3 * Reader for XMP data containing properties relevant to images.
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.
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.
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
25 * Class for reading xmp data containing properties relevant to
26 * images, and spitting out an array that FormatMetadata accepts.
28 * Note, this is not meant to recognize every possible thing you can
29 * encode in XMP. It should recognize all the properties we want.
30 * For example it doesn't have support for structures with multiple
31 * nesting levels, as none of the properties we're supporting use that
32 * feature. If it comes across properties it doesn't recognize, it should
35 * The public methods one would call in this class are
37 * Reads in xmp content.
38 * Can potentially be called multiple times with partial data each time.
39 * - parseExtended( $content )
40 * Reads XMPExtended blocks (jpeg files only).
42 * Outputs a results array.
44 * Note XMP kind of looks like rdf. They are not the same thing - XMP is
45 * encoded as a specific subset of rdf. This class can read XMP. It cannot
50 /** @var array XMP item configuration array */
53 /** @var array Array to hold the current element (and previous element, and so on) */
54 private $curItem = array();
56 /** @var bool|string The structure name when processing nested structures. */
57 private $ancestorStruct = false;
59 /** @var bool|string Temporary holder for character data that appears in xmp doc. */
60 private $charContent = false;
62 /** @var array Stores the state the xmpreader is in (see MODE_FOO constants) */
63 private $mode = array();
65 /** @var array Array to hold results */
66 private $results = array();
68 /** @var bool If we're doing a seq or bag. */
69 private $processingArray = false;
71 /** @var bool|string Used for lang alts only */
72 private $itemLang = false;
74 /** @var resource A resource handle for the XML parser */
77 /** @var bool|string Character set like 'UTF-8' */
78 private $charset = false;
81 private $extendedXMPOffset = 0;
84 * These are various mode constants.
85 * they are used to figure out what to do
86 * with an element when its encountered.
88 * For example, MODE_IGNORE is used when processing
89 * a property we're not interested in. So if a new
90 * element pops up when we're in that mode, we ignore it.
92 const MODE_INITIAL
= 0;
93 const MODE_IGNORE
= 1;
95 const MODE_LI_LANG
= 3;
98 // The following MODE constants are also used in the
99 // $items array to denote what type of property the item is.
100 const MODE_SIMPLE
= 10;
101 const MODE_STRUCT
= 11; // structure (associative array)
102 const MODE_SEQ
= 12; // ordered list
103 const MODE_BAG
= 13; // unordered list
104 const MODE_LANG
= 14;
105 const MODE_ALT
= 15; // non-language alt. Currently not implemented, and not needed atm.
106 const MODE_BAGSTRUCT
= 16; // A BAG of Structs.
108 const NS_RDF
= 'http://www.w3.org/1999/02/22-rdf-syntax-ns#';
109 const NS_XML
= 'http://www.w3.org/XML/1998/namespace';
114 * Primary job is to initialize the XMLParser
116 function __construct() {
118 if ( !function_exists( 'xml_parser_create_ns' ) ) {
119 // this should already be checked by this point
120 throw new MWException( 'XMP support requires XML Parser' );
123 $this->items
= XMPInfo
::getItems();
125 $this->resetXMLParser();
129 * Main use is if a single item has multiple xmp documents describing it.
130 * For example in jpeg's with extendedXMP
132 private function resetXMLParser() {
134 if ( $this->xmlParser
) {
136 xml_parser_free( $this->xmlParser
);
139 $this->xmlParser
= xml_parser_create_ns( 'UTF-8', ' ' );
140 xml_parser_set_option( $this->xmlParser
, XML_OPTION_CASE_FOLDING
, 0 );
141 xml_parser_set_option( $this->xmlParser
, XML_OPTION_SKIP_WHITE
, 1 );
143 xml_set_element_handler( $this->xmlParser
,
144 array( $this, 'startElement' ),
145 array( $this, 'endElement' ) );
147 xml_set_character_data_handler( $this->xmlParser
, array( $this, 'char' ) );
150 /** Destroy the xml parser
152 * Not sure if this is actually needed.
154 function __destruct() {
155 // not sure if this is needed.
156 xml_parser_free( $this->xmlParser
);
159 /** Get the result array. Do some post-processing before returning
160 * the array, and transform any metadata that is special-cased.
162 * @return array Array of results as an array of arrays suitable for
163 * FormatMetadata::getFormattedData().
165 public function getResults() {
166 // xmp-special is for metadata that affects how stuff
167 // is extracted. For example xmpNote:HasExtendedXMP.
169 // It is also used to handle photoshop:AuthorsPosition
170 // which is weird and really part of another property,
171 // see 2:85 in IPTC. See also pg 21 of IPTC4XMP standard.
172 // The location fields also use it.
174 $data = $this->results
;
176 wfRunHooks( 'XMPGetResults', array( &$data ) );
178 if ( isset( $data['xmp-special']['AuthorsPosition'] )
179 && is_string( $data['xmp-special']['AuthorsPosition'] )
180 && isset( $data['xmp-general']['Artist'][0] )
182 // Note, if there is more than one creator,
183 // this only applies to first. This also will
184 // only apply to the dc:Creator prop, not the
187 $data['xmp-general']['Artist'][0] =
188 $data['xmp-special']['AuthorsPosition'] . ', '
189 . $data['xmp-general']['Artist'][0];
192 // Go through the LocationShown and LocationCreated
193 // changing it to the non-hierarchal form used by
194 // the other location fields.
196 if ( isset( $data['xmp-special']['LocationShown'][0] )
197 && is_array( $data['xmp-special']['LocationShown'][0] )
199 // the is_array is just paranoia. It should always
201 foreach ( $data['xmp-special']['LocationShown'] as $loc ) {
202 if ( !is_array( $loc ) ) {
203 // To avoid copying over the _type meta-fields.
206 foreach ( $loc as $field => $val ) {
207 $data['xmp-general'][$field . 'Dest'][] = $val;
211 if ( isset( $data['xmp-special']['LocationCreated'][0] )
212 && is_array( $data['xmp-special']['LocationCreated'][0] )
214 // the is_array is just paranoia. It should always
216 foreach ( $data['xmp-special']['LocationCreated'] as $loc ) {
217 if ( !is_array( $loc ) ) {
218 // To avoid copying over the _type meta-fields.
221 foreach ( $loc as $field => $val ) {
222 $data['xmp-general'][$field . 'Created'][] = $val;
227 // We don't want to return the special values, since they're
228 // special and not info to be stored about the file.
229 unset( $data['xmp-special'] );
231 // Convert GPSAltitude to negative if below sea level.
232 if ( isset( $data['xmp-exif']['GPSAltitudeRef'] )
233 && isset( $data['xmp-exif']['GPSAltitude'] )
236 // Must convert to a real before multiplying by -1
237 // XMPValidate guarantees there will always be a '/' in this value.
238 list( $nom, $denom ) = explode( '/', $data['xmp-exif']['GPSAltitude'] );
239 $data['xmp-exif']['GPSAltitude'] = $nom / $denom;
241 if ( $data['xmp-exif']['GPSAltitudeRef'] == '1' ) {
242 $data['xmp-exif']['GPSAltitude'] *= -1;
244 unset( $data['xmp-exif']['GPSAltitudeRef'] );
251 * Main function to call to parse XMP. Use getResults to
254 * Also catches any errors during processing, writes them to
255 * debug log, blanks result array and returns false.
257 * @param string $content XMP data
258 * @param bool $allOfIt If this is all the data (true) or if its split up (false). Default true
259 * @param bool $reset Does xml parser need to be reset. Default false
260 * @throws MWException
261 * @return bool Success.
263 public function parse( $content, $allOfIt = true, $reset = false ) {
265 $this->resetXMLParser();
269 // detect encoding by looking for BOM which is supposed to be in processing instruction.
270 // see page 12 of http://www.adobe.com/devnet/xmp/pdfs/XMPSpecificationPart3.pdf
271 if ( !$this->charset
) {
273 if ( preg_match( '/\xEF\xBB\xBF|\xFE\xFF|\x00\x00\xFE\xFF|\xFF\xFE\x00\x00|\xFF\xFE/',
278 $this->charset
= 'UTF-16BE';
281 $this->charset
= 'UTF-16LE';
283 case "\x00\x00\xFE\xFF":
284 $this->charset
= 'UTF-32BE';
286 case "\xFF\xFE\x00\x00":
287 $this->charset
= 'UTF-32LE';
290 $this->charset
= 'UTF-8';
293 //this should be impossible to get to
294 throw new MWException( "Invalid BOM" );
297 // standard specifically says, if no bom assume utf-8
298 $this->charset
= 'UTF-8';
301 if ( $this->charset
!== 'UTF-8' ) {
302 //don't convert if already utf-8
303 wfSuppressWarnings();
304 $content = iconv( $this->charset
, 'UTF-8//IGNORE', $content );
308 $ok = xml_parse( $this->xmlParser
, $content, $allOfIt );
310 $error = xml_error_string( xml_get_error_code( $this->xmlParser
) );
311 $where = 'line: ' . xml_get_current_line_number( $this->xmlParser
)
312 . ' column: ' . xml_get_current_column_number( $this->xmlParser
)
313 . ' byte offset: ' . xml_get_current_byte_index( $this->xmlParser
);
315 wfDebugLog( 'XMP', "XMPReader::parse : Error reading XMP content: $error ($where)" );
316 $this->results
= array(); // blank if error.
319 } catch ( MWException
$e ) {
320 wfDebugLog( 'XMP', 'XMP parse error: ' . $e );
321 $this->results
= array();
329 /** Entry point for XMPExtended blocks in jpeg files
331 * @todo In serious need of testing
332 * @see http://www.adobe.ge/devnet/xmp/pdfs/XMPSpecificationPart3.pdf XMP spec part 3 page 20
333 * @param string $content XMPExtended block minus the namespace signature
334 * @return bool If it succeeded.
336 public function parseExtended( $content ) {
337 // @todo FIXME: This is untested. Hard to find example files
338 // or programs that make such files..
339 $guid = substr( $content, 0, 32 );
340 if ( !isset( $this->results
['xmp-special']['HasExtendedXMP'] )
341 ||
$this->results
['xmp-special']['HasExtendedXMP'] !== $guid
343 wfDebugLog( 'XMP', __METHOD__
.
344 " Ignoring XMPExtended block due to wrong guid (guid= '$guid')" );
348 $len = unpack( 'Nlength/Noffset', substr( $content, 32, 8 ) );
350 if ( !$len ||
$len['length'] < 4 ||
$len['offset'] < 0 ||
$len['offset'] > $len['length'] ) {
351 wfDebugLog( 'XMP', __METHOD__
. 'Error reading extended XMP block, invalid length or offset.' );
356 // we're not very robust here. we should accept it in the wrong order.
357 // To quote the XMP standard:
358 // "A JPEG writer should write the ExtendedXMP marker segments in order,
359 // immediately following the StandardXMP. However, the JPEG standard
360 // does not require preservation of marker segment order. A robust JPEG
361 // reader should tolerate the marker segments in any order."
363 // otoh the probability that an image will have more than 128k of
364 // metadata is rather low... so the probability that it will have
365 // > 128k, and be in the wrong order is very low...
367 if ( $len['offset'] !== $this->extendedXMPOffset
) {
368 wfDebugLog( 'XMP', __METHOD__
. 'Ignoring XMPExtended block due to wrong order. (Offset was '
369 . $len['offset'] . ' but expected ' . $this->extendedXMPOffset
. ')' );
374 if ( $len['offset'] === 0 ) {
375 // if we're starting the extended block, we've probably already
376 // done the XMPStandard block, so reset.
377 $this->resetXMLParser();
380 $this->extendedXMPOffset +
= $len['length'];
382 $actualContent = substr( $content, 40 );
384 if ( $this->extendedXMPOffset
=== strlen( $actualContent ) ) {
390 wfDebugLog( 'XMP', __METHOD__
. 'Parsing a XMPExtended block' );
392 return $this->parse( $actualContent, $atEnd );
396 * Character data handler
397 * Called whenever character data is found in the xmp document.
399 * does nothing if we're in MODE_IGNORE or if the data is whitespace
400 * throws an error if we're not in MODE_SIMPLE (as we're not allowed to have character
401 * data in the other modes).
403 * As an example, this happens when we encounter XMP like:
404 * <exif:DigitalZoomRatio>0/10</exif:DigitalZoomRatio>
405 * and are processing the 0/10 bit.
407 * @param XMLParser $parser XMLParser reference to the xml parser
408 * @param string $data Character data
409 * @throws MWException On invalid data
411 function char( $parser, $data ) {
413 $data = trim( $data );
414 if ( trim( $data ) === "" ) {
418 if ( !isset( $this->mode
[0] ) ) {
419 throw new MWException( 'Unexpected character data before first rdf:Description element' );
422 if ( $this->mode
[0] === self
::MODE_IGNORE
) {
426 if ( $this->mode
[0] !== self
::MODE_SIMPLE
427 && $this->mode
[0] !== self
::MODE_QDESC
429 throw new MWException( 'character data where not expected. (mode ' . $this->mode
[0] . ')' );
432 // to check, how does this handle w.s.
433 if ( $this->charContent
=== false ) {
434 $this->charContent
= $data;
436 $this->charContent
.= $data;
440 /** When we hit a closing element in MODE_IGNORE
441 * Check to see if this is the element we started to ignore,
442 * in which case we get out of MODE_IGNORE
444 * @param string $elm Namespace of element followed by a space and then tag name of element.
446 private function endElementModeIgnore( $elm ) {
447 if ( $this->curItem
[0] === $elm ) {
448 array_shift( $this->curItem
);
449 array_shift( $this->mode
);
454 * Hit a closing element when in MODE_SIMPLE.
455 * This generally means that we finished processing a
456 * property value, and now have to save the result to the
459 * For example, when processing:
460 * <exif:DigitalZoomRatio>0/10</exif:DigitalZoomRatio>
461 * this deals with when we hit </exif:DigitalZoomRatio>.
463 * Or it could be if we hit the end element of a property
464 * of a compound data structure (like a member of an array).
466 * @param string $elm Namespace, space, and tag name.
468 private function endElementModeSimple( $elm ) {
469 if ( $this->charContent
!== false ) {
470 if ( $this->processingArray
) {
471 // if we're processing an array, use the original element
472 // name instead of rdf:li.
473 list( $ns, $tag ) = explode( ' ', $this->curItem
[0], 2 );
475 list( $ns, $tag ) = explode( ' ', $elm, 2 );
477 $this->saveValue( $ns, $tag, $this->charContent
);
479 $this->charContent
= false; // reset
481 array_shift( $this->curItem
);
482 array_shift( $this->mode
);
486 * Hit a closing element in MODE_STRUCT, MODE_SEQ, MODE_BAG
487 * generally means we've finished processing a nested structure.
488 * resets some internal variables to indicate that.
490 * Note this means we hit the closing element not the "</rdf:Seq>".
492 * @par For example, when processing:
494 * <exif:ISOSpeedRatings> <rdf:Seq> <rdf:li>64</rdf:li>
495 * </rdf:Seq> </exif:ISOSpeedRatings>
498 * This method is called when we hit the "</exif:ISOSpeedRatings>" tag.
500 * @param string $elm Namespace . space . tag name.
501 * @throws MWException
503 private function endElementNested( $elm ) {
505 /* cur item must be the same as $elm, unless if in MODE_STRUCT
506 in which case it could also be rdf:Description */
507 if ( $this->curItem
[0] !== $elm
508 && !( $elm === self
::NS_RDF
. ' Description'
509 && $this->mode
[0] === self
::MODE_STRUCT
)
511 throw new MWException( "nesting mismatch. got a </$elm> but expected a </" .
512 $this->curItem
[0] . '>' );
515 // Validate structures.
516 list( $ns, $tag ) = explode( ' ', $elm, 2 );
517 if ( isset( $this->items
[$ns][$tag]['validate'] ) ) {
519 $info =& $this->items
[$ns][$tag];
520 $finalName = isset( $info['map_name'] )
521 ?
$info['map_name'] : $tag;
523 $validate = is_array( $info['validate'] ) ?
$info['validate']
524 : array( 'XMPValidate', $info['validate'] );
526 if ( !isset( $this->results
['xmp-' . $info['map_group']][$finalName] ) ) {
527 // This can happen if all the members of the struct failed validation.
528 wfDebugLog( 'XMP', __METHOD__
. " <$ns:$tag> has no valid members." );
529 } elseif ( is_callable( $validate ) ) {
530 $val =& $this->results
['xmp-' . $info['map_group']][$finalName];
531 call_user_func_array( $validate, array( $info, &$val, false ) );
532 if ( is_null( $val ) ) {
533 // the idea being the validation function will unset the variable if
535 wfDebugLog( 'XMP', __METHOD__
. " <$ns:$tag> failed validation." );
536 unset( $this->results
['xmp-' . $info['map_group']][$finalName] );
539 wfDebugLog( 'XMP', __METHOD__
. " Validation function for $finalName ("
540 . $validate[0] . '::' . $validate[1] . '()) is not callable.' );
544 array_shift( $this->curItem
);
545 array_shift( $this->mode
);
546 $this->ancestorStruct
= false;
547 $this->processingArray
= false;
548 $this->itemLang
= false;
552 * Hit a closing element in MODE_LI (either rdf:Seq, or rdf:Bag )
553 * Add information about what type of element this is.
555 * Note we still have to hit the outer "</property>"
557 * @par For example, when processing:
559 * <exif:ISOSpeedRatings> <rdf:Seq> <rdf:li>64</rdf:li>
560 * </rdf:Seq> </exif:ISOSpeedRatings>
563 * This method is called when we hit the "</rdf:Seq>".
564 * (For comparison, we call endElementModeSimple when we
565 * hit the "</rdf:li>")
567 * @param string $elm Namespace . ' ' . element name
568 * @throws MWException
570 private function endElementModeLi( $elm ) {
572 list( $ns, $tag ) = explode( ' ', $this->curItem
[0], 2 );
573 $info = $this->items
[$ns][$tag];
574 $finalName = isset( $info['map_name'] )
575 ?
$info['map_name'] : $tag;
577 array_shift( $this->mode
);
579 if ( !isset( $this->results
['xmp-' . $info['map_group']][$finalName] ) ) {
580 wfDebugLog( 'XMP', __METHOD__
. " Empty compund element $finalName." );
585 if ( $elm === self
::NS_RDF
. ' Seq' ) {
586 $this->results
['xmp-' . $info['map_group']][$finalName]['_type'] = 'ol';
587 } elseif ( $elm === self
::NS_RDF
. ' Bag' ) {
588 $this->results
['xmp-' . $info['map_group']][$finalName]['_type'] = 'ul';
589 } elseif ( $elm === self
::NS_RDF
. ' Alt' ) {
590 // extra if needed as you could theoretically have a non-language alt.
591 if ( $info['mode'] === self
::MODE_LANG
) {
592 $this->results
['xmp-' . $info['map_group']][$finalName]['_type'] = 'lang';
595 throw new MWException( __METHOD__
. " expected </rdf:seq> or </rdf:bag> but instead got $elm." );
600 * End element while in MODE_QDESC
601 * mostly when ending an element when we have a simple value
602 * that has qualifiers.
604 * Qualifiers aren't all that common, and we don't do anything
607 * @param string $elm Namespace and element
609 private function endElementModeQDesc( $elm ) {
611 if ( $elm === self
::NS_RDF
. ' value' ) {
612 list( $ns, $tag ) = explode( ' ', $this->curItem
[0], 2 );
613 $this->saveValue( $ns, $tag, $this->charContent
);
617 array_shift( $this->mode
);
618 array_shift( $this->curItem
);
623 * Handler for hitting a closing element.
625 * generally just calls a helper function depending on what
628 * Ignores the outer wrapping elements that are optional in
629 * xmp and have no meaning.
631 * @param XMLParser $parser
632 * @param string $elm Namespace . ' ' . element name
633 * @throws MWException
635 function endElement( $parser, $elm ) {
636 if ( $elm === ( self
::NS_RDF
. ' RDF' )
637 ||
$elm === 'adobe:ns:meta/ xmpmeta'
638 ||
$elm === 'adobe:ns:meta/ xapmeta'
644 if ( $elm === self
::NS_RDF
. ' type' ) {
645 // these aren't really supported properly yet.
646 // However, it appears they almost never used.
647 wfDebugLog( 'XMP', __METHOD__
. ' encountered <rdf:type>' );
650 if ( strpos( $elm, ' ' ) === false ) {
651 // This probably shouldn't happen.
652 // However, there is a bug in an adobe product
653 // that forgets the namespace on some things.
654 // (Luckily they are unimportant things).
655 wfDebugLog( 'XMP', __METHOD__
. " Encountered </$elm> which has no namespace. Skipping." );
660 if ( count( $this->mode
[0] ) === 0 ) {
661 // This should never ever happen and means
662 // there is a pretty major bug in this class.
663 throw new MWException( 'Encountered end element with no mode' );
666 if ( count( $this->curItem
) == 0 && $this->mode
[0] !== self
::MODE_INITIAL
) {
667 // just to be paranoid. Should always have a curItem, except for initially
668 // (aka during MODE_INITAL).
669 throw new MWException( "Hit end element </$elm> but no curItem" );
672 switch ( $this->mode
[0] ) {
673 case self
::MODE_IGNORE
:
674 $this->endElementModeIgnore( $elm );
676 case self
::MODE_SIMPLE
:
677 $this->endElementModeSimple( $elm );
679 case self
::MODE_STRUCT
:
682 case self
::MODE_LANG
:
683 case self
::MODE_BAGSTRUCT
:
684 $this->endElementNested( $elm );
686 case self
::MODE_INITIAL
:
687 if ( $elm === self
::NS_RDF
. ' Description' ) {
688 array_shift( $this->mode
);
690 throw new MWException( 'Element ended unexpectedly while in MODE_INITIAL' );
694 case self
::MODE_LI_LANG
:
695 $this->endElementModeLi( $elm );
697 case self
::MODE_QDESC
:
698 $this->endElementModeQDesc( $elm );
701 wfDebugLog( 'XMP', __METHOD__
. " no mode (elm = $elm)" );
707 * Hit an opening element while in MODE_IGNORE
709 * XMP is extensible, so ignore any tag we don't understand.
711 * Mostly ignores, unless we encounter the element that we are ignoring.
712 * in which case we add it to the item stack, so we can ignore things
713 * that are nested, correctly.
715 * @param string $elm Namespace . ' ' . tag name
717 private function startElementModeIgnore( $elm ) {
718 if ( $elm === $this->curItem
[0] ) {
719 array_unshift( $this->curItem
, $elm );
720 array_unshift( $this->mode
, self
::MODE_IGNORE
);
725 * Start element in MODE_BAG (unordered array)
726 * this should always be <rdf:Bag>
728 * @param string $elm Namespace . ' ' . tag
729 * @throws MWException If we have an element that's not <rdf:Bag>
731 private function startElementModeBag( $elm ) {
732 if ( $elm === self
::NS_RDF
. ' Bag' ) {
733 array_unshift( $this->mode
, self
::MODE_LI
);
735 throw new MWException( "Expected <rdf:Bag> but got $elm." );
740 * Start element in MODE_SEQ (ordered array)
741 * this should always be <rdf:Seq>
743 * @param string $elm Namespace . ' ' . tag
744 * @throws MWException If we have an element that's not <rdf:Seq>
746 private function startElementModeSeq( $elm ) {
747 if ( $elm === self
::NS_RDF
. ' Seq' ) {
748 array_unshift( $this->mode
, self
::MODE_LI
);
749 } elseif ( $elm === self
::NS_RDF
. ' Bag' ) {
751 wfDebugLog( 'XMP', __METHOD__
. ' Expected an rdf:Seq, but got an rdf:Bag. Pretending'
752 . ' it is a Seq, since some buggy software is known to screw this up.' );
753 array_unshift( $this->mode
, self
::MODE_LI
);
755 throw new MWException( "Expected <rdf:Seq> but got $elm." );
760 * Start element in MODE_LANG (language alternative)
761 * this should always be <rdf:Alt>
763 * This tag tends to be used for metadata like describe this
764 * picture, which can be translated into multiple languages.
766 * XMP supports non-linguistic alternative selections,
767 * which are really only used for thumbnails, which
768 * we don't care about.
770 * @param string $elm Namespace . ' ' . tag
771 * @throws MWException If we have an element that's not <rdf:Alt>
773 private function startElementModeLang( $elm ) {
774 if ( $elm === self
::NS_RDF
. ' Alt' ) {
775 array_unshift( $this->mode
, self
::MODE_LI_LANG
);
777 throw new MWException( "Expected <rdf:Seq> but got $elm." );
782 * Handle an opening element when in MODE_SIMPLE
784 * This should not happen often. This is for if a simple element
785 * already opened has a child element. Could happen for a
789 * <exif:DigitalZoomRatio><rdf:Description><rdf:value>0/10</rdf:value>
790 * <foo:someQualifier>Bar</foo:someQualifier> </rdf:Description>
791 * </exif:DigitalZoomRatio>
793 * This method is called when processing the <rdf:Description> element
795 * @param string $elm Namespace and tag names separated by space.
796 * @param array $attribs Attributes of the element.
797 * @throws MWException
799 private function startElementModeSimple( $elm, $attribs ) {
800 if ( $elm === self
::NS_RDF
. ' Description' ) {
801 // If this value has qualifiers
802 array_unshift( $this->mode
, self
::MODE_QDESC
);
803 array_unshift( $this->curItem
, $this->curItem
[0] );
805 if ( isset( $attribs[self
::NS_RDF
. ' value'] ) ) {
806 list( $ns, $tag ) = explode( ' ', $this->curItem
[0], 2 );
807 $this->saveValue( $ns, $tag, $attribs[self
::NS_RDF
. ' value'] );
809 } elseif ( $elm === self
::NS_RDF
. ' value' ) {
810 // This should not be here.
811 throw new MWException( __METHOD__
. ' Encountered <rdf:value> where it was unexpected.' );
813 // something else we don't recognize, like a qualifier maybe.
814 wfDebugLog( 'XMP', __METHOD__
.
815 " Encountered element <$elm> where only expecting character data as value of " .
817 array_unshift( $this->mode
, self
::MODE_IGNORE
);
818 array_unshift( $this->curItem
, $elm );
823 * Start an element when in MODE_QDESC.
824 * This generally happens when a simple element has an inner
825 * rdf:Description to hold qualifier elements.
828 * <exif:DigitalZoomRatio><rdf:Description><rdf:value>0/10</rdf:value>
829 * <foo:someQualifier>Bar</foo:someQualifier> </rdf:Description>
830 * </exif:DigitalZoomRatio>
831 * Called when processing the <rdf:value> or <foo:someQualifier>.
833 * @param string $elm Namespace and tag name separated by a space.
836 private function startElementModeQDesc( $elm ) {
837 if ( $elm === self
::NS_RDF
. ' value' ) {
838 return; // do nothing
840 // otherwise its a qualifier, which we ignore
841 array_unshift( $this->mode
, self
::MODE_IGNORE
);
842 array_unshift( $this->curItem
, $elm );
847 * Starting an element when in MODE_INITIAL
848 * This usually happens when we hit an element inside
849 * the outer rdf:Description
851 * This is generally where most properties start.
853 * @param string $ns Namespace
854 * @param string $tag Tag name (without namespace prefix)
855 * @param array $attribs Array of attributes
856 * @throws MWException
858 private function startElementModeInitial( $ns, $tag, $attribs ) {
859 if ( $ns !== self
::NS_RDF
) {
861 if ( isset( $this->items
[$ns][$tag] ) ) {
862 if ( isset( $this->items
[$ns][$tag]['structPart'] ) ) {
863 // If this element is supposed to appear only as
864 // a child of a structure, but appears here (not as
865 // a child of a struct), then something weird is
866 // happening, so ignore this element and its children.
868 wfDebugLog( 'XMP', "Encountered <$ns:$tag> outside"
869 . " of its expected parent. Ignoring." );
871 array_unshift( $this->mode
, self
::MODE_IGNORE
);
872 array_unshift( $this->curItem
, $ns . ' ' . $tag );
876 $mode = $this->items
[$ns][$tag]['mode'];
877 array_unshift( $this->mode
, $mode );
878 array_unshift( $this->curItem
, $ns . ' ' . $tag );
879 if ( $mode === self
::MODE_STRUCT
) {
880 $this->ancestorStruct
= isset( $this->items
[$ns][$tag]['map_name'] )
881 ?
$this->items
[$ns][$tag]['map_name'] : $tag;
883 if ( $this->charContent
!== false ) {
885 // Should not happen in valid XMP.
886 throw new MWException( 'tag nested in non-whitespace characters.' );
889 // This element is not on our list of allowed elements so ignore.
890 wfDebugLog( 'XMP', __METHOD__
. " Ignoring unrecognized element <$ns:$tag>." );
891 array_unshift( $this->mode
, self
::MODE_IGNORE
);
892 array_unshift( $this->curItem
, $ns . ' ' . $tag );
897 // process attributes
898 $this->doAttribs( $attribs );
902 * Hit an opening element when in a Struct (MODE_STRUCT)
903 * This is generally for fields of a compound property.
905 * Example of a struct (abbreviated; flash has more properties):
907 * <exif:Flash> <rdf:Description> <exif:Fired>True</exif:Fired>
908 * <exif:Mode>1</exif:Mode></rdf:Description></exif:Flash>
912 * <exif:Flash rdf:parseType='Resource'> <exif:Fired>True</exif:Fired>
913 * <exif:Mode>1</exif:Mode></exif:Flash>
915 * @param string $ns Namespace
916 * @param string $tag Tag name (no ns)
917 * @param array $attribs Array of attribs w/ values.
918 * @throws MWException
920 private function startElementModeStruct( $ns, $tag, $attribs ) {
921 if ( $ns !== self
::NS_RDF
) {
923 if ( isset( $this->items
[$ns][$tag] ) ) {
924 if ( isset( $this->items
[$ns][$this->ancestorStruct
]['children'] )
925 && !isset( $this->items
[$ns][$this->ancestorStruct
]['children'][$tag] )
927 // This assumes that we don't have inter-namespace nesting
928 // which we don't in all the properties we're interested in.
929 throw new MWException( " <$tag> appeared nested in <" . $this->ancestorStruct
930 . "> where it is not allowed." );
932 array_unshift( $this->mode
, $this->items
[$ns][$tag]['mode'] );
933 array_unshift( $this->curItem
, $ns . ' ' . $tag );
934 if ( $this->charContent
!== false ) {
936 // Should not happen in valid XMP.
937 throw new MWException( "tag <$tag> nested in non-whitespace characters (" .
938 $this->charContent
. ")." );
941 array_unshift( $this->mode
, self
::MODE_IGNORE
);
942 array_unshift( $this->curItem
, $elm );
948 if ( $ns === self
::NS_RDF
&& $tag === 'Description' ) {
949 $this->doAttribs( $attribs );
950 array_unshift( $this->mode
, self
::MODE_STRUCT
);
951 array_unshift( $this->curItem
, $this->curItem
[0] );
956 * opening element in MODE_LI
957 * process elements of arrays.
960 * <exif:ISOSpeedRatings> <rdf:Seq> <rdf:li>64</rdf:li>
961 * </rdf:Seq> </exif:ISOSpeedRatings>
962 * This method is called when we hit the <rdf:li> element.
964 * @param string $elm Namespace . ' ' . tagname
965 * @param array $attribs Attributes. (needed for BAGSTRUCTS)
966 * @throws MWException If gets a tag other than <rdf:li>
968 private function startElementModeLi( $elm, $attribs ) {
969 if ( ( $elm ) !== self
::NS_RDF
. ' li' ) {
970 throw new MWException( "<rdf:li> expected but got $elm." );
973 if ( !isset( $this->mode
[1] ) ) {
974 // This should never ever ever happen. Checking for it
976 throw new MWException( 'In mode Li, but no 2xPrevious mode!' );
979 if ( $this->mode
[1] === self
::MODE_BAGSTRUCT
) {
980 // This list item contains a compound (STRUCT) value.
981 array_unshift( $this->mode
, self
::MODE_STRUCT
);
982 array_unshift( $this->curItem
, $elm );
983 $this->processingArray
= true;
985 if ( !isset( $this->curItem
[1] ) ) {
987 throw new MWException( 'Can not find parent of BAGSTRUCT.' );
989 list( $curNS, $curTag ) = explode( ' ', $this->curItem
[1] );
990 $this->ancestorStruct
= isset( $this->items
[$curNS][$curTag]['map_name'] )
991 ?
$this->items
[$curNS][$curTag]['map_name'] : $curTag;
993 $this->doAttribs( $attribs );
995 // Normal BAG or SEQ containing simple values.
996 array_unshift( $this->mode
, self
::MODE_SIMPLE
);
997 // need to add curItem[0] on again since one is for the specific item
998 // and one is for the entire group.
999 array_unshift( $this->curItem
, $this->curItem
[0] );
1000 $this->processingArray
= true;
1005 * Opening element in MODE_LI_LANG.
1006 * process elements of language alternatives
1009 * <dc:title> <rdf:Alt> <rdf:li xml:lang="x-default">My house
1010 * </rdf:li> </rdf:Alt> </dc:title>
1012 * This method is called when we hit the <rdf:li> element.
1014 * @param string $elm Namespace . ' ' . tag
1015 * @param array $attribs Array of elements (most importantly xml:lang)
1016 * @throws MWException If gets a tag other than <rdf:li> or if no xml:lang
1018 private function startElementModeLiLang( $elm, $attribs ) {
1019 if ( $elm !== self
::NS_RDF
. ' li' ) {
1020 throw new MWException( __METHOD__
. " <rdf:li> expected but got $elm." );
1022 if ( !isset( $attribs[self
::NS_XML
. ' lang'] )
1023 ||
!preg_match( '/^[-A-Za-z0-9]{2,}$/D', $attribs[self
::NS_XML
. ' lang'] )
1025 throw new MWException( __METHOD__
1026 . " <rdf:li> did not contain, or has invalid xml:lang attribute in lang alternative" );
1029 // Lang is case-insensitive.
1030 $this->itemLang
= strtolower( $attribs[self
::NS_XML
. ' lang'] );
1032 // need to add curItem[0] on again since one is for the specific item
1033 // and one is for the entire group.
1034 array_unshift( $this->curItem
, $this->curItem
[0] );
1035 array_unshift( $this->mode
, self
::MODE_SIMPLE
);
1036 $this->processingArray
= true;
1040 * Hits an opening element.
1041 * Generally just calls a helper based on what MODE we're in.
1042 * Also does some initial set up for the wrapper element
1044 * @param XMLParser $parser
1045 * @param string $elm Namespace "<space>" element
1046 * @param array $attribs Attribute name => value
1047 * @throws MWException
1049 function startElement( $parser, $elm, $attribs ) {
1051 if ( $elm === self
::NS_RDF
. ' RDF'
1052 ||
$elm === 'adobe:ns:meta/ xmpmeta'
1053 ||
$elm === 'adobe:ns:meta/ xapmeta'
1057 } elseif ( $elm === self
::NS_RDF
. ' Description' ) {
1058 if ( count( $this->mode
) === 0 ) {
1060 array_unshift( $this->mode
, self
::MODE_INITIAL
);
1062 } elseif ( $elm === self
::NS_RDF
. ' type' ) {
1063 // This doesn't support rdf:type properly.
1064 // In practise I have yet to see a file that
1065 // uses this element, however it is mentioned
1066 // on page 25 of part 1 of the xmp standard.
1068 // also it seems as if exiv2 and exiftool do not support
1069 // this either (That or I misunderstand the standard)
1070 wfDebugLog( 'XMP', __METHOD__
. ' Encountered <rdf:type> which isn\'t currently supported' );
1073 if ( strpos( $elm, ' ' ) === false ) {
1074 // This probably shouldn't happen.
1075 wfDebugLog( 'XMP', __METHOD__
. " Encountered <$elm> which has no namespace. Skipping." );
1080 list( $ns, $tag ) = explode( ' ', $elm, 2 );
1082 if ( count( $this->mode
) === 0 ) {
1083 // This should not happen.
1084 throw new MWException( 'Error extracting XMP, '
1085 . "encountered <$elm> with no mode" );
1088 switch ( $this->mode
[0] ) {
1089 case self
::MODE_IGNORE
:
1090 $this->startElementModeIgnore( $elm );
1092 case self
::MODE_SIMPLE
:
1093 $this->startElementModeSimple( $elm, $attribs );
1095 case self
::MODE_INITIAL
:
1096 $this->startElementModeInitial( $ns, $tag, $attribs );
1098 case self
::MODE_STRUCT
:
1099 $this->startElementModeStruct( $ns, $tag, $attribs );
1101 case self
::MODE_BAG
:
1102 case self
::MODE_BAGSTRUCT
:
1103 $this->startElementModeBag( $elm );
1105 case self
::MODE_SEQ
:
1106 $this->startElementModeSeq( $elm );
1108 case self
::MODE_LANG
:
1109 $this->startElementModeLang( $elm );
1111 case self
::MODE_LI_LANG
:
1112 $this->startElementModeLiLang( $elm, $attribs );
1115 $this->startElementModeLi( $elm, $attribs );
1117 case self
::MODE_QDESC
:
1118 $this->startElementModeQDesc( $elm );
1121 throw new MWException( 'StartElement in unknown mode: ' . $this->mode
[0] );
1126 * Process attributes.
1127 * Simple values can be stored as either a tag or attribute
1129 * Often the initial "<rdf:Description>" tag just has all the simple
1130 * properties as attributes.
1132 * @codingStandardsIgnoreStart Long line that cannot be broken
1135 * <rdf:Description rdf:about="" xmlns:exif="http://ns.adobe.com/exif/1.0/" exif:DigitalZoomRatio="0/10">
1137 * @codingStandardsIgnoreEnd
1139 * @param array $attribs Array attribute=>value
1140 * @throws MWException
1142 private function doAttribs( $attribs ) {
1143 // first check for rdf:parseType attribute, as that can change
1144 // how the attributes are interperted.
1146 if ( isset( $attribs[self
::NS_RDF
. ' parseType'] )
1147 && $attribs[self
::NS_RDF
. ' parseType'] === 'Resource'
1148 && $this->mode
[0] === self
::MODE_SIMPLE
1150 // this is equivalent to having an inner rdf:Description
1151 $this->mode
[0] = self
::MODE_QDESC
;
1153 foreach ( $attribs as $name => $val ) {
1154 if ( strpos( $name, ' ' ) === false ) {
1155 // This shouldn't happen, but so far some old software forgets namespace
1157 wfDebugLog( 'XMP', __METHOD__
. ' Encountered non-namespaced attribute: '
1158 . " $name=\"$val\". Skipping. " );
1161 list( $ns, $tag ) = explode( ' ', $name, 2 );
1162 if ( $ns === self
::NS_RDF
) {
1163 if ( $tag === 'value' ||
$tag === 'resource' ) {
1164 // resource is for url.
1165 // value attribute is a weird way of just putting the contents.
1166 $this->char( $this->xmlParser
, $val );
1168 } elseif ( isset( $this->items
[$ns][$tag] ) ) {
1169 if ( $this->mode
[0] === self
::MODE_SIMPLE
) {
1170 throw new MWException( __METHOD__
1171 . " $ns:$tag found as attribute where not allowed" );
1173 $this->saveValue( $ns, $tag, $val );
1175 wfDebugLog( 'XMP', __METHOD__
. " Ignoring unrecognized element <$ns:$tag>." );
1181 * Given an extracted value, save it to results array
1183 * note also uses $this->ancestorStruct and
1184 * $this->processingArray to determine what name to
1185 * save the value under. (in addition to $tag).
1187 * @param string $ns Namespace of tag this is for
1188 * @param string $tag Tag name
1189 * @param string $val Value to save
1191 private function saveValue( $ns, $tag, $val ) {
1193 $info =& $this->items
[$ns][$tag];
1194 $finalName = isset( $info['map_name'] )
1195 ?
$info['map_name'] : $tag;
1196 if ( isset( $info['validate'] ) ) {
1197 $validate = is_array( $info['validate'] ) ?
$info['validate']
1198 : array( 'XMPValidate', $info['validate'] );
1200 if ( is_callable( $validate ) ) {
1201 call_user_func_array( $validate, array( $info, &$val, true ) );
1202 // the reasoning behind using &$val instead of using the return value
1203 // is to be consistent between here and validating structures.
1204 if ( is_null( $val ) ) {
1205 wfDebugLog( 'XMP', __METHOD__
. " <$ns:$tag> failed validation." );
1210 wfDebugLog( 'XMP', __METHOD__
. " Validation function for $finalName ("
1211 . $validate[0] . '::' . $validate[1] . '()) is not callable.' );
1215 if ( $this->ancestorStruct
&& $this->processingArray
) {
1216 // Aka both an array and a struct. ( self::MODE_BAGSTRUCT )
1217 $this->results
['xmp-' . $info['map_group']][$this->ancestorStruct
][][$finalName] = $val;
1218 } elseif ( $this->ancestorStruct
) {
1219 $this->results
['xmp-' . $info['map_group']][$this->ancestorStruct
][$finalName] = $val;
1220 } elseif ( $this->processingArray
) {
1221 if ( $this->itemLang
=== false ) {
1223 $this->results
['xmp-' . $info['map_group']][$finalName][] = $val;
1226 $this->results
['xmp-' . $info['map_group']][$finalName][$this->itemLang
] = $val;
1229 $this->results
['xmp-' . $info['map_group']][$finalName] = $val;