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          * @throws MWException|Exception
  56          */
  57         function __construct( $source ) {
  58                 global $wgSVGMetadataCutoff;
  59                 $this->reader = new XMLReader();
  60
  61                 // Don't use $file->getSize() since file object passed to SVGHandler::getMetadata is bogus.
  62                 $size = filesize( $source );
  63                 if ( $size === false ) {
  64                         throw new MWException( "Error getting filesize of SVG." );
  65                 }
  66
  67                 if ( $size > $wgSVGMetadataCutoff ) {
  68                         $this->debug( "SVG is $size bytes, which is bigger than $wgSVGMetadataCutoff. Truncating." );
  69                         $contents = file_get_contents( $source, false, null, -1, $wgSVGMetadataCutoff );
  70                         if ($contents === false) {
  71                                 throw new MWException( 'Error reading SVG file.' );
  72                         }
  73                         $this->reader->XML( $contents, null, LIBXML_NOERROR | LIBXML_NOWARNING );
  74                 } else {
  75                         $this->reader->open( $source, null, LIBXML_NOERROR | LIBXML_NOWARNING );
  76                 }
  77
  78                 // Expand entities, since Adobe Illustrator uses them for xmlns
  79                 // attributes (bug 31719). Note that libxml2 has some protection
  80                 // against large recursive entity expansions so this is not as
  81                 // insecure as it might appear to be.
  82                 $this->reader->setParserProperty( XMLReader::SUBST_ENTITIES, true );
  83
  84                 $this->metadata['width'] = self::DEFAULT_WIDTH;
  85                 $this->metadata['height'] = self::DEFAULT_HEIGHT;
  86
  87                 // The size in the units specified by the SVG file
  88                 // (for the metadata box)
  89                 // Per the SVG spec, if unspecified, default to '100%'
  90                 $this->metadata['originalWidth'] = '100%';
  91                 $this->metadata['originalHeight'] = '100%';
  92
  93                 // Because we cut off the end of the svg making an invalid one. Complicated
  94                 // try catch thing to make sure warnings get restored. Seems like there should
  95                 // be a better way.
  96                 wfSuppressWarnings();
  97                 try {
  98                         $this->read();
  99                 } catch( Exception $e ) {
 100                         // Note, if this happens, the width/height will be taken to be 0x0.
 101                         // Should we consider it the default 512x512 instead?
 102                         wfRestoreWarnings();
 103                         throw $e;
 104                 }
 105                 wfRestoreWarnings();
 106         }
 107
 108         /**
 109          * @return Array with the known metadata
 110          */
 111         public function getMetadata() {
 112                 return $this->metadata;
 113         }
 114
 115         /**
 116          * Read the SVG
 117          * @throws MWException
 118          * @return bool
 119          */
 120         public function read() {
 121                 $keepReading = $this->reader->read();
 122
 123                 /* Skip until first element */
 124                 while( $keepReading && $this->reader->nodeType != XmlReader::ELEMENT ) {
 125                         $keepReading = $this->reader->read();
 126                 }
 127
 128                 if ( $this->reader->localName != 'svg' || $this->reader->namespaceURI != self::NS_SVG ) {
 129                         throw new MWException( "Expected <svg> tag, got ".
 130                                 $this->reader->localName . " in NS " . $this->reader->namespaceURI );
 131                 }
 132                 $this->debug( "<svg> tag is correct." );
 133                 $this->handleSVGAttribs();
 134
 135                 $exitDepth =  $this->reader->depth;
 136                 $keepReading = $this->reader->read();
 137                 while ( $keepReading ) {
 138                         $tag = $this->reader->localName;
 139                         $type = $this->reader->nodeType;
 140                         $isSVG = ($this->reader->namespaceURI == self::NS_SVG);
 141
 142                         $this->debug( "$tag" );
 143
 144                         if ( $isSVG && $tag == 'svg' && $type == XmlReader::END_ELEMENT && $this->reader->depth <= $exitDepth ) {
 145                                 break;
 146                         } elseif ( $isSVG && $tag == 'title' ) {
 147                                 $this->readField( $tag, 'title' );
 148                         } elseif ( $isSVG && $tag == 'desc' ) {
 149                                 $this->readField( $tag, 'description' );
 150                         } elseif ( $isSVG && $tag == 'metadata' && $type == XmlReader::ELEMENT ) {
 151                                 $this->readXml( $tag, 'metadata' );
 152                         } elseif ( $isSVG && $tag == 'script' ) {
 153                                 // We normally do not allow scripted svgs.
 154                                 // However its possible to configure MW to let them
 155                                 // in, and such files should be considered animated.
 156                                 $this->metadata['animated'] = true;
 157                         } elseif ( $tag !== '#text' ) {
 158                                 $this->debug( "Unhandled top-level XML tag $tag" );
 159
 160                                 if ( !isset( $this->metadata['animated'] ) ) {
 161                                         // Recurse into children of current tag, looking for animation.
 162                                         $this->animateFilter( $tag );
 163                                 }
 164                         }
 165
 166                         // Goto next element, which is sibling of current (Skip children).
 167                         $keepReading = $this->reader->next();
 168                 }
 169
 170                 $this->reader->close();
 171
 172                 return true;
 173         }
 174
 175         /**
 176          * Read a textelement from an element
 177          *
 178          * @param String $name of the element that we are reading from
 179          * @param String $metafield that we will fill with the result
 180          */
 181         private function readField( $name, $metafield=null ) {
 182                 $this->debug ( "Read field $metafield" );
 183                 if( !$metafield || $this->reader->nodeType != XmlReader::ELEMENT ) {
 184                         return;
 185                 }
 186                 $keepReading = $this->reader->read();
 187                 while( $keepReading ) {
 188                         if( $this->reader->localName == $name && $this->reader->namespaceURI == self::NS_SVG && $this->reader->nodeType == XmlReader::END_ELEMENT ) {
 189                                 break;
 190                         } elseif( $this->reader->nodeType == XmlReader::TEXT ){
 191                                 $this->metadata[$metafield] = trim( $this->reader->value );
 192                         }
 193                         $keepReading = $this->reader->read();
 194                 }
 195         }
 196
 197         /**
 198          * Read an XML snippet from an element
 199          *
 200          * @param String $metafield that we will fill with the result
 201          * @throws MWException
 202          */
 203         private function readXml( $metafield=null ) {
 204                 $this->debug ( "Read top level metadata" );
 205                 if( !$metafield || $this->reader->nodeType != XmlReader::ELEMENT ) {
 206                         return;
 207                 }
 208                 // TODO: find and store type of xml snippet. metadata['metadataType'] = "rdf"
 209                 if( method_exists( $this->reader, 'readInnerXML' ) ) {
 210                         $this->metadata[$metafield] = trim( $this->reader->readInnerXML() );
 211                 } else {
 212                         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)." );
 213                 }
 214                 $this->reader->next();
 215         }
 216
 217         /**
 218          * Filter all children, looking for animate elements
 219          *
 220          * @param String $name of the element that we are reading from
 221          */
 222         private function animateFilter( $name ) {
 223                 $this->debug ( "animate filter for tag $name" );
 224                 if( $this->reader->nodeType != XmlReader::ELEMENT ) {
 225                         return;
 226                 }
 227                 if ( $this->reader->isEmptyElement ) {
 228                         return;
 229                 }
 230                 $exitDepth =  $this->reader->depth;
 231                 $keepReading = $this->reader->read();
 232                 while( $keepReading ) {
 233                         if( $this->reader->localName == $name && $this->reader->depth <= $exitDepth
 234                                 && $this->reader->nodeType == XmlReader::END_ELEMENT ) {
 235                                 break;
 236                         } elseif ( $this->reader->namespaceURI == self::NS_SVG && $this->reader->nodeType == XmlReader::ELEMENT ) {
 237                                 switch( $this->reader->localName ) {
 238                                         case 'script':
 239                                                 // Normally we disallow files with
 240                                                 // <script>, but its possible
 241                                                 // to configure MW to disable
 242                                                 // such checks.
 243                                         case 'animate':
 244                                         case 'set':
 245                                         case 'animateMotion':
 246                                         case 'animateColor':
 247                                         case 'animateTransform':
 248                                                 $this->debug( "HOUSTON WE HAVE ANIMATION" );
 249                                                 $this->metadata['animated'] = true;
 250                                                 break;
 251                                 }
 252                         }
 253                         $keepReading = $this->reader->read();
 254                 }
 255         }
 256
 257         private function throwXmlError( $err ) {
 258                 $this->debug( "FAILURE: $err" );
 259                 wfDebug( "SVGReader XML error: $err\n" );
 260         }
 261
 262         private function debug( $data ) {
 263                 if( $this->mDebug ) {
 264                         wfDebug( "SVGReader: $data\n" );
 265                 }
 266         }
 267
 268         private function warn( $data ) {
 269                 wfDebug( "SVGReader: $data\n" );
 270         }
 271
 272         private function notice( $data ) {
 273                 wfDebug( "SVGReader WARN: $data\n" );
 274         }
 275
 276         /**
 277          * Parse the attributes of an SVG element
 278          *
 279          * The parser has to be in the start element of "<svg>"
 280          */
 281         private function handleSVGAttribs( ) {
 282                 $defaultWidth = self::DEFAULT_WIDTH;
 283                 $defaultHeight = self::DEFAULT_HEIGHT;
 284                 $aspect = 1.0;
 285                 $width = null;
 286                 $height = null;
 287
 288                 if( $this->reader->getAttribute('viewBox') ) {
 289                         // min-x min-y width height
 290                         $viewBox = preg_split( '/\s+/', trim( $this->reader->getAttribute('viewBox') ) );
 291                         if( count( $viewBox ) == 4 ) {
 292                                 $viewWidth = $this->scaleSVGUnit( $viewBox[2] );
 293                                 $viewHeight = $this->scaleSVGUnit( $viewBox[3] );
 294                                 if( $viewWidth > 0 && $viewHeight > 0 ) {
 295                                         $aspect = $viewWidth / $viewHeight;
 296                                         $defaultHeight = $defaultWidth / $aspect;
 297                                 }
 298                         }
 299                 }
 300                 if( $this->reader->getAttribute('width') ) {
 301                         $width = $this->scaleSVGUnit( $this->reader->getAttribute('width'), $defaultWidth );
 302                         $this->metadata['originalWidth'] = $this->reader->getAttribute( 'width' );
 303                 }
 304                 if( $this->reader->getAttribute('height') ) {
 305                         $height = $this->scaleSVGUnit( $this->reader->getAttribute('height'), $defaultHeight );
 306                         $this->metadata['originalHeight'] = $this->reader->getAttribute( 'height' );
 307                 }
 308
 309                 if( !isset( $width ) && !isset( $height ) ) {
 310                         $width = $defaultWidth;
 311                         $height = $width / $aspect;
 312                 } elseif( isset( $width ) && !isset( $height ) ) {
 313                         $height = $width / $aspect;
 314                 } elseif( isset( $height ) && !isset( $width ) ) {
 315                         $width = $height * $aspect;
 316                 }
 317
 318                 if( $width > 0 && $height > 0 ) {
 319                         $this->metadata['width'] = intval( round( $width ) );
 320                         $this->metadata['height'] = intval( round( $height ) );
 321                 }
 322         }
 323
 324         /**
 325          * Return a rounded pixel equivalent for a labeled CSS/SVG length.
 326          * http://www.w3.org/TR/SVG11/coords.html#UnitIdentifiers
 327          *
 328          * @param $length String: CSS/SVG length.
 329          * @param $viewportSize: Float optional scale for percentage units...
 330          * @return float: length in pixels
 331          */
 332         static function scaleSVGUnit( $length, $viewportSize=512 ) {
 333                 static $unitLength = array(
 334                         'px' => 1.0,
 335                         'pt' => 1.25,
 336                         'pc' => 15.0,
 337                         'mm' => 3.543307,
 338                         'cm' => 35.43307,
 339                         'in' => 90.0,
 340                         'em' => 16.0, // fake it?
 341                         'ex' => 12.0, // fake it?
 342                         ''   => 1.0, // "User units" pixels by default
 343                         );
 344                 $matches = array();
 345                 if( preg_match( '/^\s*(\d+(?:\.\d+)?)(em|ex|px|pt|pc|cm|mm|in|%|)\s*$/', $length, $matches ) ) {
 346                         $length = floatval( $matches[1] );
 347                         $unit = $matches[2];
 348                         if( $unit == '%' ) {
 349                                 return $length * 0.01 * $viewportSize;
 350                         } else {
 351                                 return $length * $unitLength[$unit];
 352                         }
 353                 } else {
 354                         // Assume pixels
 355                         return floatval( $length );
 356                 }
 357         }
 358 }