includes/media/SVGMetadataExtractor.php

   1 <?php
   2 /**
   3  * Extraction of SVG image metadata.
   4  *
   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.
   9  *
  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.
  14  *
  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
  19  *
  20  * @file
  21  * @ingroup Media
  22  * @author "Derk-Jan Hartman <hartman _at_ videolan d0t org>"
  23  * @author Brion Vibber
  24  * @copyright Copyright © 2010-2010 Brion Vibber, Derk-Jan Hartman
  25  * @license http://www.gnu.org/copyleft/gpl.html GNU General Public License
  26  */
  27
  28 /**
  29  * @ingroup Media
  30  */
  31 class SVGMetadataExtractor {
  32         static function getMetadata( $filename ) {
  33                 $svg = new SVGReader( $filename );
  34                 return $svg->getMetadata();
  35         }
  36 }
  37
  38 /**
  39  * @ingroup Media
  40  */
  41 class SVGReader {
  42         const DEFAULT_WIDTH = 512;
  43         const DEFAULT_HEIGHT = 512;
  44         const NS_SVG = 'http://www.w3.org/2000/svg';
  45
  46         private $reader = null;
  47         private $mDebug = false;
  48         private $metadata = Array();
  49
  50         /**
  51          * Constructor
  52          *
  53          * Creates an SVGReader drawing from the source provided
  54          * @param $source String: URI from which to read
  55          */
  56         function __construct( $source ) {
  57                 global $wgSVGMetadataCutoff;
  58                 $this->reader = new XMLReader();
  59
  60                 // Don't use $file->getSize() since file object passed to SVGHandler::getMetadata is bogus.
  61                 $size = filesize( $source );
  62                 if ( $size === false ) {
  63                         throw new MWException( "Error getting filesize of SVG." );
  64                 }
  65
  66                 if ( $size > $wgSVGMetadataCutoff ) {
  67                         $this->debug( "SVG is $size bytes, which is bigger than $wgSVGMetadataCutoff. Truncating." );
  68                         $contents = file_get_contents( $source, false, null, -1, $wgSVGMetadataCutoff );
  69                         if ($contents === false) {
  70                                 throw new MWException( 'Error reading SVG file.' );
  71                         }
  72                         $this->reader->XML( $contents, null, LIBXML_NOERROR | LIBXML_NOWARNING );
  73                 } else {
  74                         $this->reader->open( $source, null, LIBXML_NOERROR | LIBXML_NOWARNING );
  75                 }
  76
  77                 // Expand entities, since Adobe Illustrator uses them for xmlns
  78                 // attributes (bug 31719). Note that libxml2 has some protection
  79                 // against large recursive entity expansions so this is not as
  80                 // insecure as it might appear to be.
  81                 $this->reader->setParserProperty( XMLReader::SUBST_ENTITIES, true );
  82
  83                 $this->metadata['width'] = self::DEFAULT_WIDTH;
  84                 $this->metadata['height'] = self::DEFAULT_HEIGHT;
  85
  86                 // Because we cut off the end of the svg making an invalid one. Complicated
  87                 // try catch thing to make sure warnings get restored. Seems like there should
  88                 // be a better way.
  89                 wfSuppressWarnings();
  90                 try {
  91                         $this->read();
  92                 } catch( Exception $e ) {
  93                         wfRestoreWarnings();
  94                         throw $e;
  95                 }
  96                 wfRestoreWarnings();
  97         }
  98
  99         /**
 100          * @return Array with the known metadata
 101          */
 102         public function getMetadata() {
 103                 return $this->metadata;
 104         }
 105
 106         /**
 107          * Read the SVG
 108          * @return bool
 109          */
 110         public function read() {
 111                 $keepReading = $this->reader->read();
 112
 113                 /* Skip until first element */
 114                 while( $keepReading && $this->reader->nodeType != XmlReader::ELEMENT ) {
 115                         $keepReading = $this->reader->read();
 116                 }
 117
 118                 if ( $this->reader->localName != 'svg' || $this->reader->namespaceURI != self::NS_SVG ) {
 119                         throw new MWException( "Expected <svg> tag, got ".
 120                                 $this->reader->localName . " in NS " . $this->reader->namespaceURI );
 121                 }
 122                 $this->debug( "<svg> tag is correct." );
 123                 $this->handleSVGAttribs();
 124
 125                 $exitDepth =  $this->reader->depth;
 126                 $keepReading = $this->reader->read();
 127                 while ( $keepReading ) {
 128                         $tag = $this->reader->localName;
 129                         $type = $this->reader->nodeType;
 130                         $isSVG = ($this->reader->namespaceURI == self::NS_SVG);
 131
 132                         $this->debug( "$tag" );
 133
 134                         if ( $isSVG && $tag == 'svg' && $type == XmlReader::END_ELEMENT && $this->reader->depth <= $exitDepth ) {
 135                                 break;
 136                         } elseif ( $isSVG && $tag == 'title' ) {
 137                                 $this->readField( $tag, 'title' );
 138                         } elseif ( $isSVG && $tag == 'desc' ) {
 139                                 $this->readField( $tag, 'description' );
 140                         } elseif ( $isSVG && $tag == 'metadata' && $type == XmlReader::ELEMENT ) {
 141                                 $this->readXml( $tag, 'metadata' );
 142                         } elseif ( $isSVG && $tag == 'script' ) {
 143                                 // We normally do not allow scripted svgs.
 144                                 // However its possible to configure MW to let them
 145                                 // in, and such files should be considered animated.
 146                                 $this->metadata['animated'] = true;
 147                         } elseif ( $tag !== '#text' ) {
 148                                 $this->debug( "Unhandled top-level XML tag $tag" );
 149
 150                                 if ( !isset( $this->metadata['animated'] ) ) {
 151                                         // Recurse into children of current tag, looking for animation.
 152                                         $this->animateFilter( $tag );
 153                                 }
 154                         }
 155
 156                         // Goto next element, which is sibling of current (Skip children).
 157                         $keepReading = $this->reader->next();
 158                 }
 159
 160                 $this->reader->close();
 161
 162                 return true;
 163         }
 164
 165         /**
 166          * Read a textelement from an element
 167          *
 168          * @param String $name of the element that we are reading from
 169          * @param String $metafield that we will fill with the result
 170          */
 171         private function readField( $name, $metafield=null ) {
 172                 $this->debug ( "Read field $metafield" );
 173                 if( !$metafield || $this->reader->nodeType != XmlReader::ELEMENT ) {
 174                         return;
 175                 }
 176                 $keepReading = $this->reader->read();
 177                 while( $keepReading ) {
 178                         if( $this->reader->localName == $name && $this->reader->namespaceURI == self::NS_SVG && $this->reader->nodeType == XmlReader::END_ELEMENT ) {
 179                                 break;
 180                         } elseif( $this->reader->nodeType == XmlReader::TEXT ){
 181                                 $this->metadata[$metafield] = trim( $this->reader->value );
 182                         }
 183                         $keepReading = $this->reader->read();
 184                 }
 185         }
 186
 187         /**
 188          * Read an XML snippet from an element
 189          *
 190          * @param String $metafield that we will fill with the result
 191          */
 192         private function readXml( $metafield=null ) {
 193                 $this->debug ( "Read top level metadata" );
 194                 if( !$metafield || $this->reader->nodeType != XmlReader::ELEMENT ) {
 195                         return;
 196                 }
 197                 // TODO: find and store type of xml snippet. metadata['metadataType'] = "rdf"
 198                 if( method_exists( $this->reader, 'readInnerXML' ) ) {
 199                         $this->metadata[$metafield] = trim( $this->reader->readInnerXML() );
 200                 } else {
 201                         throw new MWException( "The PHP XMLReader extension does not come with readInnerXML() method. Your libxml is probably out of date (need 2.6.20 or later)." );
 202                 }
 203                 $this->reader->next();
 204         }
 205
 206         /**
 207          * Filter all children, looking for animate elements
 208          *
 209          * @param String $name of the element that we are reading from
 210          */
 211         private function animateFilter( $name ) {
 212                 $this->debug ( "animate filter for tag $name" );
 213                 if( $this->reader->nodeType != XmlReader::ELEMENT ) {
 214                         return;
 215                 }
 216                 if ( $this->reader->isEmptyElement ) {
 217                         return;
 218                 }
 219                 $exitDepth =  $this->reader->depth;
 220                 $keepReading = $this->reader->read();
 221                 while( $keepReading ) {
 222                         if( $this->reader->localName == $name && $this->reader->depth <= $exitDepth
 223                                 && $this->reader->nodeType == XmlReader::END_ELEMENT ) {
 224                                 break;
 225                         } elseif ( $this->reader->namespaceURI == self::NS_SVG && $this->reader->nodeType == XmlReader::ELEMENT ) {
 226                                 switch( $this->reader->localName ) {
 227                                         case 'script':
 228                                                 // Normally we disallow files with
 229                                                 // <script>, but its possible
 230                                                 // to configure MW to disable
 231                                                 // such checks.
 232                                         case 'animate':
 233                                         case 'set':
 234                                         case 'animateMotion':
 235                                         case 'animateColor':
 236                                         case 'animateTransform':
 237                                                 $this->debug( "HOUSTON WE HAVE ANIMATION" );
 238                                                 $this->metadata['animated'] = true;
 239                                                 break;
 240                                 }
 241                         }
 242                         $keepReading = $this->reader->read();
 243                 }
 244         }
 245
 246         private function throwXmlError( $err ) {
 247                 $this->debug( "FAILURE: $err" );
 248                 wfDebug( "SVGReader XML error: $err\n" );
 249         }
 250
 251         private function debug( $data ) {
 252                 if( $this->mDebug ) {
 253                         wfDebug( "SVGReader: $data\n" );
 254                 }
 255         }
 256
 257         private function warn( $data ) {
 258                 wfDebug( "SVGReader: $data\n" );
 259         }
 260
 261         private function notice( $data ) {
 262                 wfDebug( "SVGReader WARN: $data\n" );
 263         }
 264
 265         /**
 266          * Parse the attributes of an SVG element
 267          *
 268          * The parser has to be in the start element of "<svg>"
 269          */
 270         private function handleSVGAttribs( ) {
 271                 $defaultWidth = self::DEFAULT_WIDTH;
 272                 $defaultHeight = self::DEFAULT_HEIGHT;
 273                 $aspect = 1.0;
 274                 $width = null;
 275                 $height = null;
 276
 277                 if( $this->reader->getAttribute('viewBox') ) {
 278                         // min-x min-y width height
 279                         $viewBox = preg_split( '/\s+/', trim( $this->reader->getAttribute('viewBox') ) );
 280                         if( count( $viewBox ) == 4 ) {
 281                                 $viewWidth = $this->scaleSVGUnit( $viewBox[2] );
 282                                 $viewHeight = $this->scaleSVGUnit( $viewBox[3] );
 283                                 if( $viewWidth > 0 && $viewHeight > 0 ) {
 284                                         $aspect = $viewWidth / $viewHeight;
 285                                         $defaultHeight = $defaultWidth / $aspect;
 286                                 }
 287                         }
 288                 }
 289                 if( $this->reader->getAttribute('width') ) {
 290                         $width = $this->scaleSVGUnit( $this->reader->getAttribute('width'), $defaultWidth );
 291                 }
 292                 if( $this->reader->getAttribute('height') ) {
 293                         $height = $this->scaleSVGUnit( $this->reader->getAttribute('height'), $defaultHeight );
 294                 }
 295
 296                 if( !isset( $width ) && !isset( $height ) ) {
 297                         $width = $defaultWidth;
 298                         $height = $width / $aspect;
 299                 } elseif( isset( $width ) && !isset( $height ) ) {
 300                         $height = $width / $aspect;
 301                 } elseif( isset( $height ) && !isset( $width ) ) {
 302                         $width = $height * $aspect;
 303                 }
 304
 305                 if( $width > 0 && $height > 0 ) {
 306                         $this->metadata['width'] = intval( round( $width ) );
 307                         $this->metadata['height'] = intval( round( $height ) );
 308                 }
 309         }
 310
 311         /**
 312          * Return a rounded pixel equivalent for a labeled CSS/SVG length.
 313          * http://www.w3.org/TR/SVG11/coords.html#UnitIdentifiers
 314          *
 315          * @param $length String: CSS/SVG length.
 316          * @param $viewportSize: Float optional scale for percentage units...
 317          * @return float: length in pixels
 318          */
 319         static function scaleSVGUnit( $length, $viewportSize=512 ) {
 320                 static $unitLength = array(
 321                         'px' => 1.0,
 322                         'pt' => 1.25,
 323                         'pc' => 15.0,
 324                         'mm' => 3.543307,
 325                         'cm' => 35.43307,
 326                         'in' => 90.0,
 327                         'em' => 16.0, // fake it?
 328                         'ex' => 12.0, // fake it?
 329                         ''   => 1.0, // "User units" pixels by default
 330                         );
 331                 $matches = array();
 332                 if( preg_match( '/^\s*(\d+(?:\.\d+)?)(em|ex|px|pt|pc|cm|mm|in|%|)\s*$/', $length, $matches ) ) {
 333                         $length = floatval( $matches[1] );
 334                         $unit = $matches[2];
 335                         if( $unit == '%' ) {
 336                                 return $length * 0.01 * $viewportSize;
 337                         } else {
 338                                 return $length * $unitLength[$unit];
 339                         }
 340                 } else {
 341                         // Assume pixels
 342                         return floatval( $length );
 343                 }
 344         }
 345 }