|
1
|
|
|
<?php |
|
2
|
|
|
|
|
3
|
|
|
namespace Cocur\Arff; |
|
4
|
|
|
|
|
5
|
|
|
use Cocur\Arff\Column\DateColumn; |
|
6
|
|
|
use Cocur\Arff\Column\NominalColumn; |
|
7
|
|
|
use Cocur\Arff\Column\NumericColumn; |
|
8
|
|
|
use Cocur\Arff\Column\StringColumn; |
|
9
|
|
|
|
|
10
|
|
|
class Reader |
|
11
|
|
|
{ |
|
12
|
|
|
/** |
|
13
|
|
|
* @param string $filename |
|
14
|
|
|
* |
|
15
|
|
|
* @return Document |
|
16
|
|
|
*/ |
|
17
|
1 |
|
public function readFile($filename) |
|
18
|
|
|
{ |
|
19
|
1 |
|
$lines = explode("\n", file_get_contents($filename)); |
|
20
|
1 |
|
$document = new Document($this->parseName($lines[0])); |
|
21
|
|
|
|
|
22
|
1 |
|
$this->parseColumns($document, $lines); |
|
23
|
1 |
|
$this->parseData($document, $lines); |
|
24
|
|
|
|
|
25
|
1 |
|
return $document; |
|
26
|
|
|
} |
|
27
|
|
|
|
|
28
|
|
|
/** |
|
29
|
|
|
* @param string $line |
|
30
|
|
|
* |
|
31
|
|
|
* @return string|null |
|
32
|
|
|
*/ |
|
33
|
|
|
protected function parseName($line) |
|
34
|
|
|
{ |
|
35
|
|
|
if (preg_match('/^@RELATION ([a-zA-Z-_\.\/\d]+)$/i', $line, $matches)) { |
|
36
|
|
|
return $matches[1]; |
|
37
|
|
|
} |
|
38
|
|
|
|
|
39
|
|
|
return null; |
|
40
|
|
|
} |
|
41
|
|
|
|
|
42
|
|
|
/** |
|
43
|
|
|
* @param Document $document |
|
44
|
|
|
* @param string[] $lines |
|
45
|
|
|
*/ |
|
46
|
|
|
protected function parseColumns(Document $document, array $lines) |
|
47
|
|
|
{ |
|
48
|
|
|
foreach ($lines as $line) { |
|
49
|
|
|
if (preg_match('/ATTRIBUTE\s([a-zA-Z0-9_-]+)\s(.*)/i', $line, $matches)) { |
|
50
|
|
|
$type = $matches[2]; |
|
51
|
|
|
$column = null; |
|
52
|
|
|
if (strcasecmp($type, 'string') === 0) { |
|
53
|
|
|
$column = new StringColumn($matches[1]); |
|
54
|
|
|
} else if (strcasecmp($type, 'numeric') === 0) { |
|
55
|
|
|
$column = new NumericColumn($matches[1]); |
|
56
|
|
|
} else if (preg_match('/^\{(.*)\}$/', $matches[2], $classMatches)) { |
|
57
|
|
|
$column = new NominalColumn($matches[1], array_map(function ($value) { |
|
58
|
|
|
return trim($value, "'"); |
|
59
|
|
|
}, preg_split( |
|
60
|
|
|
"/,(?=(?:[^\']*\'[^\']*\')*(?![^\']*\'))/", |
|
61
|
|
|
$classMatches[1] |
|
62
|
|
|
) |
|
63
|
|
|
)); |
|
64
|
|
|
} else if (preg_match('/date\s\"/', $matches[2])) { |
|
65
|
|
|
preg_match('/date\s"([A-Za-z0-9-: ]+)"/', $line, $dateMatches); |
|
66
|
|
|
$column = new DateColumn($matches[1], $dateMatches[1]); |
|
67
|
|
|
} |
|
68
|
|
|
if ($column) { |
|
69
|
|
|
$document->addColumn($column); |
|
70
|
|
|
} |
|
71
|
|
|
} |
|
72
|
|
|
} |
|
73
|
|
|
} |
|
74
|
|
|
|
|
75
|
|
|
/** |
|
76
|
|
|
* @param Document $document |
|
77
|
|
|
* @param string[] $lines |
|
78
|
|
|
*/ |
|
79
|
|
|
protected function parseData(Document $document, array $lines) |
|
80
|
|
|
{ |
|
81
|
|
|
$index = 0; |
|
82
|
|
|
while (!preg_match('/@DATA/i', $lines[$index])) { |
|
83
|
|
|
$index++; |
|
84
|
|
|
} |
|
85
|
|
|
$columns = $document->getColumns(); |
|
86
|
|
|
$columnNames = array_keys($columns); |
|
87
|
|
|
for ($i = $index+1; $i < count($lines); $i += 1) { |
|
|
|
|
|
|
88
|
|
|
$row = []; |
|
89
|
|
|
$splits = preg_split( |
|
90
|
|
|
"/,(?=(?:[^\']*\'[^\']*\')*(?![^\']*\'))/", |
|
91
|
|
|
$lines[$i], |
|
92
|
|
|
-1, |
|
93
|
|
|
PREG_SPLIT_DELIM_CAPTURE |
|
94
|
|
|
); |
|
95
|
|
|
foreach ($splits as $j => $value) { |
|
96
|
|
|
if (isset($columnNames[$j])) { |
|
97
|
|
|
$row[$columns[$columnNames[$j]]->getName()] = trim($value, "'"); |
|
98
|
|
|
} |
|
99
|
|
|
} |
|
100
|
|
|
|
|
101
|
|
|
if (count($row) != count($columnNames)) { |
|
102
|
|
|
continue; // malformed, probably and empty line |
|
103
|
|
|
} |
|
104
|
|
|
|
|
105
|
|
|
$document->addData($row); |
|
106
|
|
|
} |
|
107
|
|
|
} |
|
108
|
|
|
} |
|
109
|
|
|
|
cocur /
arff
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: