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