Use PHPUnit_Framework_TestCase in a few utils/ tests
[mediawiki.git] / includes / media / DjVu.php
blobba60af46e478495fc77c00c7866376895e8c26d5
1 <?php
2 /**
3 * Handler for DjVu images.
5 * This program is free software; you can redistribute it and/or modify
6 * it under the terms of the GNU General Public License as published by
7 * the Free Software Foundation; either version 2 of the License, or
8 * (at your option) any later version.
10 * This program is distributed in the hope that it will be useful,
11 * but WITHOUT ANY WARRANTY; without even the implied warranty of
12 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
13 * GNU General Public License for more details.
15 * You should have received a copy of the GNU General Public License along
16 * with this program; if not, write to the Free Software Foundation, Inc.,
17 * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
18 * http://www.gnu.org/copyleft/gpl.html
20 * @file
21 * @ingroup Media
24 /**
25 * Handler for DjVu images
27 * @ingroup Media
29 class DjVuHandler extends ImageHandler {
30 /**
31 * @return bool
33 function isEnabled() {
34 global $wgDjvuRenderer, $wgDjvuDump, $wgDjvuToXML;
35 if ( !$wgDjvuRenderer || ( !$wgDjvuDump && !$wgDjvuToXML ) ) {
36 wfDebug( "DjVu is disabled, please set \$wgDjvuRenderer and \$wgDjvuDump\n" );
38 return false;
39 } else {
40 return true;
44 /**
45 * @param File $file
46 * @return bool
48 function mustRender( $file ) {
49 return true;
52 /**
53 * @param File $file
54 * @return bool
56 function isMultiPage( $file ) {
57 return true;
60 /**
61 * @return array
63 function getParamMap() {
64 return array(
65 'img_width' => 'width',
66 'img_page' => 'page',
70 /**
71 * @param string $name
72 * @param mixed $value
73 * @return bool
75 function validateParam( $name, $value ) {
76 if ( $name === 'page' && trim( $value ) !== (string)intval( $value ) ) {
77 // Extra junk on the end of page, probably actually a caption
78 // e.g. [[File:Foo.djvu|thumb|Page 3 of the document shows foo]]
79 return false;
81 if ( in_array( $name, array( 'width', 'height', 'page' ) ) ) {
82 if ( $value <= 0 ) {
83 return false;
84 } else {
85 return true;
87 } else {
88 return false;
92 /**
93 * @param array $params
94 * @return bool|string
96 function makeParamString( $params ) {
97 $page = isset( $params['page'] ) ? $params['page'] : 1;
98 if ( !isset( $params['width'] ) ) {
99 return false;
102 return "page{$page}-{$params['width']}px";
106 * @param string $str
107 * @return array|bool
109 function parseParamString( $str ) {
110 $m = false;
111 if ( preg_match( '/^page(\d+)-(\d+)px$/', $str, $m ) ) {
112 return array( 'width' => $m[2], 'page' => $m[1] );
113 } else {
114 return false;
119 * @param array $params
120 * @return array
122 function getScriptParams( $params ) {
123 return array(
124 'width' => $params['width'],
125 'page' => $params['page'],
130 * @param File $image
131 * @param string $dstPath
132 * @param string $dstUrl
133 * @param array $params
134 * @param int $flags
135 * @return MediaTransformError|ThumbnailImage|TransformParameterError
137 function doTransform( $image, $dstPath, $dstUrl, $params, $flags = 0 ) {
138 global $wgDjvuRenderer, $wgDjvuPostProcessor;
140 // Fetch XML and check it, to give a more informative error message than the one which
141 // normaliseParams will inevitably give.
142 $xml = $image->getMetadata();
143 if ( !$xml ) {
144 $width = isset( $params['width'] ) ? $params['width'] : 0;
145 $height = isset( $params['height'] ) ? $params['height'] : 0;
147 return new MediaTransformError( 'thumbnail_error', $width, $height,
148 wfMessage( 'djvu_no_xml' )->text() );
151 if ( !$this->normaliseParams( $image, $params ) ) {
152 return new TransformParameterError( $params );
154 $width = $params['width'];
155 $height = $params['height'];
156 $page = $params['page'];
157 if ( $page > $this->pageCount( $image ) ) {
158 return new MediaTransformError(
159 'thumbnail_error',
160 $width,
161 $height,
162 wfMessage( 'djvu_page_error' )->text()
166 if ( $flags & self::TRANSFORM_LATER ) {
167 $params = array(
168 'width' => $width,
169 'height' => $height,
170 'page' => $page
173 return new ThumbnailImage( $image, $dstUrl, $dstPath, $params );
176 if ( !wfMkdirParents( dirname( $dstPath ), null, __METHOD__ ) ) {
177 return new MediaTransformError(
178 'thumbnail_error',
179 $width,
180 $height,
181 wfMessage( 'thumbnail_dest_directory' )->text()
185 // Get local copy source for shell scripts
186 // Thumbnail extraction is very inefficient for large files.
187 // Provide a way to pool count limit the number of downloaders.
188 if ( $image->getSize() >= 1e7 ) { // 10MB
189 $work = new PoolCounterWorkViaCallback( 'GetLocalFileCopy', sha1( $image->getName() ),
190 array(
191 'doWork' => function () use ( $image ) {
192 return $image->getLocalRefPath();
196 $srcPath = $work->execute();
197 } else {
198 $srcPath = $image->getLocalRefPath();
201 if ( $srcPath === false ) { // Failed to get local copy
202 wfDebugLog( 'thumbnail',
203 sprintf( 'Thumbnail failed on %s: could not get local copy of "%s"',
204 wfHostname(), $image->getName() ) );
206 return new MediaTransformError( 'thumbnail_error',
207 $params['width'], $params['height'],
208 wfMessage( 'filemissing' )->text()
212 # Use a subshell (brackets) to aggregate stderr from both pipeline commands
213 # before redirecting it to the overall stdout. This works in both Linux and Windows XP.
214 $cmd = '(' . wfEscapeShellArg(
215 $wgDjvuRenderer,
216 "-format=ppm",
217 "-page={$page}",
218 "-size={$params['physicalWidth']}x{$params['physicalHeight']}",
219 $srcPath );
220 if ( $wgDjvuPostProcessor ) {
221 $cmd .= " | {$wgDjvuPostProcessor}";
223 $cmd .= ' > ' . wfEscapeShellArg( $dstPath ) . ') 2>&1';
224 wfProfileIn( 'ddjvu' );
225 wfDebug( __METHOD__ . ": $cmd\n" );
226 $retval = '';
227 $err = wfShellExec( $cmd, $retval );
228 wfProfileOut( 'ddjvu' );
230 $removed = $this->removeBadFile( $dstPath, $retval );
231 if ( $retval != 0 || $removed ) {
232 $this->logErrorForExternalProcess( $retval, $err, $cmd );
233 return new MediaTransformError( 'thumbnail_error', $width, $height, $err );
234 } else {
235 $params = array(
236 'width' => $width,
237 'height' => $height,
238 'page' => $page
241 return new ThumbnailImage( $image, $dstUrl, $dstPath, $params );
246 * Cache an instance of DjVuImage in an Image object, return that instance
248 * @param File $image
249 * @param string $path
250 * @return DjVuImage
252 function getDjVuImage( $image, $path ) {
253 if ( !$image ) {
254 $deja = new DjVuImage( $path );
255 } elseif ( !isset( $image->dejaImage ) ) {
256 $deja = $image->dejaImage = new DjVuImage( $path );
257 } else {
258 $deja = $image->dejaImage;
261 return $deja;
265 * Get metadata, unserializing it if neccessary.
267 * @param File $file The DjVu file in question
268 * @return string XML metadata as a string.
269 * @throws MWException
271 private function getUnserializedMetadata( File $file ) {
272 $metadata = $file->getMetadata();
273 if ( substr( $metadata, 0, 3 ) === '<?xml' ) {
274 // Old style. Not serialized but instead just a raw string of XML.
275 return $metadata;
278 wfSuppressWarnings();
279 $unser = unserialize( $metadata );
280 wfRestoreWarnings();
281 if ( is_array( $unser ) ) {
282 if ( isset( $unser['error'] ) ) {
283 return false;
284 } elseif ( isset( $unser['xml'] ) ) {
285 return $unser['xml'];
286 } else {
287 // Should never ever reach here.
288 throw new MWException( "Error unserializing DjVu metadata." );
292 // unserialize failed. Guess it wasn't really serialized after all,
293 return $metadata;
297 * Cache a document tree for the DjVu XML metadata
298 * @param File $image
299 * @param bool $gettext DOCUMENT (Default: false)
300 * @return bool|SimpleXMLElement
302 function getMetaTree( $image, $gettext = false ) {
303 if ( $gettext && isset( $image->djvuTextTree ) ) {
304 return $image->djvuTextTree;
306 if ( !$gettext && isset( $image->dejaMetaTree ) ) {
307 return $image->dejaMetaTree;
310 $metadata = $this->getUnserializedMetadata( $image );
311 if ( !$this->isMetadataValid( $image, $metadata ) ) {
312 wfDebug( "DjVu XML metadata is invalid or missing, should have been fixed in upgradeRow\n" );
314 return false;
316 wfProfileIn( __METHOD__ );
318 wfSuppressWarnings();
319 try {
320 // Set to false rather than null to avoid further attempts
321 $image->dejaMetaTree = false;
322 $image->djvuTextTree = false;
323 $tree = new SimpleXMLElement( $metadata );
324 if ( $tree->getName() == 'mw-djvu' ) {
325 /** @var SimpleXMLElement $b */
326 foreach ( $tree->children() as $b ) {
327 if ( $b->getName() == 'DjVuTxt' ) {
328 // @todo File::djvuTextTree and File::dejaMetaTree are declared
329 // dynamically. Add a public File::$data to facilitate this?
330 $image->djvuTextTree = $b;
331 } elseif ( $b->getName() == 'DjVuXML' ) {
332 $image->dejaMetaTree = $b;
335 } else {
336 $image->dejaMetaTree = $tree;
338 } catch ( Exception $e ) {
339 wfDebug( "Bogus multipage XML metadata on '{$image->getName()}'\n" );
341 wfRestoreWarnings();
342 wfProfileOut( __METHOD__ );
343 if ( $gettext ) {
344 return $image->djvuTextTree;
345 } else {
346 return $image->dejaMetaTree;
351 * @param File $image
352 * @param string $path
353 * @return bool|array False on failure
355 function getImageSize( $image, $path ) {
356 return $this->getDjVuImage( $image, $path )->getImageSize();
359 function getThumbType( $ext, $mime, $params = null ) {
360 global $wgDjvuOutputExtension;
361 static $mime;
362 if ( !isset( $mime ) ) {
363 $magic = MimeMagic::singleton();
364 $mime = $magic->guessTypesForExtension( $wgDjvuOutputExtension );
367 return array( $wgDjvuOutputExtension, $mime );
370 function getMetadata( $image, $path ) {
371 wfDebug( "Getting DjVu metadata for $path\n" );
373 $xml = $this->getDjVuImage( $image, $path )->retrieveMetaData();
374 if ( $xml === false ) {
375 // Special value so that we don't repetitively try and decode a broken file.
376 return serialize( array( 'error' => 'Error extracting metadata' ) );
377 } else {
378 return serialize( array( 'xml' => $xml ) );
382 function getMetadataType( $image ) {
383 return 'djvuxml';
386 function isMetadataValid( $image, $metadata ) {
387 return !empty( $metadata ) && $metadata != serialize( array() );
390 function pageCount( $image ) {
391 $tree = $this->getMetaTree( $image );
392 if ( !$tree ) {
393 return false;
396 return count( $tree->xpath( '//OBJECT' ) );
399 function getPageDimensions( $image, $page ) {
400 $tree = $this->getMetaTree( $image );
401 if ( !$tree ) {
402 return false;
405 $o = $tree->BODY[0]->OBJECT[$page - 1];
406 if ( $o ) {
407 return array(
408 'width' => intval( $o['width'] ),
409 'height' => intval( $o['height'] )
411 } else {
412 return false;
417 * @param File $image
418 * @param int $page Page number to get information for
419 * @return bool|string Page text or false when no text found.
421 function getPageText( $image, $page ) {
422 $tree = $this->getMetaTree( $image, true );
423 if ( !$tree ) {
424 return false;
427 $o = $tree->BODY[0]->PAGE[$page - 1];
428 if ( $o ) {
429 $txt = $o['value'];
431 return $txt;
432 } else {
433 return false;