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         const LANG_PREFIX_MATCH = 1;
  47         const LANG_FULL_MATCH = 2;
  48
  49         /** @var null|XMLReader */
  50         private $reader = null;
  51
  52         /** @var bool */
  53         private $mDebug = false;
  54
  55         /** @var array */
  56         private $metadata = [];
  57         private $languages = [];
  58         private $languagePrefixes = [];
  59
  60         /**
  61          * Creates an SVGReader drawing from the source provided
  62          * @param string $source URI from which to read
  63          * @throws MWException|Exception
  64          */
  65         function __construct( $source ) {
  66                 global $wgSVGMetadataCutoff;
  67                 $this->reader = new XMLReader();
  68
  69                 // Don't use $file->getSize() since file object passed to SVGHandler::getMetadata is bogus.
  70                 $size = filesize( $source );
  71                 if ( $size === false ) {
  72                         throw new MWException( "Error getting filesize of SVG." );
  73                 }
  74
  75                 if ( $size > $wgSVGMetadataCutoff ) {
  76                         $this->debug( "SVG is $size bytes, which is bigger than $wgSVGMetadataCutoff. Truncating." );
  77                         $contents = file_get_contents( $source, false, null, -1, $wgSVGMetadataCutoff );
  78                         if ( $contents === false ) {
  79                                 throw new MWException( 'Error reading SVG file.' );
  80                         }
  81                         $this->reader->XML( $contents, null, LIBXML_NOERROR | LIBXML_NOWARNING );
  82                 } else {
  83                         $this->reader->open( $source, null, LIBXML_NOERROR | LIBXML_NOWARNING );
  84                 }
  85
  86                 // Expand entities, since Adobe Illustrator uses them for xmlns
  87                 // attributes (T33719). Note that libxml2 has some protection
  88                 // against large recursive entity expansions so this is not as
  89                 // insecure as it might appear to be. However, it is still extremely
  90                 // insecure. It's necessary to wrap any read() calls with
  91                 // libxml_disable_entity_loader() to avoid arbitrary local file
  92                 // inclusion, or even arbitrary code execution if the expect
  93                 // extension is installed (T48859).
  94                 $oldDisable = libxml_disable_entity_loader( true );
  95                 $this->reader->setParserProperty( XMLReader::SUBST_ENTITIES, true );
  96
  97                 $this->metadata['width'] = self::DEFAULT_WIDTH;
  98                 $this->metadata['height'] = self::DEFAULT_HEIGHT;
  99
 100                 // The size in the units specified by the SVG file
 101                 // (for the metadata box)
 102                 // Per the SVG spec, if unspecified, default to '100%'
 103                 $this->metadata['originalWidth'] = '100%';
 104                 $this->metadata['originalHeight'] = '100%';
 105
 106                 // Because we cut off the end of the svg making an invalid one. Complicated
 107                 // try catch thing to make sure warnings get restored. Seems like there should
 108                 // be a better way.
 109                 MediaWiki\suppressWarnings();
 110                 try {
 111                         $this->read();
 112                 } catch ( Exception $e ) {
 113                         // Note, if this happens, the width/height will be taken to be 0x0.
 114                         // Should we consider it the default 512x512 instead?
 115                         MediaWiki\restoreWarnings();
 116                         libxml_disable_entity_loader( $oldDisable );
 117                         throw $e;
 118                 }
 119                 MediaWiki\restoreWarnings();
 120                 libxml_disable_entity_loader( $oldDisable );
 121         }
 122
 123         /**
 124          * @return array Array with the known metadata
 125          */
 126         public function getMetadata() {
 127                 return $this->metadata;
 128         }
 129
 130         /**
 131          * Read the SVG
 132          * @throws MWException
 133          * @return bool
 134          */
 135         protected function read() {
 136                 $keepReading = $this->reader->read();
 137
 138                 /* Skip until first element */
 139                 while ( $keepReading && $this->reader->nodeType != XMLReader::ELEMENT ) {
 140                         $keepReading = $this->reader->read();
 141                 }
 142
 143                 if ( $this->reader->localName != 'svg' || $this->reader->namespaceURI != self::NS_SVG ) {
 144                         throw new MWException( "Expected <svg> tag, got " .
 145                                 $this->reader->localName . " in NS " . $this->reader->namespaceURI );
 146                 }
 147                 $this->debug( "<svg> tag is correct." );
 148                 $this->handleSVGAttribs();
 149
 150                 $exitDepth = $this->reader->depth;
 151                 $keepReading = $this->reader->read();
 152                 while ( $keepReading ) {
 153                         $tag = $this->reader->localName;
 154                         $type = $this->reader->nodeType;
 155                         $isSVG = ( $this->reader->namespaceURI == self::NS_SVG );
 156
 157                         $this->debug( "$tag" );
 158
 159                         if ( $isSVG && $tag == 'svg' && $type == XMLReader::END_ELEMENT
 160                                 && $this->reader->depth <= $exitDepth
 161                         ) {
 162                                 break;
 163                         } elseif ( $isSVG && $tag == 'title' ) {
 164                                 $this->readField( $tag, 'title' );
 165                         } elseif ( $isSVG && $tag == 'desc' ) {
 166                                 $this->readField( $tag, 'description' );
 167                         } elseif ( $isSVG && $tag == 'metadata' && $type == XMLReader::ELEMENT ) {
 168                                 $this->readXml( $tag, 'metadata' );
 169                         } elseif ( $isSVG && $tag == 'script' ) {
 170                                 // We normally do not allow scripted svgs.
 171                                 // However its possible to configure MW to let them
 172                                 // in, and such files should be considered animated.
 173                                 $this->metadata['animated'] = true;
 174                         } elseif ( $tag !== '#text' ) {
 175                                 $this->debug( "Unhandled top-level XML tag $tag" );
 176
 177                                 // Recurse into children of current tag, looking for animation and languages.
 178                                 $this->animateFilterAndLang( $tag );
 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                 $this->metadata['translations'] = $this->languages + $this->languagePrefixes;
 188
 189                 return true;
 190         }
 191
 192         /**
 193          * Read a textelement from an element
 194          *
 195          * @param string $name Name of the element that we are reading from
 196          * @param string $metafield Field that we will fill with the result
 197          */
 198         private function readField( $name, $metafield = null ) {
 199                 $this->debug( "Read field $metafield" );
 200                 if ( !$metafield || $this->reader->nodeType != XMLReader::ELEMENT ) {
 201                         return;
 202                 }
 203                 $keepReading = $this->reader->read();
 204                 while ( $keepReading ) {
 205                         if ( $this->reader->localName == $name
 206                                 && $this->reader->namespaceURI == self::NS_SVG
 207                                 && $this->reader->nodeType == XMLReader::END_ELEMENT
 208                         ) {
 209                                 break;
 210                         } elseif ( $this->reader->nodeType == XMLReader::TEXT ) {
 211                                 $this->metadata[$metafield] = trim( $this->reader->value );
 212                         }
 213                         $keepReading = $this->reader->read();
 214                 }
 215         }
 216
 217         /**
 218          * Read an XML snippet from an element
 219          *
 220          * @param string $metafield Field that we will fill with the result
 221          * @throws MWException
 222          */
 223         private function readXml( $metafield = null ) {
 224                 $this->debug( "Read top level metadata" );
 225                 if ( !$metafield || $this->reader->nodeType != XMLReader::ELEMENT ) {
 226                         return;
 227                 }
 228                 // @todo Find and store type of xml snippet. metadata['metadataType'] = "rdf"
 229                 if ( method_exists( $this->reader, 'readInnerXML' ) ) {
 230                         $this->metadata[$metafield] = trim( $this->reader->readInnerXml() );
 231                 } else {
 232                         throw new MWException( "The PHP XMLReader extension does not come " .
 233                                 "with readInnerXML() method. Your libxml is probably out of " .
 234                                 "date (need 2.6.20 or later)." );
 235                 }
 236                 $this->reader->next();
 237         }
 238
 239         /**
 240          * Filter all children, looking for animated elements.
 241          * Also get a list of languages that can be targeted.
 242          *
 243          * @param string $name Name of the element that we are reading from
 244          */
 245         private function animateFilterAndLang( $name ) {
 246                 $this->debug( "animate filter for tag $name" );
 247                 if ( $this->reader->nodeType != XMLReader::ELEMENT ) {
 248                         return;
 249                 }
 250                 if ( $this->reader->isEmptyElement ) {
 251                         return;
 252                 }
 253                 $exitDepth = $this->reader->depth;
 254                 $keepReading = $this->reader->read();
 255                 while ( $keepReading ) {
 256                         if ( $this->reader->localName == $name && $this->reader->depth <= $exitDepth
 257                                 && $this->reader->nodeType == XMLReader::END_ELEMENT
 258                         ) {
 259                                 break;
 260                         } elseif ( $this->reader->namespaceURI == self::NS_SVG
 261                                 && $this->reader->nodeType == XMLReader::ELEMENT
 262                         ) {
 263                                 $sysLang = $this->reader->getAttribute( 'systemLanguage' );
 264                                 if ( !is_null( $sysLang ) && $sysLang !== '' ) {
 265                                         // See https://www.w3.org/TR/SVG/struct.html#SystemLanguageAttribute
 266                                         $langList = explode( ',', $sysLang );
 267                                         foreach ( $langList as $langItem ) {
 268                                                 $langItem = trim( $langItem );
 269                                                 if ( Language::isWellFormedLanguageTag( $langItem ) ) {
 270                                                         $this->languages[$langItem] = self::LANG_FULL_MATCH;
 271                                                 }
 272                                                 // Note, the standard says that any prefix should work,
 273                                                 // here we do only the initial prefix, since that will catch
 274                                                 // 99% of cases, and we are going to compare against fallbacks.
 275                                                 // This differs mildly from how the spec says languages should be
 276                                                 // handled, however it matches better how the MediaWiki language
 277                                                 // preference is generally handled.
 278                                                 $dash = strpos( $langItem, '-' );
 279                                                 // Intentionally checking both !false and > 0 at the same time.
 280                                                 if ( $dash ) {
 281                                                         $itemPrefix = substr( $langItem, 0, $dash );
 282                                                         if ( Language::isWellFormedLanguageTag( $itemPrefix ) ) {
 283                                                                 $this->languagePrefixes[$itemPrefix] = self::LANG_PREFIX_MATCH;
 284                                                         }
 285                                                 }
 286                                         }
 287                                 }
 288                                 switch ( $this->reader->localName ) {
 289                                         case 'script':
 290                                                 // Normally we disallow files with
 291                                                 // <script>, but its possible
 292                                                 // to configure MW to disable
 293                                                 // such checks.
 294                                         case 'animate':
 295                                         case 'set':
 296                                         case 'animateMotion':
 297                                         case 'animateColor':
 298                                         case 'animateTransform':
 299                                                 $this->debug( "HOUSTON WE HAVE ANIMATION" );
 300                                                 $this->metadata['animated'] = true;
 301                                                 break;
 302                                 }
 303                         }
 304                         $keepReading = $this->reader->read();
 305                 }
 306         }
 307
 308         private function debug( $data ) {
 309                 if ( $this->mDebug ) {
 310                         wfDebug( "SVGReader: $data\n" );
 311                 }
 312         }
 313
 314         /**
 315          * Parse the attributes of an SVG element
 316          *
 317          * The parser has to be in the start element of "<svg>"
 318          */
 319         private function handleSVGAttribs() {
 320                 $defaultWidth = self::DEFAULT_WIDTH;
 321                 $defaultHeight = self::DEFAULT_HEIGHT;
 322                 $aspect = 1.0;
 323                 $width = null;
 324                 $height = null;
 325
 326                 if ( $this->reader->getAttribute( 'viewBox' ) ) {
 327                         // min-x min-y width height
 328                         $viewBox = preg_split( '/\s+/', trim( $this->reader->getAttribute( 'viewBox' ) ) );
 329                         if ( count( $viewBox ) == 4 ) {
 330                                 $viewWidth = $this->scaleSVGUnit( $viewBox[2] );
 331                                 $viewHeight = $this->scaleSVGUnit( $viewBox[3] );
 332                                 if ( $viewWidth > 0 && $viewHeight > 0 ) {
 333                                         $aspect = $viewWidth / $viewHeight;
 334                                         $defaultHeight = $defaultWidth / $aspect;
 335                                 }
 336                         }
 337                 }
 338                 if ( $this->reader->getAttribute( 'width' ) ) {
 339                         $width = $this->scaleSVGUnit( $this->reader->getAttribute( 'width' ), $defaultWidth );
 340                         $this->metadata['originalWidth'] = $this->reader->getAttribute( 'width' );
 341                 }
 342                 if ( $this->reader->getAttribute( 'height' ) ) {
 343                         $height = $this->scaleSVGUnit( $this->reader->getAttribute( 'height' ), $defaultHeight );
 344                         $this->metadata['originalHeight'] = $this->reader->getAttribute( 'height' );
 345                 }
 346
 347                 if ( !isset( $width ) && !isset( $height ) ) {
 348                         $width = $defaultWidth;
 349                         $height = $width / $aspect;
 350                 } elseif ( isset( $width ) && !isset( $height ) ) {
 351                         $height = $width / $aspect;
 352                 } elseif ( isset( $height ) && !isset( $width ) ) {
 353                         $width = $height * $aspect;
 354                 }
 355
 356                 if ( $width > 0 && $height > 0 ) {
 357                         $this->metadata['width'] = intval( round( $width ) );
 358                         $this->metadata['height'] = intval( round( $height ) );
 359                 }
 360         }
 361
 362         /**
 363          * Return a rounded pixel equivalent for a labeled CSS/SVG length.
 364          * https://www.w3.org/TR/SVG11/coords.html#Units
 365          *
 366          * @param string $length CSS/SVG length.
 367          * @param float|int $viewportSize Optional scale for percentage units...
 368          * @return float Length in pixels
 369          */
 370         static function scaleSVGUnit( $length, $viewportSize = 512 ) {
 371                 static $unitLength = [
 372                         'px' => 1.0,
 373                         'pt' => 1.25,
 374                         'pc' => 15.0,
 375                         'mm' => 3.543307,
 376                         'cm' => 35.43307,
 377                         'in' => 90.0,
 378                         'em' => 16.0, // fake it?
 379                         'ex' => 12.0, // fake it?
 380                         '' => 1.0, // "User units" pixels by default
 381                 ];
 382                 $matches = [];
 383                 if ( preg_match( '/^\s*(\d+(?:\.\d+)?)(em|ex|px|pt|pc|cm|mm|in|%|)\s*$/', $length, $matches ) ) {
 384                         $length = floatval( $matches[1] );
 385                         $unit = $matches[2];
 386                         if ( $unit == '%' ) {
 387                                 return $length * 0.01 * $viewportSize;
 388                         } else {
 389                                 return $length * $unitLength[$unit];
 390                         }
 391                 } else {
 392                         // Assume pixels
 393                         return floatval( $length );
 394                 }
 395         }
 396 }