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 string $source 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. However, it is still extremely
  82                 // insecure. It's necessary to wrap any read() calls with
  83                 // libxml_disable_entity_loader() to avoid arbitrary local file
  84                 // inclusion, or even arbitrary code execution if the expect
  85                 // extension is installed (bug 46859).
  86                 $oldDisable = libxml_disable_entity_loader( true );
  87                 $this->reader->setParserProperty( XMLReader::SUBST_ENTITIES, true );
  88
  89                 $this->metadata['width'] = self::DEFAULT_WIDTH;
  90                 $this->metadata['height'] = self::DEFAULT_HEIGHT;
  91
  92                 // The size in the units specified by the SVG file
  93                 // (for the metadata box)
  94                 // Per the SVG spec, if unspecified, default to '100%'
  95                 $this->metadata['originalWidth'] = '100%';
  96                 $this->metadata['originalHeight'] = '100%';
  97
  98                 // Because we cut off the end of the svg making an invalid one. Complicated
  99                 // try catch thing to make sure warnings get restored. Seems like there should
 100                 // be a better way.
 101                 wfSuppressWarnings();
 102                 try {
 103                         $this->read();
 104                 } catch ( Exception $e ) {
 105                         // Note, if this happens, the width/height will be taken to be 0x0.
 106                         // Should we consider it the default 512x512 instead?
 107                         wfRestoreWarnings();
 108                         libxml_disable_entity_loader( $oldDisable );
 109                         throw $e;
 110                 }
 111                 wfRestoreWarnings();
 112                 libxml_disable_entity_loader( $oldDisable );
 113         }
 114
 115         /**
 116          * @return Array with the known metadata
 117          */
 118         public function getMetadata() {
 119                 return $this->metadata;
 120         }
 121
 122         /**
 123          * Read the SVG
 124          * @throws MWException
 125          * @return bool
 126          */
 127         protected function read() {
 128                 $keepReading = $this->reader->read();
 129
 130                 /* Skip until first element */
 131                 while ( $keepReading && $this->reader->nodeType != XmlReader::ELEMENT ) {
 132                         $keepReading = $this->reader->read();
 133                 }
 134
 135                 if ( $this->reader->localName != 'svg' || $this->reader->namespaceURI != self::NS_SVG ) {
 136                         throw new MWException( "Expected <svg> tag, got " .
 137                                 $this->reader->localName . " in NS " . $this->reader->namespaceURI );
 138                 }
 139                 $this->debug( "<svg> tag is correct." );
 140                 $this->handleSVGAttribs();
 141
 142                 $exitDepth = $this->reader->depth;
 143                 $keepReading = $this->reader->read();
 144                 while ( $keepReading ) {
 145                         $tag = $this->reader->localName;
 146                         $type = $this->reader->nodeType;
 147                         $isSVG = ( $this->reader->namespaceURI == self::NS_SVG );
 148
 149                         $this->debug( "$tag" );
 150
 151                         if ( $isSVG && $tag == 'svg' && $type == XmlReader::END_ELEMENT && $this->reader->depth <= $exitDepth ) {
 152                                 break;
 153                         } elseif ( $isSVG && $tag == 'title' ) {
 154                                 $this->readField( $tag, 'title' );
 155                         } elseif ( $isSVG && $tag == 'desc' ) {
 156                                 $this->readField( $tag, 'description' );
 157                         } elseif ( $isSVG && $tag == 'metadata' && $type == XmlReader::ELEMENT ) {
 158                                 $this->readXml( $tag, 'metadata' );
 159                         } elseif ( $isSVG && $tag == 'script' ) {
 160                                 // We normally do not allow scripted svgs.
 161                                 // However its possible to configure MW to let them
 162                                 // in, and such files should be considered animated.
 163                                 $this->metadata['animated'] = true;
 164                         } elseif ( $tag !== '#text' ) {
 165                                 $this->debug( "Unhandled top-level XML tag $tag" );
 166
 167                                 if ( !isset( $this->metadata['animated'] ) ) {
 168                                         // Recurse into children of current tag, looking for animation.
 169                                         $this->animateFilter( $tag );
 170                                 }
 171                         }
 172
 173                         // Goto next element, which is sibling of current (Skip children).
 174                         $keepReading = $this->reader->next();
 175                 }
 176
 177                 $this->reader->close();
 178
 179                 return true;
 180         }
 181
 182         /**
 183          * Read a textelement from an element
 184          *
 185          * @param string $name of the element that we are reading from
 186          * @param string $metafield that we will fill with the result
 187          */
 188         private function readField( $name, $metafield = null ) {
 189                 $this->debug( "Read field $metafield" );
 190                 if ( !$metafield || $this->reader->nodeType != XmlReader::ELEMENT ) {
 191                         return;
 192                 }
 193                 $keepReading = $this->reader->read();
 194                 while ( $keepReading ) {
 195                         if ( $this->reader->localName == $name && $this->reader->namespaceURI == self::NS_SVG && $this->reader->nodeType == XmlReader::END_ELEMENT ) {
 196                                 break;
 197                         } elseif ( $this->reader->nodeType == XmlReader::TEXT ) {
 198                                 $this->metadata[$metafield] = trim( $this->reader->value );
 199                         }
 200                         $keepReading = $this->reader->read();
 201                 }
 202         }
 203
 204         /**
 205          * Read an XML snippet from an element
 206          *
 207          * @param string $metafield that we will fill with the result
 208          * @throws MWException
 209          */
 210         private function readXml( $metafield = null ) {
 211                 $this->debug( "Read top level metadata" );
 212                 if ( !$metafield || $this->reader->nodeType != XmlReader::ELEMENT ) {
 213                         return;
 214                 }
 215                 // TODO: find and store type of xml snippet. metadata['metadataType'] = "rdf"
 216                 if ( method_exists( $this->reader, 'readInnerXML' ) ) {
 217                         $this->metadata[$metafield] = trim( $this->reader->readInnerXML() );
 218                 } else {
 219                         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)." );
 220                 }
 221                 $this->reader->next();
 222         }
 223
 224         /**
 225          * Filter all children, looking for animate elements
 226          *
 227          * @param string $name of the element that we are reading from
 228          */
 229         private function animateFilter( $name ) {
 230                 $this->debug( "animate filter for tag $name" );
 231                 if ( $this->reader->nodeType != XmlReader::ELEMENT ) {
 232                         return;
 233                 }
 234                 if ( $this->reader->isEmptyElement ) {
 235                         return;
 236                 }
 237                 $exitDepth = $this->reader->depth;
 238                 $keepReading = $this->reader->read();
 239                 while ( $keepReading ) {
 240                         if ( $this->reader->localName == $name && $this->reader->depth <= $exitDepth
 241                                 && $this->reader->nodeType == XmlReader::END_ELEMENT ) {
 242                                 break;
 243                         } elseif ( $this->reader->namespaceURI == self::NS_SVG && $this->reader->nodeType == XmlReader::ELEMENT ) {
 244                                 switch ( $this->reader->localName ) {
 245                                         case 'script':
 246                                                 // Normally we disallow files with
 247                                                 // <script>, but its possible
 248                                                 // to configure MW to disable
 249                                                 // such checks.
 250                                         case 'animate':
 251                                         case 'set':
 252                                         case 'animateMotion':
 253                                         case 'animateColor':
 254                                         case 'animateTransform':
 255                                                 $this->debug( "HOUSTON WE HAVE ANIMATION" );
 256                                                 $this->metadata['animated'] = true;
 257                                                 break;
 258                                 }
 259                         }
 260                         $keepReading = $this->reader->read();
 261                 }
 262         }
 263
 264         private function throwXmlError( $err ) {
 265                 $this->debug( "FAILURE: $err" );
 266                 wfDebug( "SVGReader XML error: $err\n" );
 267         }
 268
 269         private function debug( $data ) {
 270                 if ( $this->mDebug ) {
 271                         wfDebug( "SVGReader: $data\n" );
 272                 }
 273         }
 274
 275         private function warn( $data ) {
 276                 wfDebug( "SVGReader: $data\n" );
 277         }
 278
 279         private function notice( $data ) {
 280                 wfDebug( "SVGReader WARN: $data\n" );
 281         }
 282
 283         /**
 284          * Parse the attributes of an SVG element
 285          *
 286          * The parser has to be in the start element of "<svg>"
 287          */
 288         private function handleSVGAttribs() {
 289                 $defaultWidth = self::DEFAULT_WIDTH;
 290                 $defaultHeight = self::DEFAULT_HEIGHT;
 291                 $aspect = 1.0;
 292                 $width = null;
 293                 $height = null;
 294
 295                 if ( $this->reader->getAttribute( 'viewBox' ) ) {
 296                         // min-x min-y width height
 297                         $viewBox = preg_split( '/\s+/', trim( $this->reader->getAttribute( 'viewBox' ) ) );
 298                         if ( count( $viewBox ) == 4 ) {
 299                                 $viewWidth = $this->scaleSVGUnit( $viewBox[2] );
 300                                 $viewHeight = $this->scaleSVGUnit( $viewBox[3] );
 301                                 if ( $viewWidth > 0 && $viewHeight > 0 ) {
 302                                         $aspect = $viewWidth / $viewHeight;
 303                                         $defaultHeight = $defaultWidth / $aspect;
 304                                 }
 305                         }
 306                 }
 307                 if ( $this->reader->getAttribute( 'width' ) ) {
 308                         $width = $this->scaleSVGUnit( $this->reader->getAttribute( 'width' ), $defaultWidth );
 309                         $this->metadata['originalWidth'] = $this->reader->getAttribute( 'width' );
 310                 }
 311                 if ( $this->reader->getAttribute( 'height' ) ) {
 312                         $height = $this->scaleSVGUnit( $this->reader->getAttribute( 'height' ), $defaultHeight );
 313                         $this->metadata['originalHeight'] = $this->reader->getAttribute( 'height' );
 314                 }
 315
 316                 if ( !isset( $width ) && !isset( $height ) ) {
 317                         $width = $defaultWidth;
 318                         $height = $width / $aspect;
 319                 } elseif ( isset( $width ) && !isset( $height ) ) {
 320                         $height = $width / $aspect;
 321                 } elseif ( isset( $height ) && !isset( $width ) ) {
 322                         $width = $height * $aspect;
 323                 }
 324
 325                 if ( $width > 0 && $height > 0 ) {
 326                         $this->metadata['width'] = intval( round( $width ) );
 327                         $this->metadata['height'] = intval( round( $height ) );
 328                 }
 329         }
 330
 331         /**
 332          * Return a rounded pixel equivalent for a labeled CSS/SVG length.
 333          * http://www.w3.org/TR/SVG11/coords.html#UnitIdentifiers
 334          *
 335          * @param string $length CSS/SVG length.
 336          * @param $viewportSize: Float optional scale for percentage units...
 337          * @return float: length in pixels
 338          */
 339         static function scaleSVGUnit( $length, $viewportSize = 512 ) {
 340                 static $unitLength = array(
 341                         'px' => 1.0,
 342                         'pt' => 1.25,
 343                         'pc' => 15.0,
 344                         'mm' => 3.543307,
 345                         'cm' => 35.43307,
 346                         'in' => 90.0,
 347                         'em' => 16.0, // fake it?
 348                         'ex' => 12.0, // fake it?
 349                         '' => 1.0, // "User units" pixels by default
 350                         );
 351                 $matches = array();
 352                 if ( preg_match( '/^\s*(\d+(?:\.\d+)?)(em|ex|px|pt|pc|cm|mm|in|%|)\s*$/', $length, $matches ) ) {
 353                         $length = floatval( $matches[1] );
 354                         $unit = $matches[2];
 355                         if ( $unit == '%' ) {
 356                                 return $length * 0.01 * $viewportSize;
 357                         } else {
 358                                 return $length * $unitLength[$unit];
 359                         }
 360                 } else {
 361                         // Assume pixels
 362                         return floatval( $length );
 363                 }
 364         }
 365 }