3 * Formatting of image metadata values into human readable form.
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
21 * @author Ævar Arnfjörð Bjarmason <avarab@gmail.com>
22 * @copyright Copyright © 2005, Ævar Arnfjörð Bjarmason, 2009 Brent Garber, 2010 Brian Wolff
23 * @license GPL-2.0-or-later
24 * @see http://exif.org/Exif2-2.PDF The Exif 2.2 specification
27 use MediaWiki\MediaWikiServices
;
28 use Wikimedia\Timestamp\TimestampException
;
31 * Format Image metadata values into a human readable form.
33 * Note lots of these messages use the prefix 'exif' even though
34 * they may not be exif properties. For example 'exif-ImageDescription'
35 * can be the Exif ImageDescription, or it could be the iptc-iim caption
36 * property, or it could be the xmp dc:description property. This
37 * is because these messages should be independent of how the data is
38 * stored, sine the user doesn't care if the description is stored in xmp,
39 * exif, etc only that its a description. (Additionally many of these properties
40 * are merged together following the MWG standard, such that for example,
41 * exif properties override XMP properties that mean the same thing if
42 * there is a conflict).
44 * It should perhaps use a prefix like 'metadata' instead, but there
45 * is already a large number of messages using the 'exif' prefix.
48 * @since 1.23 the class extends ContextSource and various formerly-public
49 * internal methods are private
51 class FormatMetadata
extends ContextSource
{
53 * Only output a single language for multi-language fields
57 protected $singleLang = false;
60 * Trigger only outputting single language for multilanguage fields
65 public function setSingleLanguage( $val ) {
66 $this->singleLang
= $val;
70 * Numbers given by Exif user agents are often magical, that is they
71 * should be replaced by a detailed explanation depending on their
72 * value which most of the time are plain integers. This function
73 * formats Exif (and other metadata) values into human readable form.
75 * This is the usual entry point for this class.
77 * @param array $tags The Exif data to format ( as returned by
78 * Exif::getFilteredData() or BitmapMetadataHandler )
79 * @param bool|IContextSource $context Context to use (optional)
82 public static function getFormattedData( $tags, $context = false ) {
83 $obj = new FormatMetadata
;
85 $obj->setContext( $context );
88 return $obj->makeFormattedData( $tags );
92 * Numbers given by Exif user agents are often magical, that is they
93 * should be replaced by a detailed explanation depending on their
94 * value which most of the time are plain integers. This function
95 * formats Exif (and other metadata) values into human readable form.
97 * @param array $tags The Exif data to format ( as returned by
98 * Exif::getFilteredData() or BitmapMetadataHandler )
102 public function makeFormattedData( $tags ) {
103 $resolutionunit = !isset( $tags['ResolutionUnit'] ) ||
$tags['ResolutionUnit'] == 2 ?
2 : 3;
104 unset( $tags['ResolutionUnit'] );
106 foreach ( $tags as $tag => &$vals ) {
107 // This seems ugly to wrap non-array's in an array just to unwrap again,
108 // especially when most of the time it is not an array
109 if ( !is_array( $tags[$tag] ) ) {
113 // _type is a special value to say what array type
114 if ( isset( $tags[$tag]['_type'] ) ) {
115 $type = $tags[$tag]['_type'];
116 unset( $vals['_type'] );
118 $type = 'ul'; // default unordered list.
121 // This is done differently as the tag is an array.
122 if ( $tag == 'GPSTimeStamp' && count( $vals ) === 3 ) {
123 // hour min sec array
125 $h = explode( '/', $vals[0] );
126 $m = explode( '/', $vals[1] );
127 $s = explode( '/', $vals[2] );
129 // this should already be validated
130 // when loaded from file, but it could
131 // come from a foreign repo, so be
142 $tags[$tag] = str_pad( intval( $h[0] / $h[1] ), 2, '0', STR_PAD_LEFT
)
143 . ':' . str_pad( intval( $m[0] / $m[1] ), 2, '0', STR_PAD_LEFT
)
144 . ':' . str_pad( intval( $s[0] / $s[1] ), 2, '0', STR_PAD_LEFT
);
147 $time = wfTimestamp( TS_MW
, '1971:01:01 ' . $tags[$tag] );
148 // the 1971:01:01 is just a placeholder, and not shown to user.
149 if ( $time && intval( $time ) > 0 ) {
150 $tags[$tag] = $this->getLanguage()->time( $time );
152 } catch ( TimestampException
$e ) {
153 // This shouldn't happen, but we've seen bad formats
154 // such as 4-digit seconds in the wild.
155 // leave $tags[$tag] as-is
160 // The contact info is a multi-valued field
161 // instead of the other props which are single
162 // valued (mostly) so handle as a special case.
163 if ( $tag === 'Contact' ) {
164 $vals = $this->collapseContactInfo( $vals );
168 foreach ( $vals as &$val ) {
183 $val = $this->exifMsg( $tag, $val );
186 /* If not recognized, display as is. */
191 case 'PhotometricInterpretation':
205 $val = $this->exifMsg( $tag, $val );
208 /* If not recognized, display as is. */
223 $val = $this->exifMsg( $tag, $val );
226 /* If not recognized, display as is. */
231 case 'PlanarConfiguration':
235 $val = $this->exifMsg( $tag, $val );
238 /* If not recognized, display as is. */
243 // TODO: YCbCrSubSampling
244 case 'YCbCrPositioning':
248 $val = $this->exifMsg( $tag, $val );
251 /* If not recognized, display as is. */
258 switch ( $resolutionunit ) {
260 $val = $this->exifMsg( 'XYResolution', 'i', $this->formatNum( $val ) );
263 $val = $this->exifMsg( 'XYResolution', 'c', $this->formatNum( $val ) );
266 /* If not recognized, display as is. */
271 // TODO: YCbCrCoefficients #p27 (see annex E)
273 case 'FlashpixVersion':
274 $val = (int)$val / 100;
281 $val = $this->exifMsg( $tag, $val );
284 /* If not recognized, display as is. */
289 case 'ComponentsConfiguration':
298 $val = $this->exifMsg( $tag, $val );
301 /* If not recognized, display as is. */
307 case 'DateTimeOriginal':
308 case 'DateTimeDigitized':
309 case 'DateTimeReleased':
310 case 'DateTimeExpires':
313 case 'DateTimeMetadata':
314 if ( $val == '0000:00:00 00:00:00' ||
$val == ' : : : : ' ) {
315 $val = $this->msg( 'exif-unknowndate' )->text();
316 } elseif ( preg_match(
317 '/^(?:\d{4}):(?:\d\d):(?:\d\d) (?:\d\d):(?:\d\d):(?:\d\d)$/D',
321 $time = wfTimestamp( TS_MW
, $val );
322 if ( $time && intval( $time ) > 0 ) {
323 $val = $this->getLanguage()->timeanddate( $time );
325 } elseif ( preg_match( '/^(?:\d{4}):(?:\d\d):(?:\d\d) (?:\d\d):(?:\d\d)$/D', $val ) ) {
326 // No second field. Still format the same
327 // since timeanddate doesn't include seconds anyways,
328 // but second still available in api
329 $time = wfTimestamp( TS_MW
, $val . ':00' );
330 if ( $time && intval( $time ) > 0 ) {
331 $val = $this->getLanguage()->timeanddate( $time );
333 } elseif ( preg_match( '/^(?:\d{4}):(?:\d\d):(?:\d\d)$/D', $val ) ) {
334 // If only the date but not the time is filled in.
335 $time = wfTimestamp( TS_MW
, substr( $val, 0, 4 )
336 . substr( $val, 5, 2 )
337 . substr( $val, 8, 2 )
339 if ( $time && intval( $time ) > 0 ) {
340 $val = $this->getLanguage()->date( $time );
343 // else it will just output $val without formatting it.
346 case 'ExposureProgram':
357 $val = $this->exifMsg( $tag, $val );
360 /* If not recognized, display as is. */
365 case 'SubjectDistance':
366 $val = $this->exifMsg( $tag, '', $this->formatNum( $val ) );
380 $val = $this->exifMsg( $tag, $val );
383 /* If not recognized, display as is. */
411 $val = $this->exifMsg( $tag, $val );
414 /* If not recognized, display as is. */
421 'fired' => $val & 0b00000001
,
422 'return' => ( $val & 0b00000110
) >> 1,
423 'mode' => ( $val & 0b00011000
) >> 3,
424 'function' => ( $val & 0b00100000
) >> 5,
425 'redeye' => ( $val & 0b01000000
) >> 6,
426 // 'reserved' => ( $val & 0b10000000 ) >> 7,
429 # We do not need to handle unknown values since all are used.
430 foreach ( $flashDecode as $subTag => $subValue ) {
431 # We do not need any message for zeroed values.
432 if ( $subTag != 'fired' && $subValue == 0 ) {
435 $fullTag = $tag . '-' . $subTag;
436 $flashMsgs[] = $this->exifMsg( $fullTag, $subValue );
438 $val = $this->getLanguage()->commaList( $flashMsgs );
441 case 'FocalPlaneResolutionUnit':
444 $val = $this->exifMsg( $tag, $val );
447 /* If not recognized, display as is. */
452 case 'SensingMethod':
461 $val = $this->exifMsg( $tag, $val );
464 /* If not recognized, display as is. */
472 $val = $this->exifMsg( $tag, $val );
475 /* If not recognized, display as is. */
483 $val = $this->exifMsg( $tag, $val );
486 /* If not recognized, display as is. */
491 case 'CustomRendered':
495 /* The following are unofficial Apple additions */
496 case 2: /* HDR (no original saved) */
497 case 3: /* HDR (original saved) */
498 case 4: /* Original (for HDR) */
499 /* Yes 5 is not present ;) */
500 case 6: /* Panorama */
501 case 7: /* Portrait HDR */
502 case 8: /* Portrait */
503 $val = $this->exifMsg( $tag, $val );
506 /* If not recognized, display as is. */
516 $val = $this->exifMsg( $tag, $val );
519 /* If not recognized, display as is. */
528 $val = $this->exifMsg( $tag, $val );
531 /* If not recognized, display as is. */
536 case 'SceneCaptureType':
542 $val = $this->exifMsg( $tag, $val );
545 /* If not recognized, display as is. */
557 $val = $this->exifMsg( $tag, $val );
560 /* If not recognized, display as is. */
570 $val = $this->exifMsg( $tag, $val );
573 /* If not recognized, display as is. */
583 $val = $this->exifMsg( $tag, $val );
586 /* If not recognized, display as is. */
596 $val = $this->exifMsg( $tag, $val );
599 /* If not recognized, display as is. */
604 case 'SubjectDistanceRange':
610 $val = $this->exifMsg( $tag, $val );
613 /* If not recognized, display as is. */
618 // The GPS...Ref values are kept for compatibility, probably won't be reached.
619 case 'GPSLatitudeRef':
620 case 'GPSDestLatitudeRef':
624 $val = $this->exifMsg( 'GPSLatitude', $val );
627 /* If not recognized, display as is. */
632 case 'GPSLongitudeRef':
633 case 'GPSDestLongitudeRef':
637 $val = $this->exifMsg( 'GPSLongitude', $val );
640 /* If not recognized, display as is. */
647 $val = $this->exifMsg( 'GPSAltitude', 'below-sealevel', $this->formatNum( -$val, 3 ) );
649 $val = $this->exifMsg( 'GPSAltitude', 'above-sealevel', $this->formatNum( $val, 3 ) );
657 $val = $this->exifMsg( $tag, $val );
660 /* If not recognized, display as is. */
665 case 'GPSMeasureMode':
669 $val = $this->exifMsg( $tag, $val );
672 /* If not recognized, display as is. */
678 case 'GPSImgDirectionRef':
679 case 'GPSDestBearingRef':
683 $val = $this->exifMsg( 'GPSDirection', $val );
686 /* If not recognized, display as is. */
692 case 'GPSDestLatitude':
693 $val = $this->formatCoords( $val, 'latitude' );
696 case 'GPSDestLongitude':
697 $val = $this->formatCoords( $val, 'longitude' );
705 $val = $this->exifMsg( 'GPSSpeed', $val );
708 /* If not recognized, display as is. */
713 case 'GPSDestDistanceRef':
718 $val = $this->exifMsg( 'GPSDestDistance', $val );
721 /* If not recognized, display as is. */
727 // See https://en.wikipedia.org/wiki/Dilution_of_precision_(GPS)
729 $val = $this->exifMsg( $tag, 'excellent', $this->formatNum( $val ) );
730 } elseif ( $val <= 5 ) {
731 $val = $this->exifMsg( $tag, 'good', $this->formatNum( $val ) );
732 } elseif ( $val <= 10 ) {
733 $val = $this->exifMsg( $tag, 'moderate', $this->formatNum( $val ) );
734 } elseif ( $val <= 20 ) {
735 $val = $this->exifMsg( $tag, 'fair', $this->formatNum( $val ) );
737 $val = $this->exifMsg( $tag, 'poor', $this->formatNum( $val ) );
741 // This is not in the Exif standard, just a special
742 // case for our purposes which enables wikis to wikify
743 // the make, model and software name to link to their articles.
746 $val = $this->exifMsg( $tag, '', $val );
750 if ( is_array( $val ) ) {
751 if ( count( $val ) > 1 ) {
752 // if its a software, version array.
753 $val = $this->msg( 'exif-software-version-value', $val[0], $val[1] )->text();
755 // https://phabricator.wikimedia.org/T178130
756 $val = $this->exifMsg( $tag, '', $val[0] );
759 $val = $this->exifMsg( $tag, '', $val );
764 // Show the pretty fraction as well as decimal version
765 $val = $this->msg( 'exif-exposuretime-format',
766 $this->formatFraction( $val ), $this->formatNum( $val ) )->text();
768 case 'ISOSpeedRatings':
769 // If its = 65535 that means its at the
770 // limit of the size of Exif::short and
772 if ( $val == '65535' ) {
773 $val = $this->exifMsg( $tag, 'overflow' );
775 $val = $this->formatNum( $val );
779 $val = $this->msg( 'exif-fnumber-format',
780 $this->formatNum( $val ) )->text();
784 case 'FocalLengthIn35mmFilm':
785 $val = $this->msg( 'exif-focallength-format',
786 $this->formatNum( $val ) )->text();
789 case 'MaxApertureValue':
790 if ( strpos( $val, '/' ) !== false ) {
791 // need to expand this earlier to calculate fNumber
792 list( $n, $d ) = explode( '/', $val );
793 if ( is_numeric( $n ) && is_numeric( $d ) ) {
797 if ( is_numeric( $val ) ) {
798 $fNumber = 2 ** ( $val / 2 );
799 if ( $fNumber !== false ) {
800 $val = $this->msg( 'exif-maxaperturevalue-value',
801 $this->formatNum( $val ),
802 $this->formatNum( $fNumber, 2 )
809 switch ( strtolower( $val ) ) {
810 // See pg 29 of IPTC photo
811 // metadata standard.
829 $val = $this->exifMsg(
835 case 'SubjectNewsCode':
836 // Essentially like iimCategory.
837 // 8 (numeric) digit hierarchical
838 // classification. We decode the
839 // first 2 digits, which provide
841 $val = $this->convertNewsCode( $val );
844 // 1-8 with 1 being highest, 5 normal
845 // 0 is reserved, and 9 is 'user-defined'.
847 if ( $val == 0 ||
$val == 9 ) {
849 } elseif ( $val < 5 && $val > 1 ) {
851 } elseif ( $val == 5 ) {
853 } elseif ( $val <= 8 && $val > 5 ) {
857 if ( $urgency !== '' ) {
858 $val = $this->exifMsg( 'urgency',
864 // Things that have a unit of pixels.
865 case 'OriginalImageHeight':
866 case 'OriginalImageWidth':
867 case 'PixelXDimension':
868 case 'PixelYDimension':
871 $val = $this->formatNum( $val ) . ' ' . $this->msg( 'unit-pixel' )->text();
874 // Do not transform fields with pure text.
875 // For some languages the formatNum()
876 // conversion results to wrong output like
877 // foo,bar@example,com or foo٫bar@example٫com.
878 // Also some 'numeric' things like Scene codes
879 // are included here as we really don't want
881 case 'ImageDescription':
885 case 'RelatedSoundFile':
886 case 'ImageUniqueID':
887 case 'SpectralSensitivity':
888 case 'GPSSatellites':
892 case 'WorldRegionDest':
894 case 'CountryCodeDest':
895 case 'ProvinceOrStateDest':
897 case 'SublocationDest':
898 case 'WorldRegionCreated':
899 case 'CountryCreated':
900 case 'CountryCodeCreated':
901 case 'ProvinceOrStateCreated':
903 case 'SublocationCreated':
905 case 'SpecialInstructions':
910 case 'FixtureIdentifier':
912 case 'LocationDestCode':
914 case 'JPEGFileComment':
915 case 'iimSupplementalCategory':
916 case 'OriginalTransmissionRef':
918 case 'dc-contributor':
927 case 'CameraOwnerName':
930 case 'RightsCertificate':
931 case 'CopyrightOwner':
934 case 'OriginalDocumentID':
936 case 'MorePermissionsUrl':
937 case 'AttributionUrl':
938 case 'PreferredAttributionName':
939 case 'PNGFileComment':
941 case 'ContentWarning':
942 case 'GIFFileComment':
944 case 'IntellectualGenre':
946 case 'OrginisationInImage':
947 case 'PersonInImage':
949 $val = htmlspecialchars( $val );
957 $val = $this->exifMsg( $tag, $val );
960 $val = htmlspecialchars( $val );
968 $val = $this->exifMsg( $tag, $val );
973 if ( $val == '-1' ) {
974 $val = $this->exifMsg( $tag, 'rejected' );
976 $val = $this->formatNum( $val );
981 $lang = Language
::fetchLanguageName( strtolower( $val ), $this->getLanguage()->getCode() );
982 $val = htmlspecialchars( $lang ?
: $val );
986 $val = $this->formatNum( $val );
990 // End formatting values, start flattening arrays.
991 $vals = $this->flattenArrayReal( $vals, $type );
998 * Flatten an array, using the content language for any messages.
1000 * @param array $vals Array of values
1001 * @param string $type Type of array (either lang, ul, ol).
1002 * lang = language assoc array with keys being the lang code
1003 * ul = unordered list, ol = ordered list
1004 * type can also come from the '_type' member of $vals.
1005 * @param bool $noHtml If to avoid returning anything resembling HTML.
1006 * (Ugly hack for backwards compatibility with old MediaWiki).
1007 * @param bool|IContextSource $context
1008 * @return string Single value (in wiki-syntax).
1011 public static function flattenArrayContentLang( $vals, $type = 'ul',
1012 $noHtml = false, $context = false
1014 $obj = new FormatMetadata
;
1016 $obj->setContext( $context );
1018 $context = new DerivativeContext( $obj->getContext() );
1019 $context->setLanguage( MediaWikiServices
::getInstance()->getContentLanguage() );
1020 $obj->setContext( $context );
1022 return $obj->flattenArrayReal( $vals, $type, $noHtml );
1026 * A function to collapse multivalued tags into a single value.
1027 * This turns an array of (for example) authors into a bulleted list.
1029 * This is public on the basis it might be useful outside of this class.
1031 * @param array $vals Array of values
1032 * @param string $type Type of array (either lang, ul, ol).
1033 * lang = language assoc array with keys being the lang code
1034 * ul = unordered list, ol = ordered list
1035 * type can also come from the '_type' member of $vals.
1036 * @param bool $noHtml If to avoid returning anything resembling HTML.
1037 * (Ugly hack for backwards compatibility with old mediawiki).
1038 * @return string Single value (in wiki-syntax).
1041 public function flattenArrayReal( $vals, $type = 'ul', $noHtml = false ) {
1042 if ( !is_array( $vals ) ) {
1043 return $vals; // do nothing if not an array;
1046 if ( isset( $vals['_type'] ) ) {
1047 $type = $vals['_type'];
1048 unset( $vals['_type'] );
1051 if ( !is_array( $vals ) ) {
1052 return $vals; // do nothing if not an array;
1053 } elseif ( count( $vals ) === 1 && $type !== 'lang' && isset( $vals[0] ) ) {
1055 } elseif ( count( $vals ) === 0 ) {
1056 wfDebug( __METHOD__
. " metadata array with 0 elements!\n" );
1058 return ""; // paranoia. This should never happen
1060 /* @todo FIXME: This should hide some of the list entries if there are
1061 * say more than four. Especially if a field is translated into 20
1062 * languages, we don't want to show them all by default
1066 // Display default, followed by ContentLanguage,
1067 // followed by the rest in no particular
1070 // Todo: hide some items if really long list.
1074 $priorityLanguages = $this->getPriorityLanguages();
1075 $defaultItem = false;
1076 $defaultLang = false;
1078 // If default is set, save it for later,
1079 // as we don't know if it's equal to
1080 // one of the lang codes. (In xmp
1081 // you specify the language for a
1082 // default property by having both
1083 // a default prop, and one in the language
1084 // that are identical)
1085 if ( isset( $vals['x-default'] ) ) {
1086 $defaultItem = $vals['x-default'];
1087 unset( $vals['x-default'] );
1089 foreach ( $priorityLanguages as $pLang ) {
1090 if ( isset( $vals[$pLang] ) ) {
1092 if ( $vals[$pLang] === $defaultItem ) {
1093 $defaultItem = false;
1096 $content .= $this->langItem(
1097 $vals[$pLang], $pLang,
1098 $isDefault, $noHtml );
1100 unset( $vals[$pLang] );
1102 if ( $this->singleLang
) {
1103 return Html
::rawElement( 'span',
1104 [ 'lang' => $pLang ], $vals[$pLang] );
1110 foreach ( $vals as $lang => $item ) {
1111 if ( $item === $defaultItem ) {
1112 $defaultLang = $lang;
1115 $content .= $this->langItem( $item,
1116 $lang, false, $noHtml );
1117 if ( $this->singleLang
) {
1118 return Html
::rawElement( 'span',
1119 [ 'lang' => $lang ], $item );
1122 if ( $defaultItem !== false ) {
1123 $content = $this->langItem( $defaultItem,
1124 $defaultLang, true, $noHtml ) .
1126 if ( $this->singleLang
) {
1127 return $defaultItem;
1134 return '<ul class="metadata-langlist">' .
1139 return "\n#" . implode( "\n#", $vals );
1142 return "<ol><li>" . implode( "</li>\n<li>", $vals ) . '</li></ol>';
1146 return "\n*" . implode( "\n*", $vals );
1149 return "<ul><li>" . implode( "</li>\n<li>", $vals ) . '</li></ul>';
1154 /** Helper function for creating lists of translations.
1156 * @param string $value Value (this is not escaped)
1157 * @param string $lang Lang code of item or false
1158 * @param bool $default If it is default value.
1159 * @param bool $noHtml If to avoid html (for back-compat)
1160 * @throws MWException
1161 * @return string Language item (Note: despite how this looks, this is
1162 * treated as wikitext, not as HTML).
1164 private function langItem( $value, $lang, $default = false, $noHtml = false ) {
1165 if ( $lang === false && $default === false ) {
1166 throw new MWException( '$lang and $default cannot both '
1171 $wrappedValue = $value;
1173 $wrappedValue = '<span class="mw-metadata-lang-value">'
1174 . $value . '</span>';
1177 if ( $lang === false ) {
1178 $msg = $this->msg( 'metadata-langitem-default', $wrappedValue );
1180 return $msg->text() . "\n\n";
1183 return '<li class="mw-metadata-lang-default">'
1188 $lowLang = strtolower( $lang );
1189 $langName = Language
::fetchLanguageName( $lowLang );
1190 if ( $langName === '' ) {
1191 // try just the base language name. (aka en-US -> en ).
1192 $langPrefix = explode( '-', $lowLang, 2 )[0];
1193 $langName = Language
::fetchLanguageName( $langPrefix );
1194 if ( $langName === '' ) {
1199 // else we have a language specified
1201 $msg = $this->msg( 'metadata-langitem', $wrappedValue, $langName, $lang );
1203 return '*' . $msg->text();
1206 $item = '<li class="mw-metadata-lang-code-'
1209 $item .= ' mw-metadata-lang-default';
1211 $item .= '" lang="' . $lang . '">';
1212 $item .= $msg->text();
1219 * Convenience function for getFormattedData()
1221 * @param string $tag The tag name to pass on
1222 * @param string $val The value of the tag
1223 * @param string $arg An argument to pass ($1)
1224 * @param string $arg2 A 2nd argument to pass ($2)
1225 * @return string The text content of "exif-$tag-$val" message in lower case
1227 private function exifMsg( $tag, $val, $arg = null, $arg2 = null ) {
1228 if ( $val === '' ) {
1233 MediaWikiServices
::getInstance()->getContentLanguage()->lc( "exif-$tag-$val" ),
1240 * Format a number, convert numbers from fractions into floating point
1241 * numbers, joins arrays of numbers with commas.
1243 * @param mixed $num The value to format
1244 * @param float|int|bool $round Digits to round to or false.
1245 * @return mixed A floating point number or whatever we were fed
1247 private function formatNum( $num, $round = false ) {
1249 if ( is_array( $num ) ) {
1251 foreach ( $num as $number ) {
1252 $out[] = $this->formatNum( $number );
1255 return $this->getLanguage()->commaList( $out );
1257 if ( preg_match( '/^(-?\d+)\/(\d+)$/', $num, $m ) ) {
1259 $newNum = $m[1] / $m[2];
1260 if ( $round !== false ) {
1261 $newNum = round( $newNum, $round );
1267 return $this->getLanguage()->formatNum( $newNum );
1269 if ( is_numeric( $num ) && $round !== false ) {
1270 $num = round( $num, $round );
1273 return $this->getLanguage()->formatNum( $num );
1278 * Format a rational number, reducing fractions
1280 * @param mixed $num The value to format
1281 * @return mixed A floating point number or whatever we were fed
1283 private function formatFraction( $num ) {
1285 if ( preg_match( '/^(-?\d+)\/(\d+)$/', $num, $m ) ) {
1286 $numerator = intval( $m[1] );
1287 $denominator = intval( $m[2] );
1288 $gcd = $this->gcd( abs( $numerator ), $denominator );
1290 // 0 shouldn't happen! ;)
1291 return $this->formatNum( $numerator / $gcd ) . '/' . $this->formatNum( $denominator / $gcd );
1295 return $this->formatNum( $num );
1299 * Calculate the greatest common divisor of two integers.
1301 * @param int $a Numerator
1302 * @param int $b Denominator
1305 private function gcd( $a, $b ) {
1307 // https://en.wikipedia.org/wiki/Euclidean_algorithm
1308 // Recursive form would be:
1312 return gcd( $b, $a % $b );
1315 $remainder = $a %
$b;
1317 // tail recursion...
1326 * Fetch the human readable version of a news code.
1327 * A news code is an 8 digit code. The first two
1328 * digits are a general classification, so we just
1331 * Note, leading 0's are significant, so this is
1332 * a string, not an int.
1334 * @param string $val The 8 digit news code.
1335 * @return string The human readable form
1337 private function convertNewsCode( $val ) {
1338 if ( !preg_match( '/^\d{8}$/D', $val ) ) {
1339 // Not a valid news code.
1343 switch ( substr( $val, 0, 2 ) ) {
1396 if ( $cat !== '' ) {
1397 $catMsg = $this->exifMsg( 'iimcategory', $cat );
1398 $val = $this->exifMsg( 'subjectnewscode', '', $val, $catMsg );
1405 * Format a coordinate value, convert numbers from floating point
1406 * into degree minute second representation.
1408 * @param int $coord Degrees, minutes and seconds
1409 * @param string $type Latitude or longitude (for if its a NWS or E)
1410 * @return mixed A floating point number or whatever we were fed
1412 private function formatCoords( $coord, $type ) {
1416 if ( $type === 'latitude' ) {
1418 } elseif ( $type === 'longitude' ) {
1423 if ( $type === 'latitude' ) {
1425 } elseif ( $type === 'longitude' ) {
1430 $deg = floor( $nCoord );
1431 $min = floor( ( $nCoord - $deg ) * 60.0 );
1432 $sec = round( ( ( $nCoord - $deg ) - $min / 60 ) * 3600, 2 );
1434 $deg = $this->formatNum( $deg );
1435 $min = $this->formatNum( $min );
1436 $sec = $this->formatNum( $sec );
1438 return $this->msg( 'exif-coordinate-format', $deg, $min, $sec, $ref, $coord )->text();
1442 * Format the contact info field into a single value.
1444 * This function might be called from
1445 * JpegHandler::convertMetadataVersion which is why it is
1448 * @param array $vals Array with fields of the ContactInfo
1449 * struct defined in the IPTC4XMP spec. Or potentially
1450 * an array with one element that is a free form text
1451 * value from the older iptc iim 1:118 prop.
1452 * @return string HTML-ish looking wikitext
1453 * @since 1.23 no longer static
1455 public function collapseContactInfo( $vals ) {
1456 if ( !( isset( $vals['CiAdrExtadr'] )
1457 ||
isset( $vals['CiAdrCity'] )
1458 ||
isset( $vals['CiAdrCtry'] )
1459 ||
isset( $vals['CiEmailWork'] )
1460 ||
isset( $vals['CiTelWork'] )
1461 ||
isset( $vals['CiAdrPcode'] )
1462 ||
isset( $vals['CiAdrRegion'] )
1463 ||
isset( $vals['CiUrlWork'] )
1465 // We don't have any sub-properties
1466 // This could happen if its using old
1467 // iptc that just had this as a free-form
1469 // Note: We run this through htmlspecialchars
1470 // partially to be consistent, and partially
1471 // because people often insert >, etc into
1472 // the metadata which should not be interpreted
1473 // but we still want to auto-link urls.
1474 foreach ( $vals as &$val ) {
1475 $val = htmlspecialchars( $val );
1478 return $this->flattenArrayReal( $vals );
1480 // We have a real ContactInfo field.
1481 // Its unclear if all these fields have to be
1482 // set, so assume they do not.
1483 $url = $tel = $street = $city = $country = '';
1484 $email = $postal = $region = '';
1486 // Also note, some of the class names this uses
1487 // are similar to those used by hCard. This is
1488 // mostly because they're sensible names. This
1489 // does not (and does not attempt to) output
1490 // stuff in the hCard microformat. However it
1491 // might output in the adr microformat.
1493 if ( isset( $vals['CiAdrExtadr'] ) ) {
1494 // Todo: This can potentially be multi-line.
1495 // Need to check how that works in XMP.
1496 $street = '<span class="extended-address">'
1498 $vals['CiAdrExtadr'] )
1501 if ( isset( $vals['CiAdrCity'] ) ) {
1502 $city = '<span class="locality">'
1503 . htmlspecialchars( $vals['CiAdrCity'] )
1506 if ( isset( $vals['CiAdrCtry'] ) ) {
1507 $country = '<span class="country-name">'
1508 . htmlspecialchars( $vals['CiAdrCtry'] )
1511 if ( isset( $vals['CiEmailWork'] ) ) {
1513 // Have to split multiple emails at commas/new lines.
1514 $splitEmails = explode( "\n", $vals['CiEmailWork'] );
1515 foreach ( $splitEmails as $e1 ) {
1516 // Also split on comma
1517 foreach ( explode( ',', $e1 ) as $e2 ) {
1518 $finalEmail = trim( $e2 );
1519 if ( $finalEmail == ',' ||
$finalEmail == '' ) {
1522 if ( strpos( $finalEmail, '<' ) !== false ) {
1523 // Don't do fancy formatting to
1524 // "My name" <foo@bar.com> style stuff
1525 $emails[] = $finalEmail;
1527 $emails[] = '[mailto:'
1529 . ' <span class="email">'
1535 $email = implode( ', ', $emails );
1537 if ( isset( $vals['CiTelWork'] ) ) {
1538 $tel = '<span class="tel">'
1539 . htmlspecialchars( $vals['CiTelWork'] )
1542 if ( isset( $vals['CiAdrPcode'] ) ) {
1543 $postal = '<span class="postal-code">'
1545 $vals['CiAdrPcode'] )
1548 if ( isset( $vals['CiAdrRegion'] ) ) {
1549 // Note this is province/state.
1550 $region = '<span class="region">'
1552 $vals['CiAdrRegion'] )
1555 if ( isset( $vals['CiUrlWork'] ) ) {
1556 $url = '<span class="url">'
1557 . htmlspecialchars( $vals['CiUrlWork'] )
1561 return $this->msg( 'exif-contact-value', $email, $url,
1562 $street, $city, $region, $postal, $country,
1568 * Get a list of fields that are visible by default.
1573 public static function getVisibleFields() {
1575 $lines = explode( "\n", wfMessage( 'metadata-fields' )->inContentLanguage()->text() );
1576 foreach ( $lines as $line ) {
1578 if ( preg_match( '/^\\*\s*(.*?)\s*$/', $line, $matches ) ) {
1579 $fields[] = $matches[1];
1582 $fields = array_map( 'strtolower', $fields );
1588 * Get an array of extended metadata. (See the imageinfo API for format.)
1590 * @param File $file File to use
1591 * @return array [<property name> => ['value' => <value>]], or [] on error
1594 public function fetchExtendedMetadata( File
$file ) {
1595 $cache = MediaWikiServices
::getInstance()->getMainWANObjectCache();
1597 // If revision deleted, exit immediately
1598 if ( $file->isDeleted( File
::DELETED_FILE
) ) {
1602 $cacheKey = $cache->makeKey(
1603 'getExtendedMetadata',
1604 $this->getLanguage()->getCode(),
1605 (int)$this->singleLang
,
1609 $cachedValue = $cache->get( $cacheKey );
1612 && Hooks
::run( 'ValidateExtendedMetadataCache', [ $cachedValue['timestamp'], $file ] )
1614 $extendedMetadata = $cachedValue['data'];
1616 $maxCacheTime = ( $file instanceof ForeignAPIFile
) ?
60 * 60 * 12 : 60 * 60 * 24 * 30;
1617 $fileMetadata = $this->getExtendedMetadataFromFile( $file );
1618 $extendedMetadata = $this->getExtendedMetadataFromHook( $file, $fileMetadata, $maxCacheTime );
1619 if ( $this->singleLang
) {
1620 $this->resolveMultilangMetadata( $extendedMetadata );
1622 $this->discardMultipleValues( $extendedMetadata );
1623 // Make sure the metadata won't break the API when an XML format is used.
1624 // This is an API-specific function so it would be cleaner to call it from
1625 // outside fetchExtendedMetadata, but this way we don't need to redo the
1626 // computation on a cache hit.
1627 $this->sanitizeArrayForAPI( $extendedMetadata );
1628 $valueToCache = [ 'data' => $extendedMetadata, 'timestamp' => wfTimestampNow() ];
1629 $cache->set( $cacheKey, $valueToCache, $maxCacheTime );
1632 return $extendedMetadata;
1636 * Get file-based metadata in standardized format.
1638 * Note that for a remote file, this might return metadata supplied by extensions.
1640 * @param File $file File to use
1641 * @return array [<property name> => ['value' => <value>]], or [] on error
1644 protected function getExtendedMetadataFromFile( File
$file ) {
1645 // If this is a remote file accessed via an API request, we already
1646 // have remote metadata so we just ignore any local one
1647 if ( $file instanceof ForeignAPIFile
) {
1648 // In case of error we pretend no metadata - this will get cached.
1649 // Might or might not be a good idea.
1650 return $file->getExtendedMetadata() ?
: [];
1653 $uploadDate = wfTimestamp( TS_ISO_8601
, $file->getTimestamp() );
1656 // This is modification time, which is close to "upload" time.
1658 'value' => $uploadDate,
1659 'source' => 'mediawiki-metadata',
1663 $title = $file->getTitle();
1665 $text = $title->getText();
1666 $pos = strrpos( $text, '.' );
1669 $name = substr( $text, 0, $pos );
1674 $fileMetadata['ObjectName'] = [
1676 'source' => 'mediawiki-metadata',
1680 return $fileMetadata;
1684 * Get additional metadata from hooks in standardized format.
1686 * @param File $file File to use
1687 * @param array $extendedMetadata
1688 * @param int &$maxCacheTime Hook handlers might use this parameter to override cache time
1690 * @return array [<property name> => ['value' => <value>]], or [] on error
1693 protected function getExtendedMetadataFromHook( File
$file, array $extendedMetadata,
1696 Hooks
::run( 'GetExtendedMetadata', [
1699 $this->getContext(),
1704 $visible = array_flip( self
::getVisibleFields() );
1705 foreach ( $extendedMetadata as $key => $value ) {
1706 if ( !isset( $visible[strtolower( $key )] ) ) {
1707 $extendedMetadata[$key]['hidden'] = '';
1711 return $extendedMetadata;
1715 * Turns an XMP-style multilang array into a single value.
1716 * If the value is not a multilang array, it is returned unchanged.
1717 * See mediawiki.org/wiki/Manual:File_metadata_handling#Multi-language_array_format
1718 * @param mixed $value
1719 * @return mixed Value in best language, null if there were no languages at all
1722 protected function resolveMultilangValue( $value ) {
1725 ||
!isset( $value['_type'] )
1726 ||
$value['_type'] != 'lang'
1728 return $value; // do nothing if not a multilang array
1731 // choose the language best matching user or site settings
1732 $priorityLanguages = $this->getPriorityLanguages();
1733 foreach ( $priorityLanguages as $lang ) {
1734 if ( isset( $value[$lang] ) ) {
1735 return $value[$lang];
1739 // otherwise go with the default language, if set
1740 if ( isset( $value['x-default'] ) ) {
1741 return $value['x-default'];
1744 // otherwise just return any one language
1745 unset( $value['_type'] );
1746 if ( !empty( $value ) ) {
1747 return reset( $value );
1750 // this should not happen; signal error
1755 * Turns an XMP-style multivalue array into a single value by dropping all but the first
1756 * value. If the value is not a multivalue array (or a multivalue array inside a multilang
1757 * array), it is returned unchanged.
1758 * See mediawiki.org/wiki/Manual:File_metadata_handling#Multi-language_array_format
1759 * @param mixed $value
1760 * @return mixed The value, or the first value if there were multiple ones
1763 protected function resolveMultivalueValue( $value ) {
1764 if ( !is_array( $value ) ) {
1766 } elseif ( isset( $value['_type'] ) && $value['_type'] === 'lang' ) {
1767 // if this is a multilang array, process fields separately
1769 foreach ( $value as $k => $v ) {
1770 $newValue[$k] = $this->resolveMultivalueValue( $v );
1773 } else { // _type is 'ul' or 'ol' or missing in which case it defaults to 'ul'
1774 $v = reset( $value );
1775 if ( key( $value ) === '_type' ) {
1776 $v = next( $value );
1783 * Takes an array returned by the getExtendedMetadata* functions,
1784 * and resolves multi-language values in it.
1785 * @param array &$metadata
1788 protected function resolveMultilangMetadata( &$metadata ) {
1789 if ( !is_array( $metadata ) ) {
1792 foreach ( $metadata as &$field ) {
1793 if ( isset( $field['value'] ) ) {
1794 $field['value'] = $this->resolveMultilangValue( $field['value'] );
1800 * Takes an array returned by the getExtendedMetadata* functions,
1801 * and turns all fields into single-valued ones by dropping extra values.
1802 * @param array &$metadata
1805 protected function discardMultipleValues( &$metadata ) {
1806 if ( !is_array( $metadata ) ) {
1809 foreach ( $metadata as $key => &$field ) {
1810 if ( $key === 'Software' ||
$key === 'Contact' ) {
1811 // we skip some fields which have composite values. They are not particularly interesting
1812 // and you can get them via the metadata / commonmetadata APIs anyway.
1815 if ( isset( $field['value'] ) ) {
1816 $field['value'] = $this->resolveMultivalueValue( $field['value'] );
1822 * Makes sure the given array is a valid API response fragment
1823 * @param array &$arr
1825 protected function sanitizeArrayForAPI( &$arr ) {
1826 if ( !is_array( $arr ) ) {
1831 foreach ( $arr as $key => &$value ) {
1832 $sanitizedKey = $this->sanitizeKeyForAPI( $key );
1833 if ( $sanitizedKey !== $key ) {
1834 if ( isset( $arr[$sanitizedKey] ) ) {
1835 // Make the sanitized keys hopefully unique.
1836 // To make it definitely unique would be too much effort, given that
1837 // sanitizing is only needed for misformatted metadata anyway, but
1838 // this at least covers the case when $arr is numeric.
1839 $sanitizedKey .= $counter;
1842 $arr[$sanitizedKey] = $arr[$key];
1843 unset( $arr[$key] );
1845 if ( is_array( $value ) ) {
1846 $this->sanitizeArrayForAPI( $value );
1850 // Handle API metadata keys (particularly "_type")
1851 $keys = array_filter( array_keys( $arr ), 'ApiResult::isMetadataKey' );
1853 ApiResult
::setPreserveKeysList( $arr, $keys );
1858 * Turns a string into a valid API identifier.
1859 * @param string $key
1863 protected function sanitizeKeyForAPI( $key ) {
1864 // drop all characters which are not valid in an XML tag name
1865 // a bunch of non-ASCII letters would be valid but probably won't
1866 // be used so we take the easy way
1867 $key = preg_replace( '/[^a-zA-z0-9_:.\-]/', '', $key );
1868 // drop characters which are invalid at the first position
1869 $key = preg_replace( '/^[\d\-.]+/', '', $key );
1875 // special case for an internal keyword
1876 if ( $key == '_element' ) {
1884 * Returns a list of languages (first is best) to use when formatting multilang fields,
1885 * based on user and site preferences.
1889 protected function getPriorityLanguages() {
1890 $priorityLanguages =
1891 Language
::getFallbacksIncludingSiteLanguage( $this->getLanguage()->getCode() );
1892 $priorityLanguages = array_merge(
1893 (array)$this->getLanguage()->getCode(),
1894 $priorityLanguages[0],
1895 $priorityLanguages[1]
1898 return $priorityLanguages;