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
  39         private $reader = null;
  40         private $mDebug = false;
  41         private $metadata = Array();
  42
  43         /**
  44          * Constructor
  45          *
  46          * Creates an SVGReader drawing from the source provided
  47          * @param $source String: URI from which to read
  48          */
  49         function __construct( $source ) {
  50                 global $wgSVGMetadataCutoff;
  51                 $this->reader = new XMLReader();
  52
  53                 // Don't use $file->getSize() since file object passed to SVGHandler::getMetadata is bogus.
  54                 $size = filesize( $source );
  55                 if ( $size === false ) {
  56                         throw new MWException( "Error getting filesize of SVG." );
  57                 }
  58
  59                 if ( $size > $wgSVGMetadataCutoff ) {
  60                         $this->debug( "SVG is $size bytes, which is bigger than $wgSVGMetadataCutoff. Truncating." );
  61                         $contents = file_get_contents( $source, false, null, -1, $wgSVGMetadataCutoff );
  62                         if ($contents === false) {
  63                                 throw new MWException( 'Error reading SVG file.' );
  64                         }
  65                         $this->reader->XML( $contents, null, LIBXML_NOERROR | LIBXML_NOWARNING );
  66                 } else {
  67                         $this->reader->open( $source, null, LIBXML_NOERROR | LIBXML_NOWARNING );
  68                 }
  69
  70                 $this->metadata['width'] = self::DEFAULT_WIDTH;
  71                 $this->metadata['height'] = self::DEFAULT_HEIGHT;
  72
  73                 // Because we cut off the end of the svg making an invalid one. Complicated
  74                 // try catch thing to make sure warnings get restored. Seems like there should
  75                 // be a better way.
  76                 wfSuppressWarnings();
  77                 try {
  78                         $this->read();
  79                 } catch( Exception $e ) {
  80                         wfRestoreWarnings();
  81                         throw $e;
  82                 }
  83                 wfRestoreWarnings();
  84         }
  85
  86         /**
  87          * @return Array with the known metadata
  88          */
  89         public function getMetadata() {
  90                 return $this->metadata;
  91         }
  92
  93         /**
  94          * Read the SVG
  95          */
  96         public function read() {
  97                 $keepReading = $this->reader->read();
  98
  99                 /* Skip until first element */
 100                 while( $keepReading && $this->reader->nodeType != XmlReader::ELEMENT ) {
 101                         $keepReading = $this->reader->read();
 102                 }
 103
 104                 if ( !$this->qualifiedNameEquals( $this->reader->name, 'svg', 'svg' ) ) {
 105                         throw new MWException( "Expected <svg> tag, got ".
 106                                 $this->reader->name );
 107                 }
 108                 $this->debug( "<svg> tag is correct." );
 109                 $this->handleSVGAttribs();
 110
 111                 $exitDepth =  $this->reader->depth;
 112                 $keepReading = $this->reader->read();
 113                 while ( $keepReading ) {
 114                         $tag = $this->reader->name;
 115                         $type = $this->reader->nodeType;
 116
 117                         $this->debug( "$tag" );
 118
 119                         if ( $this->qualifiedNameEquals( $tag, 'svg', 'svg' ) && $type == XmlReader::END_ELEMENT && $this->reader->depth <= $exitDepth ) {
 120                                 break;
 121                         } elseif ( $this->qualifiedNameEquals( $tag, 'svg', 'title' ) ) {
 122                                 $this->readField( $tag, 'title' );
 123                         } elseif ( $this->qualifiedNameEquals( $tag, 'svg', 'desc' )  ) {
 124                                 $this->readField( $tag, 'description' );
 125                         } elseif ( $this->qualifiedNameEquals( $tag, 'svg', 'metadata' ) && $type == XmlReader::ELEMENT ) {
 126                                 $this->readXml( $tag, 'metadata' );
 127                         } elseif ( $tag !== '#text' ) {
 128                                 $this->debug( "Unhandled top-level XML tag $tag" );
 129
 130                                 if ( !isset( $this->metadata['animated'] ) ) {
 131                                         // Recurse into children of current tag, looking for animation.
 132                                         $this->animateFilter( $tag );
 133                                 }
 134                         }
 135
 136                         // Goto next element, which is sibling of current (Skip children).
 137                         $keepReading = $this->reader->next();
 138                 }
 139
 140                 $this->reader->close();
 141
 142                 return true;
 143         }
 144
 145         /**
 146          * Read a textelement from an element
 147          *
 148          * @param String $name of the element that we are reading from
 149          * @param String $metafield that we will fill with the result
 150          */
 151         private function readField( $name, $metafield=null ) {
 152                 $this->debug ( "Read field $metafield" );
 153                 if( !$metafield || $this->reader->nodeType != XmlReader::ELEMENT ) {
 154                         return;
 155                 }
 156                 $keepReading = $this->reader->read();
 157                 while( $keepReading ) {
 158                         if( $this->reader->name == $name && $this->reader->nodeType == XmlReader::END_ELEMENT ) {
 159                                 break;
 160                         } elseif( $this->reader->nodeType == XmlReader::TEXT ){
 161                                 $this->metadata[$metafield] = trim( $this->reader->value );
 162                         }
 163                         $keepReading = $this->reader->read();
 164                 }
 165         }
 166
 167         /*
 168          * Read an XML snippet from an element
 169          *
 170          * @param String $metafield that we will fill with the result
 171          */
 172         private function readXml( $metafield=null ) {
 173                 $this->debug ( "Read top level metadata" );
 174                 if( !$metafield || $this->reader->nodeType != XmlReader::ELEMENT ) {
 175                         return;
 176                 }
 177                 // TODO: find and store type of xml snippet. metadata['metadataType'] = "rdf"
 178                 $this->metadata[$metafield] = $this->reader->readInnerXML();
 179                 $this->reader->next();
 180         }
 181
 182         /**
 183          * Filter all children, looking for animate elements
 184          *
 185          * @param String $name of the element that we are reading from
 186          */
 187         private function animateFilter( $name ) {
 188                 $this->debug ( "animate filter for tag $name" );
 189                 if( $this->reader->nodeType != XmlReader::ELEMENT ) {
 190                         return;
 191                 }
 192                 if ( $this->reader->isEmptyElement ) {
 193                         return;
 194                 }
 195                 $exitDepth =  $this->reader->depth;
 196                 $keepReading = $this->reader->read();
 197                 while( $keepReading ) {
 198                         if( $this->reader->name == $name && $this->reader->depth <= $exitDepth
 199                                 && $this->reader->nodeType == XmlReader::END_ELEMENT ) {
 200                                 break;
 201                         } elseif ( $this->reader->nodeType == XmlReader::ELEMENT ) {
 202                                 switch( $this->reader->name ) {
 203                                         case 'animate':
 204                                         case 'svg:animate':
 205                                         case 'set':
 206                                         case 'svg:set':
 207                                         case 'animateMotion':
 208                                         case 'svg:animateMotion':
 209                                         case 'animateColor':
 210                                         case 'svg:animateColor':
 211                                         case 'animateTransform':
 212                                         case 'svg:animateTransform':
 213                                                 $this->debug( "HOUSTON WE HAVE ANIMATION" );
 214                                                 $this->metadata['animated'] = true;
 215                                                 break;
 216                                 }
 217                         }
 218                         $keepReading = $this->reader->read();
 219                 }
 220         }
 221
 222         private function throwXmlError( $err ) {
 223                 $this->debug( "FAILURE: $err" );
 224                 wfDebug( "SVGReader XML error: $err\n" );
 225         }
 226
 227         private function debug( $data ) {
 228                 if( $this->mDebug ) {
 229                         wfDebug( "SVGReader: $data\n" );
 230                 }
 231         }
 232
 233         private function warn( $data ) {
 234                 wfDebug( "SVGReader: $data\n" );
 235         }
 236
 237         private function notice( $data ) {
 238                 wfDebug( "SVGReader WARN: $data\n" );
 239         }
 240
 241         /**
 242          * Parse the attributes of an SVG element
 243          *
 244          * The parser has to be in the start element of <svg>
 245          */
 246         private function handleSVGAttribs( ) {
 247                 $defaultWidth = self::DEFAULT_WIDTH;
 248                 $defaultHeight = self::DEFAULT_HEIGHT;
 249                 $aspect = 1.0;
 250                 $width = null;
 251                 $height = null;
 252
 253                 if( $this->reader->getAttribute('viewBox') ) {
 254                         // min-x min-y width height
 255                         $viewBox = preg_split( '/\s+/', trim( $this->reader->getAttribute('viewBox') ) );
 256                         if( count( $viewBox ) == 4 ) {
 257                                 $viewWidth = $this->scaleSVGUnit( $viewBox[2] );
 258                                 $viewHeight = $this->scaleSVGUnit( $viewBox[3] );
 259                                 if( $viewWidth > 0 && $viewHeight > 0 ) {
 260                                         $aspect = $viewWidth / $viewHeight;
 261                                         $defaultHeight = $defaultWidth / $aspect;
 262                                 }
 263                         }
 264                 }
 265                 if( $this->reader->getAttribute('width') ) {
 266                         $width = $this->scaleSVGUnit( $this->reader->getAttribute('width'), $defaultWidth );
 267                 }
 268                 if( $this->reader->getAttribute('height') ) {
 269                         $height = $this->scaleSVGUnit( $this->reader->getAttribute('height'), $defaultHeight );
 270                 }
 271
 272                 if( !isset( $width ) && !isset( $height ) ) {
 273                         $width = $defaultWidth;
 274                         $height = $width / $aspect;
 275                 } elseif( isset( $width ) && !isset( $height ) ) {
 276                         $height = $width / $aspect;
 277                 } elseif( isset( $height ) && !isset( $width ) ) {
 278                         $width = $height * $aspect;
 279                 }
 280
 281                 if( $width > 0 && $height > 0 ) {
 282                         $this->metadata['width'] = intval( round( $width ) );
 283                         $this->metadata['height'] = intval( round( $height ) );
 284                 }
 285         }
 286
 287         /**
 288          * Return a rounded pixel equivalent for a labeled CSS/SVG length.
 289          * http://www.w3.org/TR/SVG11/coords.html#UnitIdentifiers
 290          *
 291          * @param $length String: CSS/SVG length.
 292          * @param $viewportSize: Float optional scale for percentage units...
 293          * @return float: length in pixels
 294          */
 295         static function scaleSVGUnit( $length, $viewportSize=512 ) {
 296                 static $unitLength = array(
 297                         'px' => 1.0,
 298                         'pt' => 1.25,
 299                         'pc' => 15.0,
 300                         'mm' => 3.543307,
 301                         'cm' => 35.43307,
 302                         'in' => 90.0,
 303                         'em' => 16.0, // fake it?
 304                         'ex' => 12.0, // fake it?
 305                         ''   => 1.0, // "User units" pixels by default
 306                         );
 307                 $matches = array();
 308                 if( preg_match( '/^\s*(\d+(?:\.\d+)?)(em|ex|px|pt|pc|cm|mm|in|%|)\s*$/', $length, $matches ) ) {
 309                         $length = floatval( $matches[1] );
 310                         $unit = $matches[2];
 311                         if( $unit == '%' ) {
 312                                 return $length * 0.01 * $viewportSize;
 313                         } else {
 314                                 return $length * $unitLength[$unit];
 315                         }
 316                 } else {
 317                         // Assume pixels
 318                         return floatval( $length );
 319                 }
 320         }
 321
 322         /**
 323          * XML namespace check
 324          *
 325          * Check if a read node name matches the expected nodeName
 326          * @param $qualifiedName as read by XMLReader
 327          * @param $prefix the namespace prefix that you expect for this element, defaults to svg namespace
 328          * @param $localName the localName part of the element that you want to match
 329          *
 330          * @return boolean
 331          */
 332         private function qualifiedNameEquals( $qualifiedName, $prefix="svg", $localName ) {
 333                 if( ($qualifiedName == $localName && $prefix == "svg" ) ||
 334                      $qualifiedName == ($prefix . ":" . $localName) ) {
 335                         return true;
 336                 }
 337                 return false;
 338         }
 339 }