|
1
|
|
|
<?php |
|
2
|
|
|
|
|
3
|
|
|
/** |
|
4
|
|
|
* @file |
|
5
|
|
|
* This file is part of the PdfParser library. |
|
6
|
|
|
* |
|
7
|
|
|
* @author Sébastien MALOT <[email protected]> |
|
8
|
|
|
* |
|
9
|
|
|
* @date 2017-01-03 |
|
10
|
|
|
* |
|
11
|
|
|
* @license LGPLv3 |
|
12
|
|
|
* |
|
13
|
|
|
* @url <https://github.com/smalot/pdfparser> |
|
14
|
|
|
* |
|
15
|
|
|
* PdfParser is a pdf library written in PHP, extraction oriented. |
|
16
|
|
|
* Copyright (C) 2017 - Sébastien MALOT <[email protected]> |
|
17
|
|
|
* |
|
18
|
|
|
* This program is free software: you can redistribute it and/or modify |
|
19
|
|
|
* it under the terms of the GNU Lesser General Public License as published by |
|
20
|
|
|
* the Free Software Foundation, either version 3 of the License, or |
|
21
|
|
|
* (at your option) any later version. |
|
22
|
|
|
* |
|
23
|
|
|
* This program is distributed in the hope that it will be useful, |
|
24
|
|
|
* but WITHOUT ANY WARRANTY; without even the implied warranty of |
|
25
|
|
|
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
|
26
|
|
|
* GNU Lesser General Public License for more details. |
|
27
|
|
|
* |
|
28
|
|
|
* You should have received a copy of the GNU Lesser General Public License |
|
29
|
|
|
* along with this program. |
|
30
|
|
|
* If not, see <http://www.pdfparser.org/sites/default/LICENSE.txt>. |
|
31
|
|
|
*/ |
|
32
|
|
|
|
|
33
|
|
|
namespace Smalot\PdfParser; |
|
34
|
|
|
|
|
35
|
|
|
/** |
|
36
|
|
|
* Technical references : |
|
37
|
|
|
* - http://www.mactech.com/articles/mactech/Vol.15/15.09/PDFIntro/index.html |
|
38
|
|
|
* - http://framework.zend.com/issues/secure/attachment/12512/Pdf.php |
|
39
|
|
|
* - http://www.php.net/manual/en/ref.pdf.php#74211 |
|
40
|
|
|
* - http://cpansearch.perl.org/src/JV/PostScript-Font-1.10.02/lib/PostScript/ISOLatin1Encoding.pm |
|
41
|
|
|
* - http://cpansearch.perl.org/src/JV/PostScript-Font-1.10.02/lib/PostScript/ISOLatin9Encoding.pm |
|
42
|
|
|
* - http://cpansearch.perl.org/src/JV/PostScript-Font-1.10.02/lib/PostScript/StandardEncoding.pm |
|
43
|
|
|
* - http://cpansearch.perl.org/src/JV/PostScript-Font-1.10.02/lib/PostScript/WinAnsiEncoding.pm |
|
44
|
|
|
* |
|
45
|
|
|
* Class Document |
|
46
|
|
|
*/ |
|
47
|
|
|
class Document |
|
48
|
|
|
{ |
|
49
|
|
|
/** |
|
50
|
|
|
* @var PDFObject[] |
|
51
|
|
|
*/ |
|
52
|
|
|
protected $objects = []; |
|
53
|
|
|
|
|
54
|
|
|
/** |
|
55
|
|
|
* @var array |
|
56
|
|
|
*/ |
|
57
|
|
|
protected $dictionary = []; |
|
58
|
|
|
|
|
59
|
|
|
/** |
|
60
|
|
|
* @var Header |
|
61
|
|
|
*/ |
|
62
|
|
|
protected $trailer; |
|
63
|
|
|
|
|
64
|
|
|
/** |
|
65
|
|
|
* @var Metadata |
|
|
|
|
|
|
66
|
|
|
*/ |
|
67
|
|
|
protected $metadata = []; |
|
68
|
|
|
|
|
69
|
72 |
|
/** |
|
70
|
|
|
* @var array |
|
71
|
72 |
|
*/ |
|
72
|
72 |
|
protected $details; |
|
73
|
|
|
|
|
74
|
49 |
|
public function __construct() |
|
75
|
|
|
{ |
|
76
|
49 |
|
$this->trailer = new Header([], $this); |
|
77
|
|
|
} |
|
78
|
49 |
|
|
|
79
|
|
|
public function init() |
|
80
|
|
|
{ |
|
81
|
49 |
|
$this->buildDictionary(); |
|
82
|
49 |
|
|
|
83
|
49 |
|
$this->buildDetails(); |
|
84
|
|
|
|
|
85
|
49 |
|
// Propagate init to objects. |
|
86
|
|
|
foreach ($this->objects as $object) { |
|
87
|
|
|
$object->getHeader()->init(); |
|
88
|
|
|
$object->init(); |
|
89
|
|
|
} |
|
90
|
49 |
|
} |
|
91
|
|
|
|
|
92
|
|
|
/** |
|
93
|
49 |
|
* Build dictionary based on type header field. |
|
94
|
|
|
*/ |
|
95
|
49 |
|
protected function buildDictionary() |
|
96
|
|
|
{ |
|
97
|
49 |
|
// Build dictionary. |
|
98
|
|
|
$this->dictionary = []; |
|
99
|
49 |
|
|
|
100
|
49 |
|
foreach ($this->objects as $id => $object) { |
|
101
|
49 |
|
// Cache objects by type and subtype |
|
102
|
|
|
$type = $object->getHeader()->get('Type')->getContent(); |
|
103
|
|
|
|
|
104
|
|
|
if (null != $type) { |
|
105
|
|
|
if (!isset($this->dictionary[$type])) { |
|
106
|
|
|
$this->dictionary[$type] = [ |
|
107
|
49 |
|
'all' => [], |
|
108
|
|
|
'subtype' => [], |
|
109
|
49 |
|
]; |
|
110
|
49 |
|
} |
|
111
|
42 |
|
|
|
112
|
42 |
|
$this->dictionary[$type]['all'][$id] = $object; |
|
113
|
|
|
|
|
114
|
42 |
|
$subtype = $object->getHeader()->get('Subtype')->getContent(); |
|
115
|
|
|
if (null != $subtype) { |
|
116
|
|
|
if (!isset($this->dictionary[$type]['subtype'][$subtype])) { |
|
117
|
|
|
$this->dictionary[$type]['subtype'][$subtype] = []; |
|
118
|
49 |
|
} |
|
119
|
|
|
$this->dictionary[$type]['subtype'][$subtype][$id] = $object; |
|
120
|
|
|
} |
|
121
|
|
|
} |
|
122
|
|
|
} |
|
123
|
49 |
|
} |
|
124
|
|
|
|
|
125
|
|
|
/** |
|
126
|
49 |
|
* Build details array. |
|
127
|
|
|
*/ |
|
128
|
|
|
protected function buildDetails() |
|
129
|
49 |
|
{ |
|
130
|
|
|
// Build details array. |
|
131
|
40 |
|
$details = []; |
|
132
|
|
|
|
|
133
|
|
|
// Extract document info |
|
134
|
40 |
|
if ($this->trailer->has('Info')) { |
|
135
|
40 |
|
/** @var PDFObject $info */ |
|
136
|
|
|
$info = $this->trailer->get('Info'); |
|
137
|
|
|
// This could be an ElementMissing object, so we need to check for |
|
138
|
|
|
// the getHeader method first. |
|
139
|
|
|
if (null !== $info && method_exists($info, 'getHeader')) { |
|
140
|
|
|
$details = $info->getHeader()->getDetails(); |
|
141
|
49 |
|
} |
|
142
|
48 |
|
} |
|
143
|
2 |
|
|
|
144
|
2 |
|
// Retrieve the page count |
|
145
|
|
|
try { |
|
146
|
|
|
$pages = $this->getPages(); |
|
147
|
49 |
|
$details['Pages'] = \count($pages); |
|
148
|
49 |
|
} catch (\Exception $e) { |
|
149
|
|
|
$details['Pages'] = 0; |
|
150
|
1 |
|
} |
|
151
|
|
|
|
|
152
|
1 |
|
$details = array_merge($details, $this->metadata); |
|
|
|
|
|
|
153
|
|
|
|
|
154
|
|
|
$this->details = $details; |
|
155
|
|
|
} |
|
156
|
|
|
|
|
157
|
|
|
/** |
|
158
|
49 |
|
* Extract XMP Metadata |
|
159
|
|
|
*/ |
|
160
|
49 |
|
public function extractXMPMetadata(string $content): void |
|
161
|
|
|
{ |
|
162
|
49 |
|
$xml = xml_parser_create(); |
|
163
|
49 |
|
xml_parser_set_option($xml, \XML_OPTION_SKIP_WHITE, 1); |
|
164
|
|
|
|
|
165
|
|
|
if (xml_parse_into_struct($xml, $content, $values, $index)) { |
|
166
|
|
|
$detail = ''; |
|
167
|
|
|
|
|
168
|
1 |
|
foreach ($values as $val) { |
|
169
|
|
|
switch ($val['tag']) { |
|
170
|
1 |
|
case 'DC:CREATOR': |
|
171
|
|
|
$detail = ('open' == $val['type']) ? 'Author' : ''; |
|
172
|
|
|
break; |
|
173
|
|
|
|
|
174
|
|
|
case 'DC:DESCRIPTION': |
|
175
|
|
|
$detail = ('open' == $val['type']) ? 'Description' : ''; |
|
176
|
46 |
|
break; |
|
177
|
|
|
|
|
178
|
46 |
|
case 'DC:TITLE': |
|
179
|
46 |
|
$detail = ('open' == $val['type']) ? 'Title' : ''; |
|
180
|
|
|
break; |
|
181
|
|
|
|
|
182
|
3 |
|
case 'DC:SUBJECT': |
|
183
|
|
|
$detail = ('open' == $val['type']) ? 'Subject' : ''; |
|
184
|
|
|
break; |
|
185
|
50 |
|
|
|
186
|
|
|
case 'RDF:LI': |
|
187
|
50 |
|
if ($detail && 'complete' == $val['type'] && isset($val['value'])) { |
|
188
|
|
|
$this->metadata[$detail] = $val['value']; |
|
189
|
|
|
} |
|
190
|
53 |
|
break; |
|
191
|
|
|
|
|
192
|
53 |
|
case 'DC:FORMAT': |
|
193
|
12 |
|
if ('complete' == $val['type'] && isset($val['value'])) { |
|
194
|
|
|
$this->metadata['Format'] = $val['value']; |
|
195
|
|
|
} |
|
196
|
48 |
|
break; |
|
197
|
|
|
|
|
198
|
|
|
case 'PDF:KEYWORDS': |
|
199
|
|
|
if ('complete' == $val['type'] && isset($val['value'])) { |
|
200
|
|
|
$this->metadata['Keywords'] = $val['value']; |
|
201
|
|
|
} |
|
202
|
|
|
break; |
|
203
|
|
|
|
|
204
|
48 |
|
case 'PDF:PRODUCER': |
|
205
|
|
|
if ('complete' == $val['type'] && isset($val['value'])) { |
|
206
|
|
|
$this->metadata['Producer'] = $val['value']; |
|
207
|
|
|
} |
|
208
|
|
|
break; |
|
209
|
|
|
|
|
210
|
27 |
|
case 'PDFX:SOURCEMODIFIED': |
|
211
|
|
|
if ('complete' == $val['type'] && isset($val['value'])) { |
|
212
|
27 |
|
$this->metadata['SourceModified'] = $val['value']; |
|
213
|
|
|
} |
|
214
|
|
|
break; |
|
215
|
21 |
|
|
|
216
|
|
|
case 'PDFX:COMPANY': |
|
217
|
21 |
|
if ('complete' == $val['type'] && isset($val['value'])) { |
|
218
|
21 |
|
$this->metadata['Company'] = $val['value']; |
|
219
|
3 |
|
} |
|
220
|
|
|
break; |
|
221
|
|
|
|
|
222
|
18 |
|
case 'XMP:CREATEDATE': |
|
223
|
|
|
if ('complete' == $val['type'] && isset($val['value'])) { |
|
224
|
|
|
$this->metadata['CreationDate'] = $val['value']; |
|
225
|
|
|
} |
|
226
|
|
|
break; |
|
227
|
|
|
|
|
228
|
|
|
case 'XMP:CREATORTOOL': |
|
229
|
|
|
if ('complete' == $val['type'] && isset($val['value'])) { |
|
230
|
50 |
|
$this->metadata['Creator'] = $val['value']; |
|
231
|
|
|
} |
|
232
|
50 |
|
break; |
|
233
|
|
|
|
|
234
|
42 |
|
case 'XMP:MODIFYDATE': |
|
235
|
42 |
|
if ('complete' == $val['type'] && isset($val['value'])) { |
|
236
|
|
|
$this->metadata['ModDate'] = $val['value']; |
|
237
|
|
|
} |
|
238
|
42 |
|
break; |
|
239
|
42 |
|
|
|
240
|
42 |
|
case 'XMP:METADATADATE': |
|
241
|
|
|
if ('complete' == $val['type'] && isset($val['value'])) { |
|
242
|
|
|
$this->metadata['MetadataDate'] = $val['value']; |
|
243
|
|
|
} |
|
244
|
9 |
|
break; |
|
245
|
|
|
|
|
246
|
1 |
|
case 'XMPMM:DOCUMENTID': |
|
247
|
|
|
if ('complete' == $val['type'] && isset($val['value'])) { |
|
248
|
|
|
$this->metadata['DocumentUUID'] = $val['value']; |
|
249
|
1 |
|
} |
|
250
|
1 |
|
break; |
|
251
|
1 |
|
|
|
252
|
|
|
case 'XMPMM:INSTANCEID': |
|
253
|
|
|
if ('complete' == $val['type'] && isset($val['value'])) { |
|
254
|
1 |
|
$this->metadata['InstanceUUID'] = $val['value']; |
|
255
|
|
|
} |
|
256
|
|
|
break; |
|
257
|
9 |
|
} |
|
258
|
|
|
} |
|
259
|
7 |
|
} |
|
260
|
|
|
} |
|
261
|
7 |
|
|
|
262
|
|
|
public function getDictionary(): array |
|
263
|
|
|
{ |
|
264
|
3 |
|
return $this->dictionary; |
|
265
|
|
|
} |
|
266
|
|
|
|
|
267
|
12 |
|
/** |
|
268
|
|
|
* @param PDFObject[] $objects |
|
269
|
12 |
|
*/ |
|
270
|
12 |
|
public function setObjects($objects = []) |
|
271
|
|
|
{ |
|
272
|
|
|
$this->objects = (array) $objects; |
|
273
|
12 |
|
|
|
274
|
1 |
|
$this->init(); |
|
275
|
|
|
} |
|
276
|
|
|
|
|
277
|
12 |
|
/** |
|
278
|
|
|
* @return PDFObject[] |
|
279
|
|
|
*/ |
|
280
|
|
|
public function getObjects() |
|
281
|
12 |
|
{ |
|
282
|
|
|
return $this->objects; |
|
283
|
|
|
} |
|
284
|
12 |
|
|
|
285
|
12 |
|
/** |
|
286
|
|
|
* @return PDFObject|Font|Page|Element|null |
|
287
|
|
|
*/ |
|
288
|
|
|
public function getObjectById(string $id) |
|
289
|
12 |
|
{ |
|
290
|
|
|
if (isset($this->objects[$id])) { |
|
291
|
|
|
return $this->objects[$id]; |
|
292
|
|
|
} |
|
293
|
|
|
|
|
294
|
|
|
return null; |
|
295
|
|
|
} |
|
296
|
|
|
|
|
297
|
41 |
|
public function hasObjectsByType(string $type, string $subtype = null): bool |
|
298
|
|
|
{ |
|
299
|
41 |
|
return 0 < \count($this->getObjectsByType($type, $subtype)); |
|
300
|
41 |
|
} |
|
301
|
|
|
|
|
302
|
12 |
|
public function getObjectsByType(string $type, string $subtype = null): array |
|
303
|
|
|
{ |
|
304
|
12 |
|
if (!isset($this->dictionary[$type])) { |
|
305
|
|
|
return []; |
|
306
|
|
|
} |
|
307
|
|
|
|
|
308
|
|
|
if (null != $subtype) { |
|
|
|
|
|
|
309
|
|
|
if (!isset($this->dictionary[$type]['subtype'][$subtype])) { |
|
310
|
|
|
return []; |
|
311
|
|
|
} |
|
312
|
|
|
|
|
313
|
|
|
return $this->dictionary[$type]['subtype'][$subtype]; |
|
314
|
|
|
} |
|
315
|
|
|
|
|
316
|
|
|
return $this->dictionary[$type]['all']; |
|
317
|
|
|
} |
|
318
|
|
|
|
|
319
|
|
|
/** |
|
320
|
|
|
* @return Font[] |
|
321
|
|
|
*/ |
|
322
|
|
|
public function getFonts() |
|
323
|
|
|
{ |
|
324
|
|
|
return $this->getObjectsByType('Font'); |
|
325
|
|
|
} |
|
326
|
|
|
|
|
327
|
|
|
public function getFirstFont(): ?Font |
|
328
|
|
|
{ |
|
329
|
|
|
$fonts = $this->getFonts(); |
|
330
|
|
|
if ([] === $fonts) { |
|
331
|
|
|
return null; |
|
332
|
|
|
} |
|
333
|
|
|
|
|
334
|
|
|
return reset($fonts); |
|
335
|
|
|
} |
|
336
|
|
|
|
|
337
|
|
|
/** |
|
338
|
|
|
* @return Page[] |
|
339
|
|
|
* |
|
340
|
|
|
* @throws \Exception |
|
341
|
|
|
*/ |
|
342
|
|
|
public function getPages() |
|
343
|
|
|
{ |
|
344
|
|
|
if ($this->hasObjectsByType('Catalog')) { |
|
345
|
|
|
// Search for catalog to list pages. |
|
346
|
|
|
$catalogues = $this->getObjectsByType('Catalog'); |
|
347
|
|
|
$catalogue = reset($catalogues); |
|
348
|
|
|
|
|
349
|
|
|
/** @var Pages $object */ |
|
350
|
|
|
$object = $catalogue->get('Pages'); |
|
351
|
|
|
if (method_exists($object, 'getPages')) { |
|
352
|
|
|
return $object->getPages(true); |
|
353
|
|
|
} |
|
354
|
|
|
} |
|
355
|
|
|
|
|
356
|
|
|
if ($this->hasObjectsByType('Pages')) { |
|
357
|
|
|
// Search for pages to list kids. |
|
358
|
|
|
$pages = []; |
|
359
|
|
|
|
|
360
|
|
|
/** @var Pages[] $objects */ |
|
361
|
|
|
$objects = $this->getObjectsByType('Pages'); |
|
362
|
|
|
foreach ($objects as $object) { |
|
363
|
|
|
$pages = array_merge($pages, $object->getPages(true)); |
|
364
|
|
|
} |
|
365
|
|
|
|
|
366
|
|
|
return $pages; |
|
367
|
|
|
} |
|
368
|
|
|
|
|
369
|
|
|
if ($this->hasObjectsByType('Page')) { |
|
370
|
|
|
// Search for 'page' (unordered pages). |
|
371
|
|
|
$pages = $this->getObjectsByType('Page'); |
|
372
|
|
|
|
|
373
|
|
|
return array_values($pages); |
|
374
|
|
|
} |
|
375
|
|
|
|
|
376
|
|
|
throw new \Exception('Missing catalog.'); |
|
377
|
|
|
} |
|
378
|
|
|
|
|
379
|
|
|
public function getText(int $pageLimit = null): string |
|
380
|
|
|
{ |
|
381
|
|
|
$texts = []; |
|
382
|
|
|
$pages = $this->getPages(); |
|
383
|
|
|
|
|
384
|
|
|
// Only use the first X number of pages if $pageLimit is set and numeric. |
|
385
|
|
|
if (\is_int($pageLimit) && 0 < $pageLimit) { |
|
386
|
|
|
$pages = \array_slice($pages, 0, $pageLimit); |
|
387
|
|
|
} |
|
388
|
|
|
|
|
389
|
|
|
foreach ($pages as $index => $page) { |
|
390
|
|
|
/** |
|
391
|
|
|
* In some cases, the $page variable may be null. |
|
392
|
|
|
*/ |
|
393
|
|
|
if (null === $page) { |
|
394
|
|
|
continue; |
|
395
|
|
|
} |
|
396
|
|
|
if ($text = trim($page->getText())) { |
|
397
|
|
|
$texts[] = $text; |
|
398
|
|
|
} |
|
399
|
|
|
} |
|
400
|
|
|
|
|
401
|
|
|
return implode("\n\n", $texts); |
|
402
|
|
|
} |
|
403
|
|
|
|
|
404
|
|
|
public function getTrailer(): Header |
|
405
|
|
|
{ |
|
406
|
|
|
return $this->trailer; |
|
407
|
|
|
} |
|
408
|
|
|
|
|
409
|
|
|
public function setTrailer(Header $trailer) |
|
410
|
|
|
{ |
|
411
|
|
|
$this->trailer = $trailer; |
|
412
|
|
|
} |
|
413
|
|
|
|
|
414
|
|
|
public function getDetails(): array |
|
415
|
|
|
{ |
|
416
|
|
|
return $this->details; |
|
417
|
|
|
} |
|
418
|
|
|
} |
|
419
|
|
|
|
smalot /
pdfparser
Loading history...
The issue could also be caused by a filter entry in the build configuration. If the path has been excluded in your configuration, e.g.
excluded_paths: ["lib/*"], you can move it to the dependency path list as follows:For further information see https://scrutinizer-ci.com/docs/tools/php/php-scrutinizer/#list-dependency-paths