|
1
|
|
|
<?php |
|
2
|
|
|
|
|
3
|
|
|
/** |
|
4
|
|
|
* TechDivision\Import\Adapter\Goodby\Lexer |
|
5
|
|
|
* |
|
6
|
|
|
* NOTICE OF LICENSE |
|
7
|
|
|
* |
|
8
|
|
|
* This source file is subject to the Open Software License (OSL 3.0) |
|
9
|
|
|
* that is available through the world-wide-web at this URL: |
|
10
|
|
|
* http://opensource.org/licenses/osl-3.0.php |
|
11
|
|
|
* |
|
12
|
|
|
* PHP version 5 |
|
13
|
|
|
* |
|
14
|
|
|
* @author Tim Wagner <[email protected]> |
|
15
|
|
|
* @copyright 2016 TechDivision GmbH <[email protected]> |
|
16
|
|
|
* @license http://opensource.org/licenses/osl-3.0.php Open Software License (OSL 3.0) |
|
17
|
|
|
* @link https://github.com/techdivision/import |
|
18
|
|
|
* @link http://www.techdivision.com |
|
19
|
|
|
*/ |
|
20
|
|
|
|
|
21
|
|
|
namespace TechDivision\Import\Adapter\Goodby; |
|
22
|
|
|
|
|
23
|
|
|
use Goodby\CSV\Import\Protocol\LexerInterface; |
|
24
|
|
|
use Goodby\CSV\Import\Protocol\InterpreterInterface; |
|
25
|
|
|
use Goodby\CSV\Import\Standard\LexerConfig; |
|
26
|
|
|
use Goodby\CSV\Import\Standard\StreamFilter\ConvertMbstringEncoding; |
|
27
|
|
|
|
|
28
|
|
|
/** |
|
29
|
|
|
* Custom exporter implementation which resets row consistency on every import. |
|
30
|
|
|
* |
|
31
|
|
|
* @author Tim Wagner <[email protected]> |
|
32
|
|
|
* @copyright 2016 TechDivision GmbH <[email protected]> |
|
33
|
|
|
* @license http://opensource.org/licenses/osl-3.0.php Open Software License (OSL 3.0) |
|
34
|
|
|
* @link https://github.com/techdivision/import |
|
35
|
|
|
* @link http://www.techdivision.com |
|
36
|
|
|
*/ |
|
37
|
|
|
class Lexer implements LexerInterface |
|
38
|
|
|
{ |
|
39
|
|
|
|
|
40
|
|
|
/** |
|
41
|
|
|
* The exporter configuration. |
|
42
|
|
|
* |
|
43
|
|
|
* @var \Goodby\CSV\Import\Standard\LexerConfig |
|
44
|
|
|
*/ |
|
45
|
|
|
private $config; |
|
46
|
|
|
|
|
47
|
|
|
/** |
|
48
|
|
|
* Initialize the instance with the passed configuration. |
|
49
|
|
|
* |
|
50
|
|
|
* @param \Goodby\CSV\Import\Standard\LexerConfig $config The lexer configuration |
|
51
|
|
|
*/ |
|
52
|
|
|
public function __construct(LexerConfig $config = null) |
|
53
|
|
|
{ |
|
54
|
|
|
|
|
55
|
|
|
// query whether or not a configuration has been passed |
|
56
|
|
|
if ($config instanceof LexerConfig) { |
|
57
|
|
|
$this->config = $config; |
|
58
|
|
|
} else { |
|
59
|
|
|
$this->config = new LexerConfig(); |
|
60
|
|
|
} |
|
61
|
|
|
|
|
62
|
|
|
// register the encoding filter |
|
63
|
|
|
ConvertMbstringEncoding::register(); |
|
64
|
|
|
} |
|
65
|
|
|
|
|
66
|
|
|
/** |
|
67
|
|
|
* Parse the passed CSV file. |
|
68
|
|
|
* |
|
69
|
|
|
* @param string $filename The filename to parse |
|
70
|
|
|
* @param \Goodby\CSV\Import\Protocol\InterpreterInterface $interpreter The interpreter instance |
|
71
|
|
|
* |
|
72
|
|
|
* @return void |
|
73
|
|
|
*/ |
|
74
|
|
|
public function parse($filename, InterpreterInterface $interpreter) |
|
75
|
|
|
{ |
|
76
|
|
|
|
|
77
|
|
|
// for mac's office excel csv |
|
78
|
|
|
ini_set('auto_detect_line_endings', true); |
|
79
|
|
|
|
|
80
|
|
|
// initialize the configuration |
|
81
|
|
|
$delimiter = $this->config->getDelimiter(); |
|
82
|
|
|
$enclosure = $this->config->getEnclosure(); |
|
83
|
|
|
$escape = $this->config->getEscape(); |
|
84
|
|
|
$fromCharset = $this->config->getFromCharset(); |
|
85
|
|
|
$toCharset = $this->config->getToCharset(); |
|
86
|
|
|
$flags = $this->config->getFlags(); |
|
87
|
|
|
$ignoreHeader = $this->config->getIgnoreHeaderLine(); |
|
88
|
|
|
|
|
89
|
|
|
// query whether or not the charset has to be converted |
|
90
|
|
|
if ($fromCharset === null) { |
|
91
|
|
|
$url = $filename; |
|
92
|
|
|
} else { |
|
93
|
|
|
$url = ConvertMbstringEncoding::getFilterURL($filename, $fromCharset, $toCharset); |
|
94
|
|
|
} |
|
95
|
|
|
|
|
96
|
|
|
// initialize the CSV file object |
|
97
|
|
|
$csv = new \SplFileObject($url); |
|
98
|
|
|
$csv->setCsvControl($delimiter, $enclosure, $escape); |
|
99
|
|
|
$csv->setFlags($flags); |
|
100
|
|
|
|
|
101
|
|
|
// backup current locale |
|
102
|
|
|
$originalLocale = setlocale(LC_ALL, '0'); |
|
103
|
|
|
setlocale(LC_ALL, 'en_US.UTF-8'); |
|
104
|
|
|
|
|
105
|
|
|
// process each line of the CSV file |
|
106
|
|
|
foreach ($csv as $lineNumber => $line) { |
|
107
|
|
|
if ($ignoreHeader && $lineNumber == 0 || (count($line) === 1 && trim($line[0]) === '')) { |
|
108
|
|
|
continue; |
|
109
|
|
|
} |
|
110
|
|
|
$interpreter->interpret($line); |
|
111
|
|
|
} |
|
112
|
|
|
|
|
113
|
|
|
// reset locale |
|
114
|
|
|
parse_str(str_replace(';', '&', $originalLocale), $locale_array); |
|
115
|
|
|
setlocale(LC_ALL, $locale_array); |
|
116
|
|
|
} |
|
117
|
|
|
} |
|
118
|
|
|
|
techdivision /
import
