Merge "Rename autonym for 'no' from 'norsk bokmål' to 'norsk'"
[lhc/web/wiklou.git] / includes / media / IPTC.php
1 <?php
2 /**
3 * Class for some IPTC functions.
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 */
23
24 /**
25 * Class for some IPTC functions.
26 *
27 * @ingroup Media
28 */
29 class IPTC {
30 /**
31 * This takes the results of iptcparse() and puts it into a
32 * form that can be handled by mediawiki. Generally called from
33 * BitmapMetadataHandler::doApp13.
34 *
35 * @see http://www.iptc.org/std/IIM/4.1/specification/IIMV4.1.pdf
36 *
37 * @param string $rawData The app13 block from jpeg containing iptc/iim data
38 * @return array IPTC metadata array
39 */
40 static function parse( $rawData ) {
41 $parsed = iptcparse( $rawData );
42 $data = [];
43 if ( !is_array( $parsed ) ) {
44 return $data;
45 }
46
47 $c = '';
48 // charset info contained in tag 1:90.
49 if ( isset( $parsed['1#090'] ) && isset( $parsed['1#090'][0] ) ) {
50 $c = self::getCharset( $parsed['1#090'][0] );
51 if ( $c === false ) {
52 // Unknown charset. refuse to parse.
53 // note: There is a different between
54 // unknown and no charset specified.
55 return [];
56 }
57 unset( $parsed['1#090'] );
58 }
59
60 foreach ( $parsed as $tag => $val ) {
61 if ( isset( $val[0] ) && trim( $val[0] ) == '' ) {
62 wfDebugLog( 'iptc', "IPTC tag $tag had only whitespace as its value." );
63 continue;
64 }
65 switch ( $tag ) {
66 case '2#120': /*IPTC caption. mapped with exif ImageDescription*/
67 $data['ImageDescription'] = self::convIPTC( $val, $c );
68 break;
69 case '2#116': /* copyright. Mapped with exif copyright */
70 $data['Copyright'] = self::convIPTC( $val, $c );
71 break;
72 case '2#080': /* byline. Mapped with exif Artist */
73 /* merge with byline title (2:85)
74 * like how exif does it with
75 * Title, person. Not sure if this is best
76 * approach since we no longer have the two fields
77 * separate. each byline title entry corresponds to a
78 * specific byline. */
79
80 $bylines = self::convIPTC( $val, $c );
81 if ( isset( $parsed['2#085'] ) ) {
82 $titles = self::convIPTC( $parsed['2#085'], $c );
83 } else {
84 $titles = [];
85 }
86
87 $titleCount = count( $titles );
88 for ( $i = 0; $i < $titleCount; $i++ ) {
89 if ( isset( $bylines[$i] ) ) {
90 // theoretically this should always be set
91 // but doesn't hurt to be careful.
92 $bylines[$i] = $titles[$i] . ', ' . $bylines[$i];
93 }
94 }
95 $data['Artist'] = $bylines;
96 break;
97 case '2#025': /* keywords */
98 $data['Keywords'] = self::convIPTC( $val, $c );
99 break;
100 case '2#101': /* Country (shown) */
101 $data['CountryDest'] = self::convIPTC( $val, $c );
102 break;
103 case '2#095': /* state/province (shown) */
104 $data['ProvinceOrStateDest'] = self::convIPTC( $val, $c );
105 break;
106 case '2#090': /* city (Shown) */
107 $data['CityDest'] = self::convIPTC( $val, $c );
108 break;
109 case '2#092': /* sublocation (shown) */
110 $data['SublocationDest'] = self::convIPTC( $val, $c );
111 break;
112 case '2#005': /* object name/title */
113 $data['ObjectName'] = self::convIPTC( $val, $c );
114 break;
115 case '2#040': /* special instructions */
116 $data['SpecialInstructions'] = self::convIPTC( $val, $c );
117 break;
118 case '2#105': /* headline */
119 $data['Headline'] = self::convIPTC( $val, $c );
120 break;
121 case '2#110': /* credit */
122 /*"Identifies the provider of the objectdata,
123 * not necessarily the owner/creator". */
124 $data['Credit'] = self::convIPTC( $val, $c );
125 break;
126 case '2#115': /* source */
127 /* "Identifies the original owner of the intellectual content of the
128 *objectdata. This could be an agency, a member of an agency or
129 *an individual." */
130 $data['Source'] = self::convIPTC( $val, $c );
131 break;
132
133 case '2#007': /* edit status (lead, correction, etc) */
134 $data['EditStatus'] = self::convIPTC( $val, $c );
135 break;
136 case '2#015': /* category. deprecated. max 3 letters in theory, often more */
137 $data['iimCategory'] = self::convIPTC( $val, $c );
138 break;
139 case '2#020': /* category. deprecated. */
140 $data['iimSupplementalCategory'] = self::convIPTC( $val, $c );
141 break;
142 case '2#010': /*urgency (1-8. 1 most, 5 normal, 8 low priority)*/
143 $data['Urgency'] = self::convIPTC( $val, $c );
144 break;
145 case '2#022':
146 /* "Identifies objectdata that recurs often and predictably...
147 * Example: Euroweather" */
148 $data['FixtureIdentifier'] = self::convIPTC( $val, $c );
149 break;
150 case '2#026':
151 /* Content location code (iso 3166 + some custom things)
152 * ex: TUR (for turkey), XUN (for UN), XSP (outer space)
153 * See wikipedia article on iso 3166 and appendix D of iim std. */
154 $data['LocationDestCode'] = self::convIPTC( $val, $c );
155 break;
156 case '2#027':
157 /* Content location name. Full printable name
158 * of location of photo. */
159 $data['LocationDest'] = self::convIPTC( $val, $c );
160 break;
161 case '2#065':
162 /* Originating Program.
163 * Combine with Program version (2:70) if present.
164 */
165 $software = self::convIPTC( $val, $c );
166
167 if ( count( $software ) !== 1 ) {
168 // according to iim standard this cannot have multiple values
169 // so if there is more than one, something weird is happening,
170 // and we skip it.
171 wfDebugLog( 'iptc', 'IPTC: Wrong count on 2:65 Software field' );
172 break;
173 }
174
175 if ( isset( $parsed['2#070'] ) ) {
176 // if a version is set for the software.
177 $softwareVersion = self::convIPTC( $parsed['2#070'], $c );
178 unset( $parsed['2#070'] );
179 $data['Software'] = [ [ $software[0], $softwareVersion[0] ] ];
180 } else {
181 $data['Software'] = $software;
182 }
183 break;
184 case '2#075':
185 /* Object cycle.
186 * a for morning (am), p for evening, b for both */
187 $data['ObjectCycle'] = self::convIPTC( $val, $c );
188 break;
189 case '2#100':
190 /* Country/Primary location code.
191 * "Indicates the code of the country/primary location where the
192 * intellectual property of the objectdata was created"
193 * unclear how this differs from 2#026
194 */
195 $data['CountryCodeDest'] = self::convIPTC( $val, $c );
196 break;
197 case '2#103':
198 /* original transmission ref.
199 * "A code representing the location of original transmission ac-
200 * cording to practises of the provider."
201 */
202 $data['OriginalTransmissionRef'] = self::convIPTC( $val, $c );
203 break;
204 case '2#118': /*contact*/
205 $data['Contact'] = self::convIPTC( $val, $c );
206 break;
207 case '2#122':
208 /* Writer/Editor
209 * "Identification of the name of the person involved in the writing,
210 * editing or correcting the objectdata or caption/abstract."
211 */
212 $data['Writer'] = self::convIPTC( $val, $c );
213 break;
214 case '2#135': /* lang code */
215 $data['LanguageCode'] = self::convIPTC( $val, $c );
216 break;
217
218 // Start date stuff.
219 // It doesn't accept incomplete dates even though they are valid
220 // according to spec.
221 // Should potentially store timezone as well.
222 case '2#055':
223 // Date created (not date digitized).
224 // Maps to exif DateTimeOriginal
225 if ( isset( $parsed['2#060'] ) ) {
226 $time = $parsed['2#060'];
227 } else {
228 $time = [];
229 }
230 $timestamp = self::timeHelper( $val, $time, $c );
231 if ( $timestamp ) {
232 $data['DateTimeOriginal'] = $timestamp;
233 }
234 break;
235
236 case '2#062':
237 // Date converted to digital representation.
238 // Maps to exif DateTimeDigitized
239 if ( isset( $parsed['2#063'] ) ) {
240 $time = $parsed['2#063'];
241 } else {
242 $time = [];
243 }
244 $timestamp = self::timeHelper( $val, $time, $c );
245 if ( $timestamp ) {
246 $data['DateTimeDigitized'] = $timestamp;
247 }
248 break;
249
250 case '2#030':
251 // Date released.
252 if ( isset( $parsed['2#035'] ) ) {
253 $time = $parsed['2#035'];
254 } else {
255 $time = [];
256 }
257 $timestamp = self::timeHelper( $val, $time, $c );
258 if ( $timestamp ) {
259 $data['DateTimeReleased'] = $timestamp;
260 }
261 break;
262
263 case '2#037':
264 // Date expires.
265 if ( isset( $parsed['2#038'] ) ) {
266 $time = $parsed['2#038'];
267 } else {
268 $time = [];
269 }
270 $timestamp = self::timeHelper( $val, $time, $c );
271 if ( $timestamp ) {
272 $data['DateTimeExpires'] = $timestamp;
273 }
274 break;
275
276 case '2#000': /* iim version */
277 // unlike other tags, this is a 2-byte binary number.
278 // technically this is required if there is iptc data
279 // but in practise it isn't always there.
280 if ( strlen( $val[0] ) == 2 ) {
281 // if is just to be paranoid.
282 $versionValue = ord( substr( $val[0], 0, 1 ) ) * 256;
283 $versionValue += ord( substr( $val[0], 1, 1 ) );
284 $data['iimVersion'] = $versionValue;
285 }
286 break;
287
288 case '2#004':
289 // IntellectualGenere.
290 // first 4 characters are an id code
291 // That we're not really interested in.
292
293 // This prop is weird, since it's
294 // allowed to have multiple values
295 // in iim 4.1, but not in the XMP
296 // stuff. We're going to just
297 // extract the first value.
298 $con = self::convIPTC( $val, $c );
299 if ( strlen( $con[0] ) < 5 ) {
300 wfDebugLog( 'iptc', 'IPTC: '
301 . '2:04 too short. '
302 . 'Ignoring.' );
303 break;
304 }
305 $extracted = substr( $con[0], 4 );
306 $data['IntellectualGenre'] = $extracted;
307 break;
308
309 case '2#012':
310 // Subject News code - this is a compound field
311 // at the moment we only extract the subject news
312 // code, which is an 8 digit (ascii) number
313 // describing the subject matter of the content.
314 $codes = self::convIPTC( $val, $c );
315 foreach ( $codes as $ic ) {
316 $fields = explode( ':', $ic, 3 );
317
318 if ( count( $fields ) < 2 || $fields[0] !== 'IPTC' ) {
319 wfDebugLog( 'IPTC', 'IPTC: '
320 . 'Invalid 2:12 - ' . $ic );
321 break;
322 }
323 $data['SubjectNewsCode'] = $fields[1];
324 }
325 break;
326
327 // purposely does not do 2:125, 2:130, 2:131,
328 // 2:47, 2:50, 2:45, 2:42, 2:8, 2:3
329 // 2:200, 2:201, 2:202
330 // or the audio stuff (2:150 to 2:154)
331
332 case '2#070':
333 case '2#060':
334 case '2#063':
335 case '2#085':
336 case '2#038':
337 case '2#035':
338 // ignore. Handled elsewhere.
339 break;
340
341 default:
342 wfDebugLog( 'iptc', "Unsupported iptc tag: $tag. Value: " . implode( ',', $val ) );
343 break;
344 }
345 }
346
347 return $data;
348 }
349
350 /**
351 * Convert an iptc date and time tags into the exif format
352 *
353 * @todo Potentially this should also capture the timezone offset.
354 * @param array $date The date tag
355 * @param array $time The time tag
356 * @param string $c The charset
357 * @return string Date in EXIF format.
358 */
359 private static function timeHelper( $date, $time, $c ) {
360 if ( count( $date ) === 1 ) {
361 // the standard says this should always be 1
362 // just double checking.
363 list( $date ) = self::convIPTC( $date, $c );
364 } else {
365 return null;
366 }
367
368 if ( count( $time ) === 1 ) {
369 list( $time ) = self::convIPTC( $time, $c );
370 $dateOnly = false;
371 } else {
372 $time = '000000+0000'; // placeholder
373 $dateOnly = true;
374 }
375
376 if ( !( preg_match( '/\d\d\d\d\d\d[-+]\d\d\d\d/', $time )
377 && preg_match( '/\d\d\d\d\d\d\d\d/', $date )
378 && substr( $date, 0, 4 ) !== '0000'
379 && substr( $date, 4, 2 ) !== '00'
380 && substr( $date, 6, 2 ) !== '00'
381 ) ) {
382 // something wrong.
383 // Note, this rejects some valid dates according to iptc spec
384 // for example: the date 00000400 means the photo was taken in
385 // April, but the year and day is unknown. We don't process these
386 // types of incomplete dates atm.
387 wfDebugLog( 'iptc', "IPTC: invalid time ( $time ) or date ( $date )" );
388
389 return null;
390 }
391
392 $unixTS = wfTimestamp( TS_UNIX, $date . substr( $time, 0, 6 ) );
393 if ( $unixTS === false ) {
394 wfDebugLog( 'iptc', "IPTC: can't convert date to TS_UNIX: $date $time." );
395
396 return null;
397 }
398
399 $tz = ( intval( substr( $time, 7, 2 ) ) * 60 * 60 )
400 + ( intval( substr( $time, 9, 2 ) ) * 60 );
401
402 if ( substr( $time, 6, 1 ) === '-' ) {
403 $tz = -$tz;
404 }
405
406 $finalTimestamp = wfTimestamp( TS_EXIF, $unixTS + $tz );
407 if ( $finalTimestamp === false ) {
408 wfDebugLog( 'iptc', "IPTC: can't make final timestamp. Date: " . ( $unixTS + $tz ) );
409
410 return null;
411 }
412 if ( $dateOnly ) {
413 // return the date only
414 return substr( $finalTimestamp, 0, 10 );
415 } else {
416 return $finalTimestamp;
417 }
418 }
419
420 /**
421 * Helper function to convert charset for iptc values.
422 * @param string|array $data The iptc string
423 * @param string $charset The charset
424 *
425 * @return string|array
426 */
427 private static function convIPTC( $data, $charset ) {
428 if ( is_array( $data ) ) {
429 foreach ( $data as &$val ) {
430 $val = self::convIPTCHelper( $val, $charset );
431 }
432 } else {
433 $data = self::convIPTCHelper( $data, $charset );
434 }
435
436 return $data;
437 }
438
439 /**
440 * Helper function of a helper function to convert charset for iptc values.
441 * @param string|array $data The IPTC string
442 * @param string $charset The charset
443 *
444 * @return string
445 */
446 private static function convIPTCHelper( $data, $charset ) {
447 if ( $charset ) {
448 MediaWiki\suppressWarnings();
449 $data = iconv( $charset, "UTF-8//IGNORE", $data );
450 MediaWiki\restoreWarnings();
451 if ( $data === false ) {
452 $data = "";
453 wfDebugLog( 'iptc', __METHOD__ . " Error converting iptc data charset $charset to utf-8" );
454 }
455 } else {
456 // treat as utf-8 if is valid utf-8. otherwise pretend its windows-1252
457 // most of the time if there is no 1:90 tag, it is either ascii, latin1, or utf-8
458 $oldData = $data;
459 UtfNormal\Validator::quickIsNFCVerify( $data ); // make $data valid utf-8
460 if ( $data === $oldData ) {
461 return $data; // if validation didn't change $data
462 } else {
463 return self::convIPTCHelper( $oldData, 'Windows-1252' );
464 }
465 }
466
467 return trim( $data );
468 }
469
470 /**
471 * take the value of 1:90 tag and returns a charset
472 * @param string $tag 1:90 tag.
473 * @return string Charset name or "?"
474 * Warning, this function does not (and is not intended to) detect
475 * all iso 2022 escape codes. In practise, the code for utf-8 is the
476 * only code that seems to have wide use. It does detect that code.
477 */
478 static function getCharset( $tag ) {
479 // According to iim standard, charset is defined by the tag 1:90.
480 // in which there are iso 2022 escape sequences to specify the character set.
481 // the iim standard seems to encourage that all necessary escape sequences are
482 // in the 1:90 tag, but says it doesn't have to be.
483
484 // This is in need of more testing probably. This is definitely not complete.
485 // however reading the docs of some other iptc software, it appears that most iptc software
486 // only recognizes utf-8. If 1:90 tag is not present content is
487 // usually ascii or iso-8859-1 (and sometimes utf-8), but no guarantee.
488
489 // This also won't work if there are more than one escape sequence in the 1:90 tag
490 // or if something is put in the G2, or G3 charsets, etc. It will only reliably recognize utf-8.
491
492 // This is just going through the charsets mentioned in appendix C of the iim standard.
493
494 // \x1b = ESC.
495 switch ( $tag ) {
496 case "\x1b%G": // utf-8
497 // Also call things that are compatible with utf-8, utf-8 (e.g. ascii)
498 case "\x1b(B": // ascii
499 case "\x1b(@": // iso-646-IRV (ascii in latest version, $ different in older version)
500 $c = 'UTF-8';
501 break;
502 case "\x1b(A": // like ascii, but british.
503 $c = 'ISO646-GB';
504 break;
505 case "\x1b(C": // some obscure sweedish/finland encoding
506 $c = 'ISO-IR-8-1';
507 break;
508 case "\x1b(D":
509 $c = 'ISO-IR-8-2';
510 break;
511 case "\x1b(E": // some obscure danish/norway encoding
512 $c = 'ISO-IR-9-1';
513 break;
514 case "\x1b(F":
515 $c = 'ISO-IR-9-2';
516 break;
517 case "\x1b(G":
518 $c = 'SEN_850200_B'; // aka iso 646-SE; ascii-like
519 break;
520 case "\x1b(I":
521 $c = "ISO646-IT";
522 break;
523 case "\x1b(L":
524 $c = "ISO646-PT";
525 break;
526 case "\x1b(Z":
527 $c = "ISO646-ES";
528 break;
529 case "\x1b([":
530 $c = "GREEK7-OLD";
531 break;
532 case "\x1b(K":
533 $c = "ISO646-DE";
534 break;
535 case "\x1b(N": // crylic
536 $c = "ISO_5427";
537 break;
538 case "\x1b(`": // iso646-NO
539 $c = "NS_4551-1";
540 break;
541 case "\x1b(f": // iso646-FR
542 $c = "NF_Z_62-010";
543 break;
544 case "\x1b(g":
545 $c = "PT2"; // iso646-PT2
546 break;
547 case "\x1b(h":
548 $c = "ES2";
549 break;
550 case "\x1b(i": // iso646-HU
551 $c = "MSZ_7795.3";
552 break;
553 case "\x1b(w":
554 $c = "CSA_Z243.4-1985-1";
555 break;
556 case "\x1b(x":
557 $c = "CSA_Z243.4-1985-2";
558 break;
559 case "\x1b\$(B":
560 case "\x1b\$B":
561 case "\x1b&@\x1b\$B":
562 case "\x1b&@\x1b\$(B":
563 $c = "JIS_C6226-1983";
564 break;
565 case "\x1b-A": // iso-8859-1. at least for the high code characters.
566 case "\x1b(@\x1b-A":
567 case "\x1b(B\x1b-A":
568 $c = 'ISO-8859-1';
569 break;
570 case "\x1b-B": // iso-8859-2. at least for the high code characters.
571 $c = 'ISO-8859-2';
572 break;
573 case "\x1b-C": // iso-8859-3. at least for the high code characters.
574 $c = 'ISO-8859-3';
575 break;
576 case "\x1b-D": // iso-8859-4. at least for the high code characters.
577 $c = 'ISO-8859-4';
578 break;
579 case "\x1b-E": // iso-8859-5. at least for the high code characters.
580 $c = 'ISO-8859-5';
581 break;
582 case "\x1b-F": // iso-8859-6. at least for the high code characters.
583 $c = 'ISO-8859-6';
584 break;
585 case "\x1b-G": // iso-8859-7. at least for the high code characters.
586 $c = 'ISO-8859-7';
587 break;
588 case "\x1b-H": // iso-8859-8. at least for the high code characters.
589 $c = 'ISO-8859-8';
590 break;
591 case "\x1b-I": // CSN_369103. at least for the high code characters.
592 $c = 'CSN_369103';
593 break;
594 default:
595 wfDebugLog( 'iptc', __METHOD__ . 'Unknown charset in iptc 1:90: ' . bin2hex( $tag ) );
596 // at this point just give up and refuse to parse iptc?
597 $c = false;
598 }
599 return $c;
600 }
601 }