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
  35                 return $svg->getMetadata();
  36         }
  37 }
  38
  39 /**
  40  * @ingroup Media
  41  */
  42 class SVGReader {
  43         const DEFAULT_WIDTH = 512;
  44         const DEFAULT_HEIGHT = 512;
  45         const NS_SVG = 'http://www.w3.org/2000/svg';
  46
  47         /** @var null|XMLReader */
  48         private $reader = null;
  49
  50         /** @var bool */
  51         private $mDebug = false;
  52
  53         /** @var array */
  54         private $metadata = array();
  55
  56         /**
  57          * Constructor
  58          *
  59          * Creates an SVGReader drawing from the source provided
  60          * @param string $source URI from which to read
  61          * @throws MWException|Exception
  62          */
  63         function __construct( $source ) {
  64                 global $wgSVGMetadataCutoff;
  65                 $this->reader = new XMLReader();
  66
  67                 // Don't use $file->getSize() since file object passed to SVGHandler::getMetadata is bogus.
  68                 $size = filesize( $source );
  69                 if ( $size === false ) {
  70                         throw new MWException( "Error getting filesize of SVG." );
  71                 }
  72
  73                 if ( $size > $wgSVGMetadataCutoff ) {
  74                         $this->debug( "SVG is $size bytes, which is bigger than $wgSVGMetadataCutoff. Truncating." );
  75                         $contents = file_get_contents( $source, false, null, -1, $wgSVGMetadataCutoff );
  76                         if ( $contents === false ) {
  77                                 throw new MWException( 'Error reading SVG file.' );
  78                         }
  79                         $this->reader->XML( $contents, null, LIBXML_NOERROR | LIBXML_NOWARNING );
  80                 } else {
  81                         $this->reader->open( $source, null, LIBXML_NOERROR | LIBXML_NOWARNING );
  82                 }
  83
  84                 // Expand entities, since Adobe Illustrator uses them for xmlns
  85                 // attributes (bug 31719). Note that libxml2 has some protection
  86                 // against large recursive entity expansions so this is not as
  87                 // insecure as it might appear to be. However, it is still extremely
  88                 // insecure. It's necessary to wrap any read() calls with
  89                 // libxml_disable_entity_loader() to avoid arbitrary local file
  90                 // inclusion, or even arbitrary code execution if the expect
  91                 // extension is installed (bug 46859).
  92                 $oldDisable = libxml_disable_entity_loader( true );
  93                 $this->reader->setParserProperty( XMLReader::SUBST_ENTITIES, true );
  94
  95                 $this->metadata['width'] = self::DEFAULT_WIDTH;
  96                 $this->metadata['height'] = self::DEFAULT_HEIGHT;
  97
  98                 // The size in the units specified by the SVG file
  99                 // (for the metadata box)
 100                 // Per the SVG spec, if unspecified, default to '100%'
 101                 $this->metadata['originalWidth'] = '100%';
 102                 $this->metadata['originalHeight'] = '100%';
 103
 104                 // Because we cut off the end of the svg making an invalid one. Complicated
 105                 // try catch thing to make sure warnings get restored. Seems like there should
 106                 // be a better way.
 107                 wfSuppressWarnings();
 108                 try {
 109                         $this->read();
 110                 } catch ( Exception $e ) {
 111                         // Note, if this happens, the width/height will be taken to be 0x0.
 112                         // Should we consider it the default 512x512 instead?
 113                         wfRestoreWarnings();
 114                         libxml_disable_entity_loader( $oldDisable );
 115                         throw $e;
 116                 }
 117                 wfRestoreWarnings();
 118                 libxml_disable_entity_loader( $oldDisable );
 119         }
 120
 121         /**
 122          * @return Array with the known metadata
 123          */
 124         public function getMetadata() {
 125                 return $this->metadata;
 126         }
 127
 128         /**
 129          * Read the SVG
 130          * @throws MWException
 131          * @return bool
 132          */
 133         protected function read() {
 134                 $keepReading = $this->reader->read();
 135
 136                 /* Skip until first element */
 137                 while ( $keepReading && $this->reader->nodeType != XmlReader::ELEMENT ) {
 138                         $keepReading = $this->reader->read();
 139                 }
 140
 141                 if ( $this->reader->localName != 'svg' || $this->reader->namespaceURI != self::NS_SVG ) {
 142                         throw new MWException( "Expected <svg> tag, got " .
 143                                 $this->reader->localName . " in NS " . $this->reader->namespaceURI );
 144                 }
 145                 $this->debug( "<svg> tag is correct." );
 146                 $this->handleSVGAttribs();
 147
 148                 $exitDepth = $this->reader->depth;
 149                 $keepReading = $this->reader->read();
 150                 while ( $keepReading ) {
 151                         $tag = $this->reader->localName;
 152                         $type = $this->reader->nodeType;
 153                         $isSVG = ( $this->reader->namespaceURI == self::NS_SVG );
 154
 155                         $this->debug( "$tag" );
 156
 157                         if ( $isSVG && $tag == 'svg' && $type == XmlReader::END_ELEMENT
 158                                 && $this->reader->depth <= $exitDepth
 159                         ) {
 160                                 break;
 161                         } elseif ( $isSVG && $tag == 'title' ) {
 162                                 $this->readField( $tag, 'title' );
 163                         } elseif ( $isSVG && $tag == 'desc' ) {
 164                                 $this->readField( $tag, 'description' );
 165                         } elseif ( $isSVG && $tag == 'metadata' && $type == XmlReader::ELEMENT ) {
 166                                 $this->readXml( $tag, 'metadata' );
 167                         } elseif ( $isSVG && $tag == 'script' ) {
 168                                 // We normally do not allow scripted svgs.
 169                                 // However its possible to configure MW to let them
 170                                 // in, and such files should be considered animated.
 171                                 $this->metadata['animated'] = true;
 172                         } elseif ( $tag !== '#text' ) {
 173                                 $this->debug( "Unhandled top-level XML tag $tag" );
 174
 175                                 if ( !isset( $this->metadata['animated'] ) ) {
 176                                         // Recurse into children of current tag, looking for animation.
 177                                         $this->animateFilter( $tag );
 178                                 }
 179                         }
 180
 181                         // Goto next element, which is sibling of current (Skip children).
 182                         $keepReading = $this->reader->next();
 183                 }
 184
 185                 $this->reader->close();
 186
 187                 return true;
 188         }
 189
 190         /**
 191          * Read a textelement from an element
 192          *
 193          * @param string $name Name of the element that we are reading from
 194          * @param string $metafield Field that we will fill with the result
 195          */
 196         private function readField( $name, $metafield = null ) {
 197                 $this->debug( "Read field $metafield" );
 198                 if ( !$metafield || $this->reader->nodeType != XmlReader::ELEMENT ) {
 199                         return;
 200                 }
 201                 $keepReading = $this->reader->read();
 202                 while ( $keepReading ) {
 203                         if ( $this->reader->localName == $name
 204                                 && $this->reader->namespaceURI == self::NS_SVG
 205                                 && $this->reader->nodeType == XmlReader::END_ELEMENT
 206                         ) {
 207                                 break;
 208                         } elseif ( $this->reader->nodeType == XmlReader::TEXT ) {
 209                                 $this->metadata[$metafield] = trim( $this->reader->value );
 210                         }
 211                         $keepReading = $this->reader->read();
 212                 }
 213         }
 214
 215         /**
 216          * Read an XML snippet from an element
 217          *
 218          * @param string $metafield Field that we will fill with the result
 219          * @throws MWException
 220          */
 221         private function readXml( $metafield = null ) {
 222                 $this->debug( "Read top level metadata" );
 223                 if ( !$metafield || $this->reader->nodeType != XmlReader::ELEMENT ) {
 224                         return;
 225                 }
 226                 // @todo Find and store type of xml snippet. metadata['metadataType'] = "rdf"
 227                 if ( method_exists( $this->reader, 'readInnerXML' ) ) {
 228                         $this->metadata[$metafield] = trim( $this->reader->readInnerXML() );
 229                 } else {
 230                         throw new MWException( "The PHP XMLReader extension does not come " .
 231                                 "with readInnerXML() method. Your libxml is probably out of " .
 232                                 "date (need 2.6.20 or later)." );
 233                 }
 234                 $this->reader->next();
 235         }
 236
 237         /**
 238          * Filter all children, looking for animate elements
 239          *
 240          * @param string $name Name of the element that we are reading from
 241          */
 242         private function animateFilter( $name ) {
 243                 $this->debug( "animate filter for tag $name" );
 244                 if ( $this->reader->nodeType != XmlReader::ELEMENT ) {
 245                         return;
 246                 }
 247                 if ( $this->reader->isEmptyElement ) {
 248                         return;
 249                 }
 250                 $exitDepth = $this->reader->depth;
 251                 $keepReading = $this->reader->read();
 252                 while ( $keepReading ) {
 253                         if ( $this->reader->localName == $name && $this->reader->depth <= $exitDepth
 254                                 && $this->reader->nodeType == XmlReader::END_ELEMENT
 255                         ) {
 256                                 break;
 257                         } elseif ( $this->reader->namespaceURI ==
 258                                 self::NS_SVG && $this->reader->nodeType == XmlReader::ELEMENT
 259                         ) {
 260                                 switch ( $this->reader->localName ) {
 261                                         case 'script':
 262                                                 // Normally we disallow files with
 263                                                 // <script>, but its possible
 264                                                 // to configure MW to disable
 265                                                 // such checks.
 266                                         case 'animate':
 267                                         case 'set':
 268                                         case 'animateMotion':
 269                                         case 'animateColor':
 270                                         case 'animateTransform':
 271                                                 $this->debug( "HOUSTON WE HAVE ANIMATION" );
 272                                                 $this->metadata['animated'] = true;
 273                                                 break;
 274                                 }
 275                         }
 276                         $keepReading = $this->reader->read();
 277                 }
 278         }
 279
 280         // @todo FIXME: Unused, remove?
 281         private function throwXmlError( $err ) {
 282                 $this->debug( "FAILURE: $err" );
 283                 wfDebug( "SVGReader XML error: $err\n" );
 284         }
 285
 286         private function debug( $data ) {
 287                 if ( $this->mDebug ) {
 288                         wfDebug( "SVGReader: $data\n" );
 289                 }
 290         }
 291
 292         // @todo FIXME: Unused, remove?
 293         private function warn( $data ) {
 294                 wfDebug( "SVGReader: $data\n" );
 295         }
 296
 297         // @todo FIXME: Unused, remove?
 298         private function notice( $data ) {
 299                 wfDebug( "SVGReader WARN: $data\n" );
 300         }
 301
 302         /**
 303          * Parse the attributes of an SVG element
 304          *
 305          * The parser has to be in the start element of "<svg>"
 306          */
 307         private function handleSVGAttribs() {
 308                 $defaultWidth = self::DEFAULT_WIDTH;
 309                 $defaultHeight = self::DEFAULT_HEIGHT;
 310                 $aspect = 1.0;
 311                 $width = null;
 312                 $height = null;
 313
 314                 if ( $this->reader->getAttribute( 'viewBox' ) ) {
 315                         // min-x min-y width height
 316                         $viewBox = preg_split( '/\s+/', trim( $this->reader->getAttribute( 'viewBox' ) ) );
 317                         if ( count( $viewBox ) == 4 ) {
 318                                 $viewWidth = $this->scaleSVGUnit( $viewBox[2] );
 319                                 $viewHeight = $this->scaleSVGUnit( $viewBox[3] );
 320                                 if ( $viewWidth > 0 && $viewHeight > 0 ) {
 321                                         $aspect = $viewWidth / $viewHeight;
 322                                         $defaultHeight = $defaultWidth / $aspect;
 323                                 }
 324                         }
 325                 }
 326                 if ( $this->reader->getAttribute( 'width' ) ) {
 327                         $width = $this->scaleSVGUnit( $this->reader->getAttribute( 'width' ), $defaultWidth );
 328                         $this->metadata['originalWidth'] = $this->reader->getAttribute( 'width' );
 329                 }
 330                 if ( $this->reader->getAttribute( 'height' ) ) {
 331                         $height = $this->scaleSVGUnit( $this->reader->getAttribute( 'height' ), $defaultHeight );
 332                         $this->metadata['originalHeight'] = $this->reader->getAttribute( 'height' );
 333                 }
 334
 335                 if ( !isset( $width ) && !isset( $height ) ) {
 336                         $width = $defaultWidth;
 337                         $height = $width / $aspect;
 338                 } elseif ( isset( $width ) && !isset( $height ) ) {
 339                         $height = $width / $aspect;
 340                 } elseif ( isset( $height ) && !isset( $width ) ) {
 341                         $width = $height * $aspect;
 342                 }
 343
 344                 if ( $width > 0 && $height > 0 ) {
 345                         $this->metadata['width'] = intval( round( $width ) );
 346                         $this->metadata['height'] = intval( round( $height ) );
 347                 }
 348         }
 349
 350         /**
 351          * Return a rounded pixel equivalent for a labeled CSS/SVG length.
 352          * http://www.w3.org/TR/SVG11/coords.html#UnitIdentifiers
 353          *
 354          * @param string $length CSS/SVG length.
 355          * @param float|int $viewportSize Optional scale for percentage units...
 356          * @return float Length in pixels
 357          */
 358         static function scaleSVGUnit( $length, $viewportSize = 512 ) {
 359                 static $unitLength = array(
 360                         'px' => 1.0,
 361                         'pt' => 1.25,
 362                         'pc' => 15.0,
 363                         'mm' => 3.543307,
 364                         'cm' => 35.43307,
 365                         'in' => 90.0,
 366                         'em' => 16.0, // fake it?
 367                         'ex' => 12.0, // fake it?
 368                         '' => 1.0, // "User units" pixels by default
 369                 );
 370                 $matches = array();
 371                 if ( preg_match( '/^\s*(\d+(?:\.\d+)?)(em|ex|px|pt|pc|cm|mm|in|%|)\s*$/', $length, $matches ) ) {
 372                         $length = floatval( $matches[1] );
 373                         $unit = $matches[2];
 374                         if ( $unit == '%' ) {
 375                                 return $length * 0.01 * $viewportSize;
 376                         } else {
 377                                 return $length * $unitLength[$unit];
 378                         }
 379                 } else {
 380                         // Assume pixels
 381                         return floatval( $length );
 382                 }
 383         }
 384 }