|
1
|
|
|
<?php |
|
2
|
|
|
|
|
3
|
|
|
namespace PiedWeb\TextAnalyzer; |
|
4
|
|
|
|
|
5
|
|
|
class Analyzer |
|
6
|
|
|
{ |
|
7
|
|
|
protected $onlyInSentence; |
|
8
|
|
|
protected $expressionMaxWords; |
|
9
|
|
|
protected $keepTrail; |
|
10
|
|
|
|
|
11
|
|
|
protected $text; |
|
12
|
|
|
protected $expressions = []; |
|
13
|
|
|
protected $wordNumber = 0; |
|
14
|
|
|
protected $trail = []; |
|
15
|
|
|
|
|
16
|
|
|
/** |
|
17
|
|
|
* @param string|\simple_html_dom $text |
|
18
|
|
|
*/ |
|
19
|
3 |
|
public static function get( |
|
20
|
|
|
$text, |
|
21
|
|
|
bool $onlySentence = false, |
|
22
|
|
|
int $expressionMaxWords = 5, |
|
23
|
|
|
int $keepTrail = 3 |
|
24
|
|
|
) { |
|
25
|
3 |
|
$self = new self($text); |
|
26
|
|
|
|
|
27
|
3 |
|
$self->onlyInSentence = $onlySentence; |
|
28
|
3 |
|
$self->expressionMaxWords = $expressionMaxWords; |
|
29
|
3 |
|
$self->keepTrail = $keepTrail; |
|
30
|
|
|
|
|
31
|
3 |
|
return $self->exec(); |
|
32
|
|
|
} |
|
33
|
|
|
|
|
34
|
|
|
/** |
|
35
|
|
|
* @param string|\simple_html_dom $text |
|
36
|
|
|
*/ |
|
37
|
3 |
|
protected function __construct($text) |
|
38
|
|
|
{ |
|
39
|
3 |
|
$text = is_string($text) ? CleanText::stripHtmlTags($text) : CleanText::stipHtmlTagsFromDom($text); |
|
40
|
3 |
|
$text = CleanText::fixEncoding($text); |
|
41
|
|
|
|
|
42
|
3 |
|
$text = CleanText::removeDate($text); |
|
43
|
|
|
|
|
44
|
3 |
|
if ($this->onlyInSentence) { |
|
45
|
|
|
$text = CleanText::keepOnlySentence($text); |
|
46
|
|
|
} |
|
47
|
|
|
|
|
48
|
3 |
|
$this->text = $text; |
|
49
|
3 |
|
} |
|
50
|
|
|
|
|
51
|
3 |
|
protected function incrementWordNumber(int $value) |
|
52
|
|
|
{ |
|
53
|
3 |
|
$this->wordNumber = $this->wordNumber + $value; |
|
54
|
3 |
|
} |
|
55
|
|
|
|
|
56
|
3 |
|
protected function exec() |
|
57
|
|
|
{ |
|
58
|
3 |
|
if ($this->onlyInSentence) { |
|
59
|
|
|
$sentences = []; |
|
60
|
|
|
foreach (explode(chr(10), $this->text) as $paragraph) { |
|
61
|
|
|
$sentences = array_merge($sentences, CleanText::getSentences($paragraph)); |
|
62
|
|
|
} |
|
63
|
|
|
} else { |
|
64
|
3 |
|
$sentences = explode(chr(10), trim($this->text)); |
|
65
|
|
|
} |
|
66
|
|
|
|
|
67
|
3 |
|
foreach ($sentences as $sentence) { |
|
68
|
3 |
|
$this->extract($sentence); |
|
69
|
|
|
} |
|
70
|
|
|
|
|
71
|
3 |
|
arsort($this->expressions); |
|
72
|
|
|
|
|
73
|
3 |
|
return new Analysis($this->expressions, $this->wordNumber, $this->trail); |
|
74
|
|
|
} |
|
75
|
|
|
|
|
76
|
3 |
|
protected function extract(string $sentence) |
|
77
|
|
|
{ |
|
78
|
3 |
|
$sentence = CleanText::removePunctuation($sentence); |
|
79
|
|
|
|
|
80
|
3 |
|
$words = explode(' ', trim(strtolower($sentence))); |
|
81
|
|
|
|
|
82
|
3 |
|
$wordsKey = array_keys($words); |
|
83
|
3 |
|
foreach ($wordsKey as $key) { |
|
84
|
3 |
|
for ($wordNumber = 1; $wordNumber <= $this->expressionMaxWords; ++$wordNumber) { |
|
85
|
3 |
|
$expression = ''; |
|
86
|
3 |
|
for ($i = 0; $i < $wordNumber; ++$i) { |
|
87
|
3 |
|
if (isset($words[$key + $i])) { |
|
88
|
|
|
$expression .= ($i > 0 ? ' ' : '').$words[$key + $i]; |
|
89
|
|
|
} |
|
90
|
|
|
} |
|
91
|
3 |
|
|
|
92
|
|
|
$expression = $this->cleanExpr($expression, $wordNumber); |
|
93
|
3 |
|
|
|
94
|
3 |
|
if (empty($expression) |
|
95
|
3 |
|
|| ((substr_count($expression, ' ') + 1) != $wordNumber) // We avoid sur-pondération |
|
96
|
|
|
|| ! preg_match('/[a-z]/', $expression) // We avoid number or symbol only result |
|
97
|
3 |
|
) { |
|
98
|
3 |
|
if (1 === $wordNumber) { |
|
99
|
|
|
$this->incrementWordNumber(-1); |
|
100
|
|
|
} |
|
101
|
3 |
|
} else { |
|
102
|
3 |
|
$plus = 1 + substr_count(CleanText::removeStopWords($expression), ' '); |
|
103
|
3 |
|
$this->expressions[$expression] = (isset($this->expressions[$expression]) ?? 0) + $plus; |
|
104
|
3 |
|
if ($this->keepTrail > 0 && $this->expressions[$expression] > $this->keepTrail) { |
|
105
|
|
|
$this->trail[$expression][] = $sentence; |
|
106
|
|
|
} |
|
107
|
|
|
} |
|
108
|
3 |
|
} |
|
109
|
|
|
$this->incrementWordNumber(1); |
|
110
|
3 |
|
} |
|
111
|
|
|
} |
|
112
|
3 |
|
|
|
113
|
|
|
protected function cleanExpr($expression, $wordNumber) |
|
114
|
3 |
|
{ |
|
115
|
3 |
|
if ($wordNumber <= 2) { |
|
116
|
|
|
$expression = trim(CleanText::removeStopWords(' '.$expression.' ')); |
|
117
|
3 |
|
} else { |
|
118
|
3 |
|
$expression = CleanText::removeStopWordsAtExtremity($expression); |
|
119
|
3 |
|
$expression = CleanText::removeStopWordsAtExtremity($expression); |
|
120
|
3 |
|
if (false === strpos($expression, ' ')) { |
|
121
|
|
|
$expression = trim(CleanText::removeStopWords(' '.$expression.' ')); |
|
122
|
|
|
} |
|
123
|
|
|
} |
|
124
|
|
|
|
|
125
|
3 |
|
// Last Clean |
|
126
|
3 |
|
$expression = trim(preg_replace('/\s+/', ' ', $expression)); |
|
127
|
3 |
|
if ('' == htmlentities($expression)) { //Avoid � |
|
128
|
|
|
$expression = ''; |
|
129
|
|
|
} |
|
130
|
3 |
|
|
|
131
|
|
|
return $expression; |
|
132
|
|
|
} |
|
133
|
|
|
} |
|
134
|
|
|
|
PiedWeb /
TextAnalyzer
