|
1
|
|
|
<?php |
|
2
|
|
|
|
|
3
|
|
|
declare(strict_types=1); |
|
4
|
|
|
|
|
5
|
|
|
namespace Yiisoft\Strings; |
|
6
|
|
|
|
|
7
|
|
|
/** |
|
8
|
|
|
* `CombinedRegexp` optimizes matching of multiple regular expressions. |
|
9
|
|
|
* Read more about the concept in |
|
10
|
|
|
* {@see https://nikic.github.io/2014/02/18/Fast-request-routing-using-regular-expressions.html}. |
|
11
|
|
|
*/ |
|
12
|
|
|
final class CombinedRegexp |
|
13
|
|
|
{ |
|
14
|
|
|
private const REGEXP_DELIMITER = '/'; |
|
15
|
|
|
private const QUOTE_REPLACER = '\\/'; |
|
16
|
|
|
|
|
17
|
|
|
/** |
|
18
|
|
|
* @var string[] |
|
19
|
|
|
*/ |
|
20
|
|
|
private array $patterns; |
|
21
|
|
|
private string $compiledPattern; |
|
22
|
|
|
|
|
23
|
|
|
/** |
|
24
|
|
|
* @param string[] $patterns Regular expressions to combine. |
|
25
|
|
|
* @param string $flags Flags to apply to all regular expressions. |
|
26
|
|
|
*/ |
|
27
|
19 |
|
public function __construct( |
|
28
|
|
|
array $patterns, |
|
29
|
|
|
string $flags = '' |
|
30
|
|
|
) { |
|
31
|
19 |
|
if (count($patterns) === 0) { |
|
32
|
1 |
|
throw new \InvalidArgumentException('At least one pattern should be specified.'); |
|
33
|
|
|
} |
|
34
|
18 |
|
$this->patterns = $patterns; |
|
35
|
18 |
|
$this->compiledPattern = $this->compilePatterns($patterns) . $flags; |
|
36
|
|
|
} |
|
37
|
|
|
|
|
38
|
|
|
/** |
|
39
|
|
|
* @return string The compiled pattern. |
|
40
|
|
|
*/ |
|
41
|
18 |
|
public function getCompiledPattern(): string |
|
42
|
|
|
{ |
|
43
|
18 |
|
return $this->compiledPattern; |
|
44
|
|
|
} |
|
45
|
|
|
|
|
46
|
|
|
/** |
|
47
|
|
|
* Returns `true` whether the given string matches any of the patterns, `false` - otherwise. |
|
48
|
|
|
*/ |
|
49
|
12 |
|
public function matches(string $string): bool |
|
50
|
|
|
{ |
|
51
|
12 |
|
return preg_match($this->compiledPattern, $string) === 1; |
|
52
|
|
|
} |
|
53
|
|
|
|
|
54
|
|
|
/** |
|
55
|
|
|
* Returns pattern that matches the given string. |
|
56
|
|
|
* @throws \Exception if the string does not match any of the patterns. |
|
57
|
|
|
*/ |
|
58
|
5 |
|
public function getMatchingPattern(string $string): string |
|
59
|
|
|
{ |
|
60
|
5 |
|
return $this->patterns[$this->getMatchingPatternPosition($string)]; |
|
61
|
|
|
} |
|
62
|
|
|
|
|
63
|
|
|
/** |
|
64
|
|
|
* Returns position of the pattern that matches the given string. |
|
65
|
|
|
* @throws \Exception if the string does not match any of the patterns. |
|
66
|
|
|
*/ |
|
67
|
9 |
|
public function getMatchingPatternPosition(string $string): int |
|
68
|
|
|
{ |
|
69
|
9 |
|
$match = preg_match($this->compiledPattern, $string, $matches); |
|
70
|
9 |
|
if ($match !== 1) { |
|
71
|
1 |
|
throw new \Exception( |
|
72
|
1 |
|
sprintf( |
|
73
|
1 |
|
'Failed to match pattern "%s" with string "%s".', |
|
74
|
1 |
|
$this->getCompiledPattern(), |
|
75
|
1 |
|
$string, |
|
76
|
1 |
|
) |
|
77
|
1 |
|
); |
|
78
|
|
|
} |
|
79
|
|
|
|
|
80
|
8 |
|
return count($matches) - 1; |
|
81
|
|
|
} |
|
82
|
|
|
|
|
83
|
|
|
/** |
|
84
|
|
|
* @param string[] $patterns |
|
85
|
|
|
*/ |
|
86
|
18 |
|
private function compilePatterns(array $patterns): string |
|
87
|
|
|
{ |
|
88
|
18 |
|
$quotedPatterns = []; |
|
89
|
|
|
|
|
90
|
|
|
/** |
|
91
|
|
|
* Possible mutant escaping, but it's ok for our case. |
|
92
|
|
|
* It doesn't matter where to place `()` in the pattern: |
|
93
|
|
|
* https://regex101.com/r/lE1Q1S/1, https://regex101.com/r/rWg7Fj/1 |
|
94
|
|
|
*/ |
|
95
|
18 |
|
for ($i = 0; $i < count($patterns); $i++) { |
|
|
|
|
|
|
96
|
18 |
|
$quotedPatterns[] = $patterns[$i] . str_repeat('()', $i); |
|
97
|
|
|
} |
|
98
|
18 |
|
$combinedRegexps = '(?|' . strtr( |
|
99
|
18 |
|
implode('|', $quotedPatterns), |
|
100
|
18 |
|
[self::REGEXP_DELIMITER => self::QUOTE_REPLACER] |
|
101
|
18 |
|
) . ')'; |
|
102
|
|
|
|
|
103
|
18 |
|
return self::REGEXP_DELIMITER . $combinedRegexps . self::REGEXP_DELIMITER; |
|
104
|
|
|
} |
|
105
|
|
|
} |
|
106
|
|
|
|
If the size of the collection does not change during the iteration, it is generally a good practice to compute it beforehand, and not on each iteration: