includes/media/SVGMetadataExtractor.php

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