|[\x80-\xBF]{3,})/S',
            '',
            $input
        );
        $input = preg_replace('/\xE0[\x80-\x9F][\x80-\xBF]'.
            '|\xED[\xA0-\xBF][\x80-\xBF]/S', '', $input);
        //And no other control character is acceptable either
        return preg_replace('/[[:cntrl:]]/', '', $input);
    }

    /**
     * Remove all non numeric characters from string
     * @param string $string
     * @return string
     */
    public static function onlyNumbers($string)
    {
        return preg_replace("/[^0-9]/", "", $string);
    }
    
    /**
     * Remove unwanted attributes, prefixes, sulfixes and other control
     * characters like \r \n \s \t
     * @param string $string
     * @param boolean $removeEncodingTag remove encoding tag from a xml
     * @return string
     */
    public static function clearXmlString($string, $removeEncodingTag = false)
    {
        $aFind = array(
            'xmlns:default="http://www.w3.org/2000/09/xmldsig#"',
            ' standalone="no"',
            'default:',
            ':default',
            "\n",
            "\r",
            "\t"
        );
        $retXml = str_replace($aFind, "", $string);
        $retXml = preg_replace('/(\>)\s*(\<)/m', '$1$2', $retXml);
        if ($removeEncodingTag) {
            $retXml = self::deleteAllBetween($retXml, '<?xml', '?>');
        }
        return $retXml;
    }
    
    /**
     * Remove all characters between markers
     * @param string $string
     * @param string $beginning
     * @param string $end
     * @return string
     */
    public static function deleteAllBetween($string, $beginning, $end)
    {
        $beginningPos = strpos($string, $beginning);
        $endPos = strpos($string, $end);
        if ($beginningPos === false || $endPos === false) {
            return $string;
        }
        $textToDelete = substr($string, $beginningPos, ($endPos + strlen($end)) - $beginningPos);
        return str_replace($textToDelete, '', $string);
    }
    
    /**
     * Clears the xml after adding the protocol, removing repeated namespaces
     * @param string $string
     * @return string
     */
    public static function clearProtocoledXML($string)
    {
        $procXML = self::clearXmlString($string);
        $aApp = array('nfe','cte','mdfe');
        foreach ($aApp as $app) {
            $procXML = str_replace(
                'xmlns="http://www.portalfiscal.inf.br/'.$app.'" xmlns="http://www.w3.org/2000/09/xmldsig#"',
                'xmlns="http://www.portalfiscal.inf.br/'.$app.'"',
                $procXML
            );
        }
        return $procXML;
    }
    
    /**
     * Remove some alien chars from txt
     * @param string $txt
     * @return string
     */
    public static function removeSomeAlienCharsfromTxt($txt)
    {
        //remove CRs and TABs
        $txt = str_replace(["\r","\t"], "", $txt);
        //remove multiple spaces
        $txt = preg_replace('/(?:\s\s+)/', ' ', $txt);
        //remove spaces at begin and end of fields
        $txt = str_replace(["| "," |"], "|", $txt);
        return $txt;
    }
    
    /**
     * Creates a string ramdomically with the specified length
     * @param int $length
     * @return string
     */
    public static function randomString($length)
    {
        $keyspace = '0123456789abcdefghijklmnopqr'
            . 'stuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ';
        $str = '';
        $max = mb_strlen($keyspace, '8bit') - 1;
        for ($i = 0; $i < $length; ++$i) {
            $str .= $keyspace[rand(0, $max)];
        }
        return $str;
    }
}


1		<?php
2
3		namespace NFePHP\Common;
4
5		/**
6		* Classe auxiliar para o tratamento de strings
7		* @category NFePHP
8		* @package NFePHP\Common\Strings
9		* @copyright Copyright (c) 2008-2017
10		* @license http://www.gnu.org/licenses/lesser.html LGPL v3
11		* @author Roberto L. Machado <linux dot rlm at gmail dot com>
12		* @link http://github.com/nfephp-org/nfephp for the canonical source repository
13		*/
14
15		use ForceUTF8\Encoding;
16
17		class Strings
18		{
19		/**
20		* Replace all specials characters from string and retuns only 128 basics
21		* NOTE: only for UTF-8
22		* @param string $string
23		* @return string
24		*/
25	1	public static function replaceSpecialsChars($string)
26		{
27	1	$string = self::squashCharacters($string);
28	1	$string = str_replace('&', 'e', $string);
29	1	$string = preg_replace("/[^a-zA-Z0-9 @#,-_.;:$%\/]/", "", $string);
30	1	return preg_replace("/[<>]/", "", $string);
31		}
32
33		/**
34		* Clear inputs for build in XML
35		* Only UTF-8 characters is acceptable
36		* & isolated, less than, greater than, quotation marks and apostrophes
37		* should be replaced by their html equivalent
38		* Carriage Return, Tab and Line Feed is not acceptable in strings
39		* Multiple spaces is not acceptable in strings
40		* And no other control character is acceptable either
41		* @param string\|null $input
42		* @return string\|null
43		*/
44	1	public static function replaceUnacceptableCharacters($input)
45		{
46	1	if (empty($input)) {
47		return $input;
48		}
49		//& isolated, less than, greater than, quotation marks and apostrophes
50		//should be replaced by their html equivalent
51	1	$input = str_replace(
52	1	['& ','<','>','"',"'"],
53	1	['& ','<','>','"','''],
54		$input
55		);
56	1	$input = self::normalize($input);
57	1	return trim($input);
58		}
59
60		/**
61		* Converts all UTF-8 remains in ASCII
62		* @param string $input
63		* @return string
64		*/
65		public static function toASCII($input)
66		{
67		$input = self::normalize($input);
68		$input = self::squashCharacters($input);
69		return mb_convert_encoding($input, 'ascii');
70		}
71
72		/**
73		* Replaces all accented characters of their ASCII equivalents
74		* @param string $input
75		* @return string
76		*/
77	1	public static function squashCharacters($input)
78		{
79	1	$input = trim($input);
80	1	$aFind = ['á','à','ã','â','é','ê','í','ó','ô','õ','ú','ü',
81		'ç','Á','À','Ã','Â','É','Ê','Í','Ó','Ô','Õ','Ú','Ü','Ç'];
82	1	$aSubs = ['a','a','a','a','e','e','i','o','o','o','u','u',
83		'c','A','A','A','A','E','E','I','O','O','O','U','U','C'];
84	1	return str_replace($aFind, $aSubs, $input);
85		}
86
87		/**
88		* Replace all non-UTF-8 chars to UTF-8
89		* Remove all control chars
90		* Remove all multiple spaces
91		* @param string $input
92		* @return string
93		*/
94	1	public static function normalize($input)
95		{
96		//Carriage Return, Tab and Line Feed is not acceptable in strings
97	1	$input = str_replace(["\r","\t","\n"], "", $input);
98		//Multiple spaces is not acceptable in strings
99	1	$input = preg_replace('/(?:\s\s+)/', ' ', $input);
100		//Only UTF-8 characters is acceptable
101	1	$input = Encoding::fixUTF8($input);
102	1	$input = preg_replace(
103		'/[\x00-\x08\x10\x0B\x0C\x0E-\x19\x7F]'.
104		'\|[\x00-\x7F][\x80-\xBF]+'.
105		'\|([\xC0\xC1]\|[\xF0-\xFF])[\x80-\xBF]*'.
106		'\|[\xC2-\xDF]((?![\x80-\xBF])\|[\x80-\xBF]{2,})'.
107	1	'\|[\xE0-\xEF](([\x80-\xBF](?![\x80-\xBF]))\|(?![\x80-\xBF]{2})\|[\x80-\xBF]{3,})/S',
108	1	'',
109		$input
110		);
111	1	$input = preg_replace('/\xE0[\x80-\x9F][\x80-\xBF]'.
112	1	'\|\xED[\xA0-\xBF][\x80-\xBF]/S', '', $input);
113		//And no other control character is acceptable either
114	1	return preg_replace('/[[:cntrl:]]/', '', $input);
115		}
116
117		/**
118		* Remove all non numeric characters from string
119		* @param string $string
120		* @return string
121		*/
122	1	public static function onlyNumbers($string)
123		{
124	1	return preg_replace("/[^0-9]/", "", $string);
125		}
126
127		/**
128		* Remove unwanted attributes, prefixes, sulfixes and other control
129		* characters like \r \n \s \t
130		* @param string $string
131		* @param boolean $removeEncodingTag remove encoding tag from a xml
132		* @return string
133		*/
134	2	public static function clearXmlString($string, $removeEncodingTag = false)
135		{
136		$aFind = array(
137	2	'xmlns:default="http://www.w3.org/2000/09/xmldsig#"',
138		' standalone="no"',
139		'default:',
140		':default',
141		"\n",
142		"\r",
143		"\t"
144		);
145	2	$retXml = str_replace($aFind, "", $string);
146	2	$retXml = preg_replace('/(\>)\s*(\<)/m', '$1$2', $retXml);
147	2	if ($removeEncodingTag) {
148	1	$retXml = self::deleteAllBetween($retXml, '<?xml', '?>');
149		}
150	2	return $retXml;
151		}
152
153		/**
154		* Remove all characters between markers
155		* @param string $string
156		* @param string $beginning
157		* @param string $end
158		* @return string
159		*/
160	2	public static function deleteAllBetween($string, $beginning, $end)
161		{
162	2	$beginningPos = strpos($string, $beginning);
163	2	$endPos = strpos($string, $end);
164	2	if ($beginningPos === false \|\| $endPos === false) {
165		return $string;
166		}
167	2	$textToDelete = substr($string, $beginningPos, ($endPos + strlen($end)) - $beginningPos);
168	2	return str_replace($textToDelete, '', $string);
169		}
170
171		/**
172		* Clears the xml after adding the protocol, removing repeated namespaces
173		* @param string $string
174		* @return string
175		*/
176	1	public static function clearProtocoledXML($string)
177		{
178	1	$procXML = self::clearXmlString($string);
179	1	$aApp = array('nfe','cte','mdfe');
180	1	foreach ($aApp as $app) {
181	1	$procXML = str_replace(
182	1	'xmlns="http://www.portalfiscal.inf.br/'.$app.'" xmlns="http://www.w3.org/2000/09/xmldsig#"',
183	1	'xmlns="http://www.portalfiscal.inf.br/'.$app.'"',
184		$procXML
185		);
186		}
187	1	return $procXML;
188		}
189
190		/**
191		* Remove some alien chars from txt
192		* @param string $txt
193		* @return string
194		*/
195	1	public static function removeSomeAlienCharsfromTxt($txt)
196		{
197		//remove CRs and TABs
198	1	$txt = str_replace(["\r","\t"], "", $txt);
199		//remove multiple spaces
200	1	$txt = preg_replace('/(?:\s\s+)/', ' ', $txt);
201		//remove spaces at begin and end of fields
202	1	$txt = str_replace(["\| "," \|"], "\|", $txt);
203	1	return $txt;
204		}
205
206		/**
207		* Creates a string ramdomically with the specified length
208		* @param int $length
209		* @return string
210		*/
211	1	public static function randomString($length)
212		{
213		$keyspace = '0123456789abcdefghijklmnopqr'
214	1	. 'stuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ';
215	1	$str = '';
216	1	$max = mb_strlen($keyspace, '8bit') - 1;
217	1	for ($i = 0; $i < $length; ++$i) {
218	1	$str .= $keyspace[rand(0, $max)];
219		}
220	1	return $str;
221		}
222		}
223

nfephp-org / sped-common

Push — master ( 051020...3a085f )

Strings::normalize() A

Complexity

Size

Duplication

Code Coverage

Importance

Duplication Side-by-Side

Filter issues like