|
1
|
|
|
<?php |
|
2
|
|
|
|
|
3
|
|
|
namespace SilverStripe\TextExtraction\Extractor; |
|
4
|
|
|
|
|
5
|
|
|
use Exception; |
|
6
|
|
|
use GuzzleHttp\Client; |
|
7
|
|
|
use GuzzleHttp\Psr7\Response; |
|
8
|
|
|
use InvalidArgumentException; |
|
9
|
|
|
use Psr\Log\LoggerInterface; |
|
10
|
|
|
use SilverStripe\Assets\File; |
|
11
|
|
|
use SilverStripe\Core\Injector\Injector; |
|
12
|
|
|
|
|
13
|
|
|
/** |
|
14
|
|
|
* Text extractor that calls an Apache Solr instance |
|
15
|
|
|
* and extracts content via the "ExtractingRequestHandler" endpoint. |
|
16
|
|
|
* Does not alter the Solr index itself, but uses it purely |
|
17
|
|
|
* for its file parsing abilities. |
|
18
|
|
|
* |
|
19
|
|
|
* @author ischommer |
|
20
|
|
|
* @see http://wiki.apache.org/solr/ExtractingRequestHandler |
|
21
|
|
|
*/ |
|
22
|
|
|
class SolrCellTextExtractor extends FileTextExtractor |
|
23
|
|
|
{ |
|
24
|
|
|
/** |
|
25
|
|
|
* Base URL to use for Solr text extraction. |
|
26
|
|
|
* E.g. http://localhost:8983/solr/update/extract |
|
27
|
|
|
* |
|
28
|
|
|
* @config |
|
29
|
|
|
* @var string |
|
30
|
|
|
*/ |
|
31
|
|
|
private static $base_url; |
|
|
|
|
|
|
32
|
|
|
|
|
33
|
|
|
/** |
|
34
|
|
|
* @var int |
|
35
|
|
|
* @config |
|
36
|
|
|
*/ |
|
37
|
|
|
private static $priority = 75; |
|
|
|
|
|
|
38
|
|
|
|
|
39
|
|
|
/** |
|
40
|
|
|
* @var Client |
|
41
|
|
|
*/ |
|
42
|
|
|
protected $httpClient; |
|
43
|
|
|
|
|
44
|
|
|
/** |
|
45
|
|
|
* @return Client |
|
46
|
|
|
*/ |
|
47
|
|
|
public function getHttpClient() |
|
48
|
|
|
{ |
|
49
|
|
|
if (!$this->httpClient) { |
|
50
|
|
|
$this->httpClient = new Client(); |
|
51
|
|
|
} |
|
52
|
|
|
|
|
53
|
|
|
return $this->httpClient; |
|
54
|
|
|
} |
|
55
|
|
|
|
|
56
|
|
|
/** |
|
57
|
|
|
* @param Client $client |
|
58
|
|
|
* @return $this |
|
59
|
|
|
*/ |
|
60
|
|
|
public function setHttpClient(Client $client) |
|
61
|
|
|
{ |
|
62
|
|
|
$this->httpClient = $client; |
|
63
|
|
|
return $this; |
|
64
|
|
|
} |
|
65
|
|
|
|
|
66
|
|
|
/** |
|
67
|
|
|
* @return string |
|
68
|
|
|
*/ |
|
69
|
|
|
public function isAvailable() |
|
70
|
|
|
{ |
|
71
|
|
|
$url = $this->config()->get('base_url'); |
|
72
|
|
|
|
|
73
|
|
|
return (bool) $url; |
|
|
|
|
|
|
74
|
|
|
} |
|
75
|
|
|
|
|
76
|
|
|
/** |
|
77
|
|
|
* @param string $extension |
|
78
|
|
|
* @return bool |
|
79
|
|
|
*/ |
|
80
|
|
|
public function supportsExtension($extension) |
|
81
|
|
|
{ |
|
82
|
|
|
return in_array( |
|
83
|
|
|
strtolower($extension), |
|
84
|
|
|
[ |
|
85
|
|
|
'pdf', 'doc', 'docx', 'xls', 'xlsx', |
|
86
|
|
|
'epub', 'rtf', 'odt', 'fodt', 'ods', 'fods', |
|
87
|
|
|
'ppt', 'pptx', 'odp', 'fodp', 'csv' |
|
88
|
|
|
] |
|
89
|
|
|
); |
|
90
|
|
|
} |
|
91
|
|
|
|
|
92
|
|
|
/** |
|
93
|
|
|
* @param string $mime |
|
94
|
|
|
* @return bool |
|
95
|
|
|
*/ |
|
96
|
|
|
public function supportsMime($mime) |
|
97
|
|
|
{ |
|
98
|
|
|
// Rely on supportsExtension |
|
99
|
|
|
return false; |
|
100
|
|
|
} |
|
101
|
|
|
|
|
102
|
|
|
/** |
|
103
|
|
|
* @param File|string $file |
|
104
|
|
|
* @return string |
|
105
|
|
|
* @throws InvalidArgumentException |
|
106
|
|
|
*/ |
|
107
|
|
|
public function getContent($file) |
|
108
|
|
|
{ |
|
109
|
|
|
if (!$file || (is_string($file) && !file_exists($file))) { |
|
110
|
|
|
// no file |
|
111
|
|
|
return ''; |
|
112
|
|
|
} |
|
113
|
|
|
|
|
114
|
|
|
$fileName = $file instanceof File ? $file->getFilename() : basename($file); |
|
115
|
|
|
$client = $this->getHttpClient(); |
|
116
|
|
|
|
|
117
|
|
|
// Get and validate base URL |
|
118
|
|
|
$baseUrl = $this->config()->get('base_url'); |
|
119
|
|
|
if (!$this->config()->get('base_url')) { |
|
120
|
|
|
throw new InvalidArgumentException('SolrCellTextExtractor.base_url not specified'); |
|
121
|
|
|
} |
|
122
|
|
|
|
|
123
|
|
|
try { |
|
124
|
|
|
$stream = $file instanceof File ? $file->getStream() : fopen($file, 'r'); |
|
125
|
|
|
/** @var Response $response */ |
|
126
|
|
|
$response = $client |
|
127
|
|
|
->post($baseUrl, [ |
|
128
|
|
|
'multipart' => [ |
|
129
|
|
|
['name' => 'extractOnly', 'contents' => 'true'], |
|
130
|
|
|
['name' => 'extractFormat', 'contents' => 'text'], |
|
131
|
|
|
['name' => 'myfile', 'contents' => $stream], |
|
132
|
|
|
] |
|
133
|
|
|
]); |
|
134
|
|
|
} catch (InvalidArgumentException $e) { |
|
135
|
|
|
$msg = sprintf( |
|
136
|
|
|
'Error extracting text from "%s" (message: %s)', |
|
137
|
|
|
$fileName, |
|
138
|
|
|
$e->getMessage() |
|
139
|
|
|
); |
|
140
|
|
|
Injector::inst()->get(LoggerInterface::class)->notice($msg); |
|
141
|
|
|
return null; |
|
142
|
|
|
} catch (Exception $e) { |
|
143
|
|
|
// Catch other errors that Tika can throw via Guzzle but are not caught and break Solr search |
|
144
|
|
|
// query in some cases. |
|
145
|
|
|
$msg = sprintf( |
|
146
|
|
|
'Tika server error attempting to extract from "%s" (message: %s)', |
|
147
|
|
|
$fileName, |
|
148
|
|
|
$e->getMessage() |
|
149
|
|
|
); |
|
150
|
|
|
Injector::inst()->get(LoggerInterface::class)->notice($msg); |
|
151
|
|
|
return null; |
|
152
|
|
|
} |
|
153
|
|
|
|
|
154
|
|
|
$matches = []; |
|
155
|
|
|
// Use preg match to avoid SimpleXML running out of memory on large text nodes |
|
156
|
|
|
preg_match( |
|
157
|
|
|
sprintf('/\<str name\="%s"\>(.*?)\<\/str\>/s', preg_quote($fileName)), |
|
158
|
|
|
(string)$response->getBody(), |
|
159
|
|
|
$matches |
|
160
|
|
|
); |
|
161
|
|
|
|
|
162
|
|
|
return $matches ? $matches[1] : null; |
|
163
|
|
|
} |
|
164
|
|
|
} |
|
165
|
|
|
|