Merge "Inject LinkRendererFactory into Parser"
[lhc/web/wiklou.git] / includes / media / DjVuImage.php
1 <?php
2 /**
3 * DjVu image handler.
4 *
5 * Copyright © 2006 Brion Vibber <brion@pobox.com>
6 * https://www.mediawiki.org/
7 *
8 * This program is free software; you can redistribute it and/or modify
9 * it under the terms of the GNU General Public License as published by
10 * the Free Software Foundation; either version 2 of the License, or
11 * (at your option) any later version.
12 *
13 * This program is distributed in the hope that it will be useful,
14 * but WITHOUT ANY WARRANTY; without even the implied warranty of
15 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
16 * GNU General Public License for more details.
17 *
18 * You should have received a copy of the GNU General Public License along
19 * with this program; if not, write to the Free Software Foundation, Inc.,
20 * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
21 * http://www.gnu.org/copyleft/gpl.html
22 *
23 * @file
24 * @ingroup Media
25 */
26
27 use MediaWiki\Shell\Shell;
28
29 /**
30 * Support for detecting/validating DjVu image files and getting
31 * some basic file metadata (resolution etc)
32 *
33 * File format docs are available in source package for DjVuLibre:
34 * http://djvulibre.djvuzone.org/
35 *
36 * @ingroup Media
37 */
38 class DjVuImage {
39 /**
40 * @const DJVUTXT_MEMORY_LIMIT Memory limit for the DjVu description software
41 */
42 const DJVUTXT_MEMORY_LIMIT = 300000;
43
44 /**
45 * @param string $filename The DjVu file name.
46 */
47 function __construct( $filename ) {
48 $this->mFilename = $filename;
49 }
50
51 /**
52 * Check if the given file is indeed a valid DjVu image file
53 * @return bool
54 */
55 public function isValid() {
56 $info = $this->getInfo();
57
58 return $info !== false;
59 }
60
61 /**
62 * Return data in the style of getimagesize()
63 * @return array|bool Array or false on failure
64 */
65 public function getImageSize() {
66 $data = $this->getInfo();
67
68 if ( $data !== false ) {
69 $width = $data['width'];
70 $height = $data['height'];
71
72 return [ $width, $height, 'DjVu',
73 "width=\"$width\" height=\"$height\"" ];
74 }
75
76 return false;
77 }
78
79 // ---------
80
81 /**
82 * For debugging; dump the IFF chunk structure
83 */
84 function dump() {
85 $file = fopen( $this->mFilename, 'rb' );
86 $header = fread( $file, 12 );
87 $arr = unpack( 'a4magic/a4chunk/NchunkLength', $header );
88 $chunk = $arr['chunk'];
89 $chunkLength = $arr['chunkLength'];
90 echo "$chunk $chunkLength\n";
91 $this->dumpForm( $file, $chunkLength, 1 );
92 fclose( $file );
93 }
94
95 private function dumpForm( $file, $length, $indent ) {
96 $start = ftell( $file );
97 $secondary = fread( $file, 4 );
98 echo str_repeat( ' ', $indent * 4 ) . "($secondary)\n";
99 while ( ftell( $file ) - $start < $length ) {
100 $chunkHeader = fread( $file, 8 );
101 if ( $chunkHeader == '' ) {
102 break;
103 }
104 $arr = unpack( 'a4chunk/NchunkLength', $chunkHeader );
105 $chunk = $arr['chunk'];
106 $chunkLength = $arr['chunkLength'];
107 echo str_repeat( ' ', $indent * 4 ) . "$chunk $chunkLength\n";
108
109 if ( $chunk == 'FORM' ) {
110 $this->dumpForm( $file, $chunkLength, $indent + 1 );
111 } else {
112 fseek( $file, $chunkLength, SEEK_CUR );
113 if ( ( $chunkLength & 1 ) == 1 ) {
114 // Padding byte between chunks
115 fseek( $file, 1, SEEK_CUR );
116 }
117 }
118 }
119 }
120
121 function getInfo() {
122 Wikimedia\suppressWarnings();
123 $file = fopen( $this->mFilename, 'rb' );
124 Wikimedia\restoreWarnings();
125 if ( $file === false ) {
126 wfDebug( __METHOD__ . ": missing or failed file read\n" );
127
128 return false;
129 }
130
131 $header = fread( $file, 16 );
132 $info = false;
133
134 if ( strlen( $header ) < 16 ) {
135 wfDebug( __METHOD__ . ": too short file header\n" );
136 } else {
137 $arr = unpack( 'a4magic/a4form/NformLength/a4subtype', $header );
138
139 $subtype = $arr['subtype'];
140 if ( $arr['magic'] != 'AT&T' ) {
141 wfDebug( __METHOD__ . ": not a DjVu file\n" );
142 } elseif ( $subtype == 'DJVU' ) {
143 // Single-page document
144 $info = $this->getPageInfo( $file );
145 } elseif ( $subtype == 'DJVM' ) {
146 // Multi-page document
147 $info = $this->getMultiPageInfo( $file, $arr['formLength'] );
148 } else {
149 wfDebug( __METHOD__ . ": unrecognized DJVU file type '{$arr['subtype']}'\n" );
150 }
151 }
152 fclose( $file );
153
154 return $info;
155 }
156
157 private function readChunk( $file ) {
158 $header = fread( $file, 8 );
159 if ( strlen( $header ) < 8 ) {
160 return [ false, 0 ];
161 } else {
162 $arr = unpack( 'a4chunk/Nlength', $header );
163
164 return [ $arr['chunk'], $arr['length'] ];
165 }
166 }
167
168 private function skipChunk( $file, $chunkLength ) {
169 fseek( $file, $chunkLength, SEEK_CUR );
170
171 if ( ( $chunkLength & 0x01 ) == 1 && !feof( $file ) ) {
172 // padding byte
173 fseek( $file, 1, SEEK_CUR );
174 }
175 }
176
177 private function getMultiPageInfo( $file, $formLength ) {
178 // For now, we'll just look for the first page in the file
179 // and report its information, hoping others are the same size.
180 $start = ftell( $file );
181 do {
182 list( $chunk, $length ) = $this->readChunk( $file );
183 if ( !$chunk ) {
184 break;
185 }
186
187 if ( $chunk == 'FORM' ) {
188 $subtype = fread( $file, 4 );
189 if ( $subtype == 'DJVU' ) {
190 wfDebug( __METHOD__ . ": found first subpage\n" );
191
192 return $this->getPageInfo( $file );
193 }
194 $this->skipChunk( $file, $length - 4 );
195 } else {
196 wfDebug( __METHOD__ . ": skipping '$chunk' chunk\n" );
197 $this->skipChunk( $file, $length );
198 }
199 } while ( $length != 0 && !feof( $file ) && ftell( $file ) - $start < $formLength );
200
201 wfDebug( __METHOD__ . ": multi-page DJVU file contained no pages\n" );
202
203 return false;
204 }
205
206 private function getPageInfo( $file ) {
207 list( $chunk, $length ) = $this->readChunk( $file );
208 if ( $chunk != 'INFO' ) {
209 wfDebug( __METHOD__ . ": expected INFO chunk, got '$chunk'\n" );
210
211 return false;
212 }
213
214 if ( $length < 9 ) {
215 wfDebug( __METHOD__ . ": INFO should be 9 or 10 bytes, found $length\n" );
216
217 return false;
218 }
219 $data = fread( $file, $length );
220 if ( strlen( $data ) < $length ) {
221 wfDebug( __METHOD__ . ": INFO chunk cut off\n" );
222
223 return false;
224 }
225
226 $arr = unpack(
227 'nwidth/' .
228 'nheight/' .
229 'Cminor/' .
230 'Cmajor/' .
231 'vresolution/' .
232 'Cgamma', $data );
233
234 # Newer files have rotation info in byte 10, but we don't use it yet.
235
236 return [
237 'width' => $arr['width'],
238 'height' => $arr['height'],
239 'version' => "{$arr['major']}.{$arr['minor']}",
240 'resolution' => $arr['resolution'],
241 'gamma' => $arr['gamma'] / 10.0 ];
242 }
243
244 /**
245 * Return an XML string describing the DjVu image
246 * @return string|bool
247 */
248 function retrieveMetaData() {
249 global $wgDjvuToXML, $wgDjvuDump, $wgDjvuTxt;
250
251 if ( !$this->isValid() ) {
252 return false;
253 }
254
255 if ( isset( $wgDjvuDump ) ) {
256 # djvudump is faster as of version 3.5
257 # https://sourceforge.net/p/djvu/bugs/71/
258 $cmd = Shell::escape( $wgDjvuDump ) . ' ' . Shell::escape( $this->mFilename );
259 $dump = wfShellExec( $cmd );
260 $xml = $this->convertDumpToXML( $dump );
261 } elseif ( isset( $wgDjvuToXML ) ) {
262 $cmd = Shell::escape( $wgDjvuToXML ) . ' --without-anno --without-text ' .
263 Shell::escape( $this->mFilename );
264 $xml = wfShellExec( $cmd );
265 } else {
266 $xml = null;
267 }
268 # Text layer
269 if ( isset( $wgDjvuTxt ) ) {
270 $cmd = Shell::escape( $wgDjvuTxt ) . ' --detail=page ' . Shell::escape( $this->mFilename );
271 wfDebug( __METHOD__ . ": $cmd\n" );
272 $retval = '';
273 $txt = wfShellExec( $cmd, $retval, [], [ 'memory' => self::DJVUTXT_MEMORY_LIMIT ] );
274 if ( $retval == 0 ) {
275 # Strip some control characters
276 $txt = preg_replace( "/[\013\035\037]/", "", $txt );
277 $reg = <<<EOR
278 /\(page\s[\d-]*\s[\d-]*\s[\d-]*\s[\d-]*\s*"
279 ((?> # Text to match is composed of atoms of either:
280 \\\\. # - any escaped character
281 | # - any character different from " and \
282 [^"\\\\]+
283 )*?)
284 "\s*\)
285 | # Or page can be empty ; in this case, djvutxt dumps ()
286 \(\s*()\)/sx
287 EOR;
288 $txt = preg_replace_callback( $reg, [ $this, 'pageTextCallback' ], $txt );
289 $txt = "<DjVuTxt>\n<HEAD></HEAD>\n<BODY>\n" . $txt . "</BODY>\n</DjVuTxt>\n";
290 $xml = preg_replace( "/<DjVuXML>/", "<mw-djvu><DjVuXML>", $xml, 1 ) .
291 $txt .
292 '</mw-djvu>';
293 }
294 }
295
296 return $xml;
297 }
298
299 function pageTextCallback( $matches ) {
300 # Get rid of invalid UTF-8, strip control characters
301 $val = htmlspecialchars( UtfNormal\Validator::cleanUp( stripcslashes( $matches[1] ) ) );
302 $val = str_replace( [ "\n", '�' ], [ '&#10;', '' ], $val );
303 return '<PAGE value="' . $val . '" />';
304 }
305
306 /**
307 * Hack to temporarily work around djvutoxml bug
308 * @param string $dump
309 * @return string
310 */
311 function convertDumpToXML( $dump ) {
312 if ( strval( $dump ) == '' ) {
313 return false;
314 }
315
316 $xml = <<<EOT
317 <?xml version="1.0" ?>
318 <!DOCTYPE DjVuXML PUBLIC "-//W3C//DTD DjVuXML 1.1//EN" "pubtext/DjVuXML-s.dtd">
319 <DjVuXML>
320 <HEAD></HEAD>
321 <BODY>
322 EOT;
323
324 $dump = str_replace( "\r", '', $dump );
325 $line = strtok( $dump, "\n" );
326 $m = false;
327 $good = false;
328 if ( preg_match( '/^( *)FORM:DJVU/', $line, $m ) ) {
329 # Single-page
330 if ( $this->parseFormDjvu( $line, $xml ) ) {
331 $good = true;
332 } else {
333 return false;
334 }
335 } elseif ( preg_match( '/^( *)FORM:DJVM/', $line, $m ) ) {
336 # Multi-page
337 $parentLevel = strlen( $m[1] );
338 # Find DIRM
339 $line = strtok( "\n" );
340 while ( $line !== false ) {
341 $childLevel = strspn( $line, ' ' );
342 if ( $childLevel <= $parentLevel ) {
343 # End of chunk
344 break;
345 }
346
347 if ( preg_match( '/^ *DIRM.*indirect/', $line ) ) {
348 wfDebug( "Indirect multi-page DjVu document, bad for server!\n" );
349
350 return false;
351 }
352 if ( preg_match( '/^ *FORM:DJVU/', $line ) ) {
353 # Found page
354 if ( $this->parseFormDjvu( $line, $xml ) ) {
355 $good = true;
356 } else {
357 return false;
358 }
359 }
360 $line = strtok( "\n" );
361 }
362 }
363 if ( !$good ) {
364 return false;
365 }
366
367 $xml .= "</BODY>\n</DjVuXML>\n";
368
369 return $xml;
370 }
371
372 function parseFormDjvu( $line, &$xml ) {
373 $parentLevel = strspn( $line, ' ' );
374 $line = strtok( "\n" );
375
376 # Find INFO
377 while ( $line !== false ) {
378 $childLevel = strspn( $line, ' ' );
379 if ( $childLevel <= $parentLevel ) {
380 # End of chunk
381 break;
382 }
383
384 if ( preg_match(
385 '/^ *INFO *\[\d*\] *DjVu *(\d+)x(\d+), *\w*, *(\d+) *dpi, *gamma=([0-9.-]+)/',
386 $line,
387 $m
388 ) ) {
389 $xml .= Xml::tags(
390 'OBJECT',
391 [
392 # 'data' => '',
393 # 'type' => 'image/x.djvu',
394 'height' => $m[2],
395 'width' => $m[1],
396 # 'usemap' => '',
397 ],
398 "\n" .
399 Xml::element( 'PARAM', [ 'name' => 'DPI', 'value' => $m[3] ] ) . "\n" .
400 Xml::element( 'PARAM', [ 'name' => 'GAMMA', 'value' => $m[4] ] ) . "\n"
401 ) . "\n";
402
403 return true;
404 }
405 $line = strtok( "\n" );
406 }
407
408 # Not found
409 return false;
410 }
411 }