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