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