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