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 ( $tag !== '#text' ) {
 143                                 $this->debug( "Unhandled top-level XML tag $tag" );
 144
 145                                 if ( !isset( $this->metadata['animated'] ) ) {
 146                                         // Recurse into children of current tag, looking for animation.
 147                                         $this->animateFilter( $tag );
 148                                 }
 149                         }
 150
 151                         // Goto next element, which is sibling of current (Skip children).
 152                         $keepReading = $this->reader->next();
 153                 }
 154
 155                 $this->reader->close();
 156
 157                 return true;
 158         }
 159
 160         /**
 161          * Read a textelement from an element
 162          *
 163          * @param String $name of the element that we are reading from
 164          * @param String $metafield that we will fill with the result
 165          */
 166         private function readField( $name, $metafield=null ) {
 167                 $this->debug ( "Read field $metafield" );
 168                 if( !$metafield || $this->reader->nodeType != XmlReader::ELEMENT ) {
 169                         return;
 170                 }
 171                 $keepReading = $this->reader->read();
 172                 while( $keepReading ) {
 173                         if( $this->reader->localName == $name && $this->reader->namespaceURI == self::NS_SVG && $this->reader->nodeType == XmlReader::END_ELEMENT ) {
 174                                 break;
 175                         } elseif( $this->reader->nodeType == XmlReader::TEXT ){
 176                                 $this->metadata[$metafield] = trim( $this->reader->value );
 177                         }
 178                         $keepReading = $this->reader->read();
 179                 }
 180         }
 181
 182         /**
 183          * Read an XML snippet from an element
 184          *
 185          * @param String $metafield that we will fill with the result
 186          */
 187         private function readXml( $metafield=null ) {
 188                 $this->debug ( "Read top level metadata" );
 189                 if( !$metafield || $this->reader->nodeType != XmlReader::ELEMENT ) {
 190                         return;
 191                 }
 192                 // TODO: find and store type of xml snippet. metadata['metadataType'] = "rdf"
 193                 if( method_exists( $this->reader, 'readInnerXML' ) ) {
 194                         $this->metadata[$metafield] = trim( $this->reader->readInnerXML() );
 195                 } else {
 196                         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)." );
 197                 }
 198                 $this->reader->next();
 199         }
 200
 201         /**
 202          * Filter all children, looking for animate elements
 203          *
 204          * @param String $name of the element that we are reading from
 205          */
 206         private function animateFilter( $name ) {
 207                 $this->debug ( "animate filter for tag $name" );
 208                 if( $this->reader->nodeType != XmlReader::ELEMENT ) {
 209                         return;
 210                 }
 211                 if ( $this->reader->isEmptyElement ) {
 212                         return;
 213                 }
 214                 $exitDepth =  $this->reader->depth;
 215                 $keepReading = $this->reader->read();
 216                 while( $keepReading ) {
 217                         if( $this->reader->localName == $name && $this->reader->depth <= $exitDepth
 218                                 && $this->reader->nodeType == XmlReader::END_ELEMENT ) {
 219                                 break;
 220                         } elseif ( $this->reader->namespaceURI == self::NS_SVG && $this->reader->nodeType == XmlReader::ELEMENT ) {
 221                                 switch( $this->reader->localName ) {
 222                                         case 'animate':
 223                                         case 'set':
 224                                         case 'animateMotion':
 225                                         case 'animateColor':
 226                                         case 'animateTransform':
 227                                                 $this->debug( "HOUSTON WE HAVE ANIMATION" );
 228                                                 $this->metadata['animated'] = true;
 229                                                 break;
 230                                 }
 231                         }
 232                         $keepReading = $this->reader->read();
 233                 }
 234         }
 235
 236         private function throwXmlError( $err ) {
 237                 $this->debug( "FAILURE: $err" );
 238                 wfDebug( "SVGReader XML error: $err\n" );
 239         }
 240
 241         private function debug( $data ) {
 242                 if( $this->mDebug ) {
 243                         wfDebug( "SVGReader: $data\n" );
 244                 }
 245         }
 246
 247         private function warn( $data ) {
 248                 wfDebug( "SVGReader: $data\n" );
 249         }
 250
 251         private function notice( $data ) {
 252                 wfDebug( "SVGReader WARN: $data\n" );
 253         }
 254
 255         /**
 256          * Parse the attributes of an SVG element
 257          *
 258          * The parser has to be in the start element of <svg>
 259          */
 260         private function handleSVGAttribs( ) {
 261                 $defaultWidth = self::DEFAULT_WIDTH;
 262                 $defaultHeight = self::DEFAULT_HEIGHT;
 263                 $aspect = 1.0;
 264                 $width = null;
 265                 $height = null;
 266
 267                 if( $this->reader->getAttribute('viewBox') ) {
 268                         // min-x min-y width height
 269                         $viewBox = preg_split( '/\s+/', trim( $this->reader->getAttribute('viewBox') ) );
 270                         if( count( $viewBox ) == 4 ) {
 271                                 $viewWidth = $this->scaleSVGUnit( $viewBox[2] );
 272                                 $viewHeight = $this->scaleSVGUnit( $viewBox[3] );
 273                                 if( $viewWidth > 0 && $viewHeight > 0 ) {
 274                                         $aspect = $viewWidth / $viewHeight;
 275                                         $defaultHeight = $defaultWidth / $aspect;
 276                                 }
 277                         }
 278                 }
 279                 if( $this->reader->getAttribute('width') ) {
 280                         $width = $this->scaleSVGUnit( $this->reader->getAttribute('width'), $defaultWidth );
 281                 }
 282                 if( $this->reader->getAttribute('height') ) {
 283                         $height = $this->scaleSVGUnit( $this->reader->getAttribute('height'), $defaultHeight );
 284                 }
 285
 286                 if( !isset( $width ) && !isset( $height ) ) {
 287                         $width = $defaultWidth;
 288                         $height = $width / $aspect;
 289                 } elseif( isset( $width ) && !isset( $height ) ) {
 290                         $height = $width / $aspect;
 291                 } elseif( isset( $height ) && !isset( $width ) ) {
 292                         $width = $height * $aspect;
 293                 }
 294
 295                 if( $width > 0 && $height > 0 ) {
 296                         $this->metadata['width'] = intval( round( $width ) );
 297                         $this->metadata['height'] = intval( round( $height ) );
 298                 }
 299         }
 300
 301         /**
 302          * Return a rounded pixel equivalent for a labeled CSS/SVG length.
 303          * http://www.w3.org/TR/SVG11/coords.html#UnitIdentifiers
 304          *
 305          * @param $length String: CSS/SVG length.
 306          * @param $viewportSize: Float optional scale for percentage units...
 307          * @return float: length in pixels
 308          */
 309         static function scaleSVGUnit( $length, $viewportSize=512 ) {
 310                 static $unitLength = array(
 311                         'px' => 1.0,
 312                         'pt' => 1.25,
 313                         'pc' => 15.0,
 314                         'mm' => 3.543307,
 315                         'cm' => 35.43307,
 316                         'in' => 90.0,
 317                         'em' => 16.0, // fake it?
 318                         'ex' => 12.0, // fake it?
 319                         ''   => 1.0, // "User units" pixels by default
 320                         );
 321                 $matches = array();
 322                 if( preg_match( '/^\s*(\d+(?:\.\d+)?)(em|ex|px|pt|pc|cm|mm|in|%|)\s*$/', $length, $matches ) ) {
 323                         $length = floatval( $matches[1] );
 324                         $unit = $matches[2];
 325                         if( $unit == '%' ) {
 326                                 return $length * 0.01 * $viewportSize;
 327                         } else {
 328                                 return $length * $unitLength[$unit];
 329                         }
 330                 } else {
 331                         // Assume pixels
 332                         return floatval( $length );
 333                 }
 334         }
 335 }