Merge "API: Make jsonfm the default output format"
[lhc/web/wiklou.git] / includes / api / ApiQueryBacklinks.php
1 <?php
2 /**
3 *
4 *
5 * Created on Oct 16, 2006
6 *
7 * Copyright © 2006 Yuri Astrakhan "<Firstname><Lastname>@gmail.com"
8 *
9 * This program is free software; you can redistribute it and/or modify
10 * it under the terms of the GNU General Public License as published by
11 * the Free Software Foundation; either version 2 of the License, or
12 * (at your option) any later version.
13 *
14 * This program is distributed in the hope that it will be useful,
15 * but WITHOUT ANY WARRANTY; without even the implied warranty of
16 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
17 * GNU General Public License for more details.
18 *
19 * You should have received a copy of the GNU General Public License along
20 * with this program; if not, write to the Free Software Foundation, Inc.,
21 * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
22 * http://www.gnu.org/copyleft/gpl.html
23 *
24 * @file
25 */
26
27 /**
28 * This is a three-in-one module to query:
29 * * backlinks - links pointing to the given page,
30 * * embeddedin - what pages transclude the given page within themselves,
31 * * imageusage - what pages use the given image
32 *
33 * @ingroup API
34 */
35 class ApiQueryBacklinks extends ApiQueryGeneratorBase {
36
37 /**
38 * @var Title
39 */
40 private $rootTitle;
41
42 private $params, $cont, $redirect;
43 private $bl_ns, $bl_from, $bl_from_ns, $bl_table, $bl_code, $bl_title, $bl_fields, $hasNS;
44
45 /**
46 * Maps ns and title to pageid
47 *
48 * @var array
49 */
50 private $pageMap = array();
51 private $resultArr;
52
53 private $redirTitles = array();
54 private $continueStr = null;
55
56 // output element name, database column field prefix, database table
57 private $backlinksSettings = array(
58 'backlinks' => array(
59 'code' => 'bl',
60 'prefix' => 'pl',
61 'linktbl' => 'pagelinks',
62 'helpurl' => 'https://www.mediawiki.org/wiki/API:Backlinks',
63 ),
64 'embeddedin' => array(
65 'code' => 'ei',
66 'prefix' => 'tl',
67 'linktbl' => 'templatelinks',
68 'helpurl' => 'https://www.mediawiki.org/wiki/API:Embeddedin',
69 ),
70 'imageusage' => array(
71 'code' => 'iu',
72 'prefix' => 'il',
73 'linktbl' => 'imagelinks',
74 'helpurl' => 'https://www.mediawiki.org/wiki/API:Imageusage',
75 )
76 );
77
78 public function __construct( ApiQuery $query, $moduleName ) {
79 $settings = $this->backlinksSettings[$moduleName];
80 $prefix = $settings['prefix'];
81 $code = $settings['code'];
82 $this->resultArr = array();
83
84 parent::__construct( $query, $moduleName, $code );
85 $this->bl_ns = $prefix . '_namespace';
86 $this->bl_from = $prefix . '_from';
87 $this->bl_from_ns = $prefix . '_from_namespace';
88 $this->bl_table = $settings['linktbl'];
89 $this->bl_code = $code;
90 $this->helpUrl = $settings['helpurl'];
91
92 $this->hasNS = $moduleName !== 'imageusage';
93 if ( $this->hasNS ) {
94 $this->bl_title = $prefix . '_title';
95 $this->bl_fields = array(
96 $this->bl_ns,
97 $this->bl_title
98 );
99 } else {
100 $this->bl_title = $prefix . '_to';
101 $this->bl_fields = array(
102 $this->bl_title
103 );
104 }
105 }
106
107 public function execute() {
108 $this->run();
109 }
110
111 public function getCacheMode( $params ) {
112 return 'public';
113 }
114
115 public function executeGenerator( $resultPageSet ) {
116 $this->run( $resultPageSet );
117 }
118
119 /**
120 * @param ApiPageSet $resultPageSet
121 * @return void
122 */
123 private function runFirstQuery( $resultPageSet = null ) {
124 $this->addTables( array( $this->bl_table, 'page' ) );
125 $this->addWhere( "{$this->bl_from}=page_id" );
126 if ( is_null( $resultPageSet ) ) {
127 $this->addFields( array( 'page_id', 'page_title', 'page_namespace' ) );
128 } else {
129 $this->addFields( $resultPageSet->getPageTableFields() );
130 }
131 $this->addFields( array( 'page_is_redirect', 'from_ns' => 'page_namespace' ) );
132
133 $this->addWhereFld( $this->bl_title, $this->rootTitle->getDBkey() );
134 if ( $this->hasNS ) {
135 $this->addWhereFld( $this->bl_ns, $this->rootTitle->getNamespace() );
136 }
137 $this->addWhereFld( $this->bl_from_ns, $this->params['namespace'] );
138
139 if ( count( $this->cont ) >= 2 ) {
140 $op = $this->params['dir'] == 'descending' ? '<' : '>';
141 if ( count( $this->params['namespace'] ) > 1 ) {
142 $this->addWhere(
143 "{$this->bl_from_ns} $op {$this->cont[0]} OR " .
144 "({$this->bl_from_ns} = {$this->cont[0]} AND " .
145 "{$this->bl_from} $op= {$this->cont[1]})"
146 );
147 } else {
148 $this->addWhere( "{$this->bl_from} $op= {$this->cont[1]}" );
149 }
150 }
151
152 if ( $this->params['filterredir'] == 'redirects' ) {
153 $this->addWhereFld( 'page_is_redirect', 1 );
154 } elseif ( $this->params['filterredir'] == 'nonredirects' && !$this->redirect ) {
155 // bug 22245 - Check for !redirect, as filtering nonredirects, when
156 // getting what links to them is contradictory
157 $this->addWhereFld( 'page_is_redirect', 0 );
158 }
159
160 $this->addOption( 'LIMIT', $this->params['limit'] + 1 );
161 $sort = ( $this->params['dir'] == 'descending' ? ' DESC' : '' );
162 $orderBy = array();
163 if ( count( $this->params['namespace'] ) > 1 ) {
164 $orderBy[] = $this->bl_from_ns . $sort;
165 }
166 $orderBy[] = $this->bl_from . $sort;
167 $this->addOption( 'ORDER BY', $orderBy );
168 $this->addOption( 'STRAIGHT_JOIN' );
169
170 $res = $this->select( __METHOD__ );
171 $count = 0;
172 foreach ( $res as $row ) {
173 if ( ++$count > $this->params['limit'] ) {
174 // We've reached the one extra which shows that there are
175 // additional pages to be had. Stop here...
176 // Continue string may be overridden at a later step
177 $this->continueStr = "{$row->from_ns}|{$row->page_id}";
178 break;
179 }
180
181 // Fill in continuation fields for later steps
182 if ( count( $this->cont ) < 2 ) {
183 $this->cont[] = $row->from_ns;
184 $this->cont[] = $row->page_id;
185 }
186
187 $this->pageMap[$row->page_namespace][$row->page_title] = $row->page_id;
188 $t = Title::makeTitle( $row->page_namespace, $row->page_title );
189 if ( $row->page_is_redirect ) {
190 $this->redirTitles[] = $t;
191 }
192
193 if ( is_null( $resultPageSet ) ) {
194 $a = array( 'pageid' => intval( $row->page_id ) );
195 ApiQueryBase::addTitleInfo( $a, $t );
196 if ( $row->page_is_redirect ) {
197 $a['redirect'] = '';
198 }
199 // Put all the results in an array first
200 $this->resultArr[$a['pageid']] = $a;
201 } else {
202 $resultPageSet->processDbRow( $row );
203 }
204 }
205 }
206
207 /**
208 * @param ApiPageSet $resultPageSet
209 * @return void
210 */
211 private function runSecondQuery( $resultPageSet = null ) {
212 $db = $this->getDB();
213 $this->addTables( array( 'page', $this->bl_table ) );
214 $this->addWhere( "{$this->bl_from}=page_id" );
215
216 if ( is_null( $resultPageSet ) ) {
217 $this->addFields( array( 'page_id', 'page_title', 'page_namespace', 'page_is_redirect' ) );
218 } else {
219 $this->addFields( $resultPageSet->getPageTableFields() );
220 }
221
222 $this->addFields( array( $this->bl_title, 'from_ns' => 'page_namespace' ) );
223 if ( $this->hasNS ) {
224 $this->addFields( $this->bl_ns );
225 }
226
227 // We can't use LinkBatch here because $this->hasNS may be false
228 $titleWhere = array();
229 $allRedirNs = array();
230 $allRedirDBkey = array();
231 /** @var $t Title */
232 foreach ( $this->redirTitles as $t ) {
233 $redirNs = $t->getNamespace();
234 $redirDBkey = $t->getDBkey();
235 $titleWhere[] = "{$this->bl_title} = " . $db->addQuotes( $redirDBkey ) .
236 ( $this->hasNS ? " AND {$this->bl_ns} = {$redirNs}" : '' );
237 $allRedirNs[$redirNs] = true;
238 $allRedirDBkey[$redirDBkey] = true;
239 }
240 $this->addWhere( $db->makeList( $titleWhere, LIST_OR ) );
241 $this->addWhereFld( 'page_namespace', $this->params['namespace'] );
242
243 if ( count( $this->cont ) >= 6 ) {
244 $op = $this->params['dir'] == 'descending' ? '<' : '>';
245
246 $where = "{$this->bl_from} $op= {$this->cont[5]}";
247 // Don't bother with namespace, title, or from_namespace if it's
248 // otherwise constant in the where clause.
249 if ( count( $this->params['namespace'] ) > 1 ) {
250 $where = "{$this->bl_from_ns} $op {$this->cont[4]} OR " .
251 "({$this->bl_from_ns} = {$this->cont[4]} AND ($where))";
252 }
253 if ( count( $allRedirDBkey ) > 1 ) {
254 $title = $db->addQuotes( $this->cont[3] );
255 $where = "{$this->bl_title} $op $title OR " .
256 "({$this->bl_title} = $title AND ($where))";
257 }
258 if ( $this->hasNS && count( $allRedirNs ) > 1 ) {
259 $where = "{$this->bl_ns} $op {$this->cont[2]} OR " .
260 "({$this->bl_ns} = {$this->cont[2]} AND ($where))";
261 }
262
263 $this->addWhere( $where );
264 }
265 if ( $this->params['filterredir'] == 'redirects' ) {
266 $this->addWhereFld( 'page_is_redirect', 1 );
267 } elseif ( $this->params['filterredir'] == 'nonredirects' ) {
268 $this->addWhereFld( 'page_is_redirect', 0 );
269 }
270
271 $this->addOption( 'LIMIT', $this->params['limit'] + 1 );
272 $orderBy = array();
273 $sort = ( $this->params['dir'] == 'descending' ? ' DESC' : '' );
274 // Don't order by namespace/title/from_namespace if it's constant in the WHERE clause
275 if ( $this->hasNS && count( $allRedirNs ) > 1 ) {
276 $orderBy[] = $this->bl_ns . $sort;
277 }
278 if ( count( $allRedirDBkey ) > 1 ) {
279 $orderBy[] = $this->bl_title . $sort;
280 }
281 if ( count( $this->params['namespace'] ) > 1 ) {
282 $orderBy[] = $this->bl_from_ns . $sort;
283 }
284 $orderBy[] = $this->bl_from . $sort;
285 $this->addOption( 'ORDER BY', $orderBy );
286 $this->addOption( 'USE INDEX', array( 'page' => 'PRIMARY' ) );
287
288 $res = $this->select( __METHOD__ );
289 $count = 0;
290 foreach ( $res as $row ) {
291 $ns = $this->hasNS ? $row->{$this->bl_ns} : NS_FILE;
292
293 if ( ++$count > $this->params['limit'] ) {
294 // We've reached the one extra which shows that there are
295 // additional pages to be had. Stop here...
296 // Note we must keep the parameters for the first query constant
297 // This may be overridden at a later step
298 $title = $row->{$this->bl_title};
299 $this->continueStr = join( '|', array_slice( $this->cont, 0, 2 ) ) .
300 "|$ns|$title|{$row->from_ns}|{$row->page_id}";
301 break;
302 }
303
304 // Fill in continuation fields for later steps
305 if ( count( $this->cont ) < 6 ) {
306 $this->cont[] = $ns;
307 $this->cont[] = $row->{$this->bl_title};
308 $this->cont[] = $row->from_ns;
309 $this->cont[] = $row->page_id;
310 }
311
312 if ( is_null( $resultPageSet ) ) {
313 $a['pageid'] = intval( $row->page_id );
314 ApiQueryBase::addTitleInfo( $a, Title::makeTitle( $row->page_namespace, $row->page_title ) );
315 if ( $row->page_is_redirect ) {
316 $a['redirect'] = '';
317 }
318 $parentID = $this->pageMap[$ns][$row->{$this->bl_title}];
319 // Put all the results in an array first
320 $this->resultArr[$parentID]['redirlinks'][$row->page_id] = $a;
321 } else {
322 $resultPageSet->processDbRow( $row );
323 }
324 }
325 }
326
327 /**
328 * @param ApiPageSet $resultPageSet
329 * @return void
330 */
331 private function run( $resultPageSet = null ) {
332 $this->params = $this->extractRequestParams( false );
333 $this->redirect = isset( $this->params['redirect'] ) && $this->params['redirect'];
334 $userMax = ( $this->redirect ? ApiBase::LIMIT_BIG1 / 2 : ApiBase::LIMIT_BIG1 );
335 $botMax = ( $this->redirect ? ApiBase::LIMIT_BIG2 / 2 : ApiBase::LIMIT_BIG2 );
336
337 $result = $this->getResult();
338
339 if ( $this->params['limit'] == 'max' ) {
340 $this->params['limit'] = $this->getMain()->canApiHighLimits() ? $botMax : $userMax;
341 $result->setParsedLimit( $this->getModuleName(), $this->params['limit'] );
342 } else {
343 $this->params['limit'] = intval( $this->params['limit'] );
344 $this->validateLimit( 'limit', $this->params['limit'], 1, $userMax, $botMax );
345 }
346
347 $this->rootTitle = $this->getTitleOrPageId( $this->params )->getTitle();
348
349 // only image titles are allowed for the root in imageinfo mode
350 if ( !$this->hasNS && $this->rootTitle->getNamespace() !== NS_FILE ) {
351 $this->dieUsage(
352 "The title for {$this->getModuleName()} query must be a file",
353 'bad_image_title'
354 );
355 }
356
357 // Parse and validate continuation parameter
358 $this->cont = array();
359 if ( $this->params['continue'] !== null ) {
360 $db = $this->getDB();
361 $cont = explode( '|', $this->params['continue'] );
362
363 switch ( count( $cont ) ) {
364 case 8:
365 // redirect page ID for result adding
366 $this->cont[7] = (int)$cont[7];
367 $this->dieContinueUsageIf( $cont[7] !== (string)$this->cont[7] );
368
369 /* Fall through */
370
371 case 7:
372 // top-level page ID for result adding
373 $this->cont[6] = (int)$cont[6];
374 $this->dieContinueUsageIf( $cont[6] !== (string)$this->cont[6] );
375
376 /* Fall through */
377
378 case 6:
379 // ns for 2nd query (even for imageusage)
380 $this->cont[2] = (int)$cont[2];
381 $this->dieContinueUsageIf( $cont[2] !== (string)$this->cont[2] );
382
383 // title for 2nd query
384 $this->cont[3] = $cont[3];
385
386 // from_ns for 2nd query
387 $this->cont[4] = (int)$cont[4];
388 $this->dieContinueUsageIf( $cont[4] !== (string)$this->cont[4] );
389
390 // from_id for 1st query
391 $this->cont[5] = (int)$cont[5];
392 $this->dieContinueUsageIf( $cont[5] !== (string)$this->cont[5] );
393
394 /* Fall through */
395
396 case 2:
397 // from_ns for 1st query
398 $this->cont[0] = (int)$cont[0];
399 $this->dieContinueUsageIf( $cont[0] !== (string)$this->cont[0] );
400
401 // from_id for 1st query
402 $this->cont[1] = (int)$cont[1];
403 $this->dieContinueUsageIf( $cont[1] !== (string)$this->cont[1] );
404
405 break;
406
407 default:
408 $this->dieContinueUsageIf( true );
409 }
410
411 ksort( $this->cont );
412 }
413
414 $this->runFirstQuery( $resultPageSet );
415 if ( $this->redirect && count( $this->redirTitles ) ) {
416 $this->resetQueryParams();
417 $this->runSecondQuery( $resultPageSet );
418 }
419
420 // Fill in any missing fields in case it's needed below
421 $this->cont += array( 0, 0, 0, '', 0, 0, 0 );
422
423 if ( is_null( $resultPageSet ) ) {
424 // Try to add the result data in one go and pray that it fits
425 $code = $this->bl_code;
426 $data = array_map( function ( $arr ) use ( $result, $code ) {
427 if ( isset( $arr['redirlinks'] ) ) {
428 $arr['redirlinks'] = array_values( $arr['redirlinks'] );
429 $result->setIndexedTagName( $arr['redirlinks'], $code );
430 }
431 return $arr;
432 }, array_values( $this->resultArr ) );
433 $fit = $result->addValue( 'query', $this->getModuleName(), $data );
434 if ( !$fit ) {
435 // It didn't fit. Add elements one by one until the
436 // result is full.
437 ksort( $this->resultArr );
438 if ( count( $this->cont ) >= 7 ) {
439 $startAt = $this->cont[6];
440 } else {
441 reset( $this->resultArr );
442 $startAt = key( $this->resultArr );
443 }
444 $idx = 0;
445 foreach ( $this->resultArr as $pageID => $arr ) {
446 if ( $pageID < $startAt ) {
447 continue;
448 }
449
450 // Add the basic entry without redirlinks first
451 $fit = $result->addValue(
452 array( 'query', $this->getModuleName() ),
453 $idx, array_diff_key( $arr, array( 'redirlinks' => '' ) ) );
454 if ( !$fit ) {
455 $this->continueStr = join( '|', array_slice( $this->cont, 0, 6 ) ) .
456 "|$pageID";
457 break;
458 }
459
460 $hasRedirs = false;
461 $redirLinks = isset( $arr['redirlinks'] ) ? (array)$arr['redirlinks'] : array();
462 ksort( $redirLinks );
463 if ( count( $this->cont ) >= 8 && $pageID == $startAt ) {
464 $redirStartAt = $this->cont[7];
465 } else {
466 reset( $redirLinks );
467 $redirStartAt = key( $redirLinks );
468 }
469 foreach ( $redirLinks as $key => $redir ) {
470 if ( $key < $redirStartAt ) {
471 continue;
472 }
473
474 $fit = $result->addValue(
475 array( 'query', $this->getModuleName(), $idx, 'redirlinks' ),
476 null, $redir );
477 if ( !$fit ) {
478 $this->continueStr = join( '|', array_slice( $this->cont, 0, 6 ) ) .
479 "|$pageID|$key";
480 break;
481 }
482 $hasRedirs = true;
483 }
484 if ( $hasRedirs ) {
485 $result->setIndexedTagName_internal(
486 array( 'query', $this->getModuleName(), $idx, 'redirlinks' ),
487 $this->bl_code );
488 }
489 if ( !$fit ) {
490 break;
491 }
492
493 $idx++;
494 }
495 }
496
497 $result->setIndexedTagName_internal(
498 array( 'query', $this->getModuleName() ),
499 $this->bl_code
500 );
501 }
502 if ( !is_null( $this->continueStr ) ) {
503 $this->setContinueEnumParameter( 'continue', $this->continueStr );
504 }
505 }
506
507 public function getAllowedParams() {
508 $retval = array(
509 'title' => array(
510 ApiBase::PARAM_TYPE => 'string',
511 ),
512 'pageid' => array(
513 ApiBase::PARAM_TYPE => 'integer',
514 ),
515 'continue' => null,
516 'namespace' => array(
517 ApiBase::PARAM_ISMULTI => true,
518 ApiBase::PARAM_TYPE => 'namespace'
519 ),
520 'dir' => array(
521 ApiBase::PARAM_DFLT => 'ascending',
522 ApiBase::PARAM_TYPE => array(
523 'ascending',
524 'descending'
525 )
526 ),
527 'filterredir' => array(
528 ApiBase::PARAM_DFLT => 'all',
529 ApiBase::PARAM_TYPE => array(
530 'all',
531 'redirects',
532 'nonredirects'
533 )
534 ),
535 'limit' => array(
536 ApiBase::PARAM_DFLT => 10,
537 ApiBase::PARAM_TYPE => 'limit',
538 ApiBase::PARAM_MIN => 1,
539 ApiBase::PARAM_MAX => ApiBase::LIMIT_BIG1,
540 ApiBase::PARAM_MAX2 => ApiBase::LIMIT_BIG2
541 )
542 );
543 if ( $this->getModuleName() == 'embeddedin' ) {
544 return $retval;
545 }
546 $retval['redirect'] = false;
547
548 return $retval;
549 }
550
551 public function getParamDescription() {
552 $retval = array(
553 'title' => "Title to search. Cannot be used together with {$this->bl_code}pageid",
554 'pageid' => "Pageid to search. Cannot be used together with {$this->bl_code}title",
555 'continue' => 'When more results are available, use this to continue',
556 'namespace' => 'The namespace to enumerate',
557 'dir' => 'The direction in which to list',
558 );
559 if ( $this->getModuleName() != 'embeddedin' ) {
560 return array_merge( $retval, array(
561 'redirect' => 'If linking page is a redirect, find all pages ' .
562 'that link to that redirect as well. Maximum limit is halved.',
563 'filterredir' => 'How to filter for redirects. If set to ' .
564 "nonredirects when {$this->bl_code}redirect is enabled, " .
565 'this is only applied to the second level',
566 'limit' => 'How many total pages to return. If ' .
567 "{$this->bl_code}redirect is enabled, limit applies to each " .
568 'level separately (which means you may get up to 2 * limit results).'
569 ) );
570 }
571
572 return array_merge( $retval, array(
573 'filterredir' => 'How to filter for redirects',
574 'limit' => 'How many total pages to return'
575 ) );
576 }
577
578 public function getDescription() {
579 switch ( $this->getModuleName() ) {
580 case 'backlinks':
581 return 'Find all pages that link to the given page.';
582 case 'embeddedin':
583 return 'Find all pages that embed (transclude) the given title.';
584 case 'imageusage':
585 return 'Find all pages that use the given image title.';
586 default:
587 ApiBase::dieDebug( __METHOD__, 'Unknown module name.' );
588 }
589 }
590
591 public function getExamples() {
592 static $examples = array(
593 'backlinks' => array(
594 'api.php?action=query&list=backlinks&bltitle=Main%20Page',
595 'api.php?action=query&generator=backlinks&gbltitle=Main%20Page&prop=info'
596 ),
597 'embeddedin' => array(
598 'api.php?action=query&list=embeddedin&eititle=Template:Stub',
599 'api.php?action=query&generator=embeddedin&geititle=Template:Stub&prop=info'
600 ),
601 'imageusage' => array(
602 'api.php?action=query&list=imageusage&iutitle=File:Albert%20Einstein%20Head.jpg',
603 'api.php?action=query&generator=imageusage&giutitle=File:Albert%20Einstein%20Head.jpg&prop=info'
604 )
605 );
606
607 return $examples[$this->getModuleName()];
608 }
609
610 public function getHelpUrls() {
611 return $this->helpUrl;
612 }
613 }