xref: /dokuwiki/inc/search.php (revision 8c7c53b0321a3cd3116b8d3b2ad27863a38dece7) 
1<?php
2/**
3 * DokuWiki search functions
4 *
5 * @license    GPL 2 (http://www.gnu.org/licenses/gpl.html)
6 * @author     Andreas Gohr <andi@splitbrain.org>
7 */
8use dokuwiki\Utf8\PhpString;
9use dokuwiki\File\MediaFile;
10use dokuwiki\Utf8\Sort;
11
12/**
13 * Recurse directory
14 *
15 * This function recurses into a given base directory
16 * and calls the supplied function for each file and directory
17 *
18 * @param   array    &$data The results of the search are stored here
19 * @param   string    $base Where to start the search
20 * @param   callback  $func Callback (function name or array with object,method)
21 * @param   array     $opts option array will be given to the Callback
22 * @param   string    $dir  Current directory beyond $base
23 * @param   int       $lvl  Recursion Level
24 * @param   mixed     $sort 'natural' to use natural order sorting (default);
25 *                          'date' to sort by filemtime; leave empty to skip sorting.
26 * @author  Andreas Gohr <andi@splitbrain.org>
27 */
28function search(&$data,$base,$func,$opts,$dir='',$lvl=1,$sort='natural'){
29    $dirs   = [];
30    $files  = [];
31    $filepaths = [];
32
33    // safeguard against runaways #1452
34    if($base == '' || $base == '/') {
35        throw new RuntimeException('No valid $base passed to search() - possible misconfiguration or bug');
36    }
37
38    //read in directories and files
39    $dh = @opendir($base.'/'.$dir);
40    if(!$dh) return;
41    while(($file = readdir($dh)) !== false){
42        if(preg_match('/^[\._]/', $file)) continue; //skip hidden files and upper dirs
43        if(is_dir($base.'/'.$dir.'/'.$file)){
44            $dirs[] = $dir.'/'.$file;
45            continue;
46        }
47        $files[] = $dir.'/'.$file;
48        $filepaths[] = $base.'/'.$dir.'/'.$file;
49    }
50    closedir($dh);
51    if (!empty($sort)) {
52        if ($sort == 'date') {
53            @array_multisort(array_map('filemtime', $filepaths), SORT_NUMERIC, SORT_DESC, $files);
54        } else /* natural */ {
55            Sort::asortFN($files);
56        }
57        Sort::asortFN($dirs);
58    }
59
60    //give directories to userfunction then recurse
61    foreach($dirs as $dir){
62        if (call_user_func_array($func, [&$data, $base, $dir, 'd', $lvl, $opts])){
63            search($data, $base, $func, $opts, $dir, $lvl+1, $sort);
64        }
65    }
66    //now handle the files
67    foreach($files as $file){
68        call_user_func_array($func, [&$data, $base, $file, 'f', $lvl, $opts]);
69    }
70}
71
72/**
73 * The following functions are userfunctions to use with the search
74 * function above. This function is called for every found file or
75 * directory. When a directory is given to the function it has to
76 * decide if this directory should be traversed (true) or not (false)
77 * The function has to accept the following parameters:
78 *
79 * array &$data  - Reference to the result data structure
80 * string $base  - Base usually $conf['datadir']
81 * string $file  - current file or directory relative to $base
82 * string $type  - Type either 'd' for directory or 'f' for file
83 * int    $lvl   - Current recursion depht
84 * array  $opts  - option array as given to search()
85 *
86 * return values for files are ignored
87 *
88 * All functions should check the ACL for document READ rights
89 * namespaces (directories) are NOT checked (when sneaky_index is 0) as this
90 * would break the recursion (You can have an nonreadable dir over a readable
91 * one deeper nested) also make sure to check the file type (for example
92 * in case of lockfiles).
93 */
94
95/**
96 * Searches for pages beginning with the given query
97 *
98 * @author Andreas Gohr <andi@splitbrain.org>
99 *
100 * @param array $data
101 * @param string $base
102 * @param string $file
103 * @param string $type
104 * @param integer $lvl
105 * @param array $opts
106 *
107 * @return bool
108 */
109function search_qsearch(&$data,$base,$file,$type,$lvl,$opts){
110    $opts = [
111        'idmatch'   => '(^|:)'.preg_quote($opts['query'], '/').'/',
112        'listfiles' => true,
113        'pagesonly' => true
114    ];
115    return search_universal($data, $base, $file, $type, $lvl, $opts);
116}
117
118/**
119 * Build the browsable index of pages
120 *
121 * $opts['ns'] is the currently viewed namespace
122 *
123 * @author  Andreas Gohr <andi@splitbrain.org>
124 *
125 * @param array $data
126 * @param string $base
127 * @param string $file
128 * @param string $type
129 * @param integer $lvl
130 * @param array $opts
131 *
132 * @return bool
133 */
134function search_index(&$data,$base,$file,$type,$lvl,$opts){
135    global $conf;
136    $ns = $opts['ns'] ?? '';
137    $opts = [
138        'pagesonly' => true,
139        'listdirs' => true,
140        'listfiles' => empty($opts['nofiles']),
141        'sneakyacl' => $conf['sneaky_index'],
142        // Hacky, should rather use recmatch
143        'depth' => preg_match('#^'.preg_quote($file, '#').'(/|$)#', '/'.$ns) ? 0 : -1,
144    ];
145
146    return search_universal($data, $base, $file, $type, $lvl, $opts);
147}
148
149/**
150 * List all namespaces
151 *
152 * @author  Andreas Gohr <andi@splitbrain.org>
153 *
154 * @param array $data
155 * @param string $base
156 * @param string $file
157 * @param string $type
158 * @param integer $lvl
159 * @param array $opts
160 *
161 * @return bool
162 */
163function search_namespaces(&$data,$base,$file,$type,$lvl,$opts){
164    $opts = ['listdirs' => true];
165    return search_universal($data, $base, $file, $type, $lvl, $opts);
166}
167
168/**
169 * List all mediafiles in a namespace
170 *   $opts['depth']     recursion level, 0 for all
171 *   $opts['showmsg']   shows message if invalid media id is used
172 *   $opts['skipacl']   skip acl checking
173 *   $opts['pattern']   check given pattern
174 *   $opts['hash']      add hashes to result list
175 *
176 * @author  Andreas Gohr <andi@splitbrain.org>
177 *
178 * @param array $data
179 * @param string $base
180 * @param string $file
181 * @param string $type
182 * @param integer $lvl
183 * @param array $opts
184 *
185 * @return bool
186 */
187function search_media(&$data,$base,$file,$type,$lvl,$opts){
188
189    //we do nothing with directories
190    if($type == 'd') {
191        if(empty($opts['depth'])) return true; // recurse forever
192        $depth = substr_count($file, '/');
193        if($depth >= $opts['depth']) return false; // depth reached
194        return true;
195    }
196
197    $info         = [];
198    $info['id']   = pathID($file, true);
199    if($info['id'] != cleanID($info['id'])){
200        if(!empty($opts['showmsg']))
201            msg(hsc($info['id']).' is not a valid file name for DokuWiki - skipped', -1);
202        return false; // skip non-valid files
203    }
204
205    //check ACL for namespace (we have no ACL for mediafiles)
206    $info['perm'] = auth_quickaclcheck(getNS($info['id']).':*');
207    if(empty($opts['skipacl']) && $info['perm'] < AUTH_READ){
208        return false;
209    }
210
211    //check pattern filter
212    if(!empty($opts['pattern']) && !@preg_match($opts['pattern'], $info['id'])){
213        return false;
214    }
215
216    $info['file']     = PhpString::basename($file);
217    $info['size']     = filesize($base.'/'.$file);
218    $info['mtime']    = filemtime($base.'/'.$file);
219    $info['writable'] = is_writable($base.'/'.$file);
220    if(preg_match("/\.(jpe?g|gif|png)$/", $file)){
221        $info['isimg'] = true;
222        $info['meta']  = new JpegMeta($base.'/'.$file);
223    }else{
224        $info['isimg'] = false;
225    }
226    if(!empty($opts['hash'])){
227        $info['hash'] = md5(io_readFile(mediaFN($info['id']), false));
228    }
229
230    $data[] = $info;
231
232    return false;
233}
234
235/**
236 * List all mediafiles in a namespace
237 *   $opts['depth']     recursion level, 0 for all
238 *   $opts['showmsg']   shows message if invalid media id is used
239 *   $opts['skipacl']   skip acl checking
240 *   $opts['pattern']   check given pattern
241 *   $opts['hash']      add hashes to result list
242 *
243 * @todo This is a temporary copy of search_media returning a list of MediaFile intances
244 *
245 * @param array $data
246 * @param string $base
247 * @param string $file
248 * @param string $type
249 * @param integer $lvl
250 * @param array $opts
251 *
252 * @return bool
253 */
254function search_mediafiles(&$data,$base,$file,$type,$lvl,$opts){
255
256    //we do nothing with directories
257    if($type == 'd') {
258        if(empty($opts['depth'])) return true; // recurse forever
259        $depth = substr_count($file, '/');
260        if($depth >= $opts['depth']) return false; // depth reached
261        return true;
262    }
263
264    $id   = pathID($file, true);
265    if($id != cleanID($id)){
266        if($opts['showmsg'])
267            msg(hsc($id).' is not a valid file name for DokuWiki - skipped', -1);
268        return false; // skip non-valid files
269    }
270
271    //check ACL for namespace (we have no ACL for mediafiles)
272    $info['perm'] = auth_quickaclcheck(getNS($id).':*');
273    if(empty($opts['skipacl']) && $info['perm'] < AUTH_READ){
274        return false;
275    }
276
277    //check pattern filter
278    if(!empty($opts['pattern']) && !@preg_match($opts['pattern'], $id)){
279        return false;
280    }
281
282    $data[] = new MediaFile($id);
283    return false;
284}
285
286
287/**
288 * This function just lists documents (for RSS namespace export)
289 *
290 * @author  Andreas Gohr <andi@splitbrain.org>
291 *
292 * @param array $data
293 * @param string $base
294 * @param string $file
295 * @param string $type
296 * @param integer $lvl
297 * @param array $opts
298 *
299 * @return bool
300 */
301function search_list(&$data,$base,$file,$type,$lvl,$opts){
302    //we do nothing with directories
303    if($type == 'd') return false;
304    //only search txt files
305    if(substr($file, -4) == '.txt'){
306        //check ACL
307        $id = pathID($file);
308        if(auth_quickaclcheck($id) < AUTH_READ){
309            return false;
310        }
311        $data[]['id'] = $id;
312    }
313    return false;
314}
315
316/**
317 * Quicksearch for searching matching pagenames
318 *
319 * $opts['query'] is the search query
320 *
321 * @author  Andreas Gohr <andi@splitbrain.org>
322 *
323 * @param array $data
324 * @param string $base
325 * @param string $file
326 * @param string $type
327 * @param integer $lvl
328 * @param array $opts
329 *
330 * @return bool
331 */
332function search_pagename(&$data,$base,$file,$type,$lvl,$opts){
333    //we do nothing with directories
334    if($type == 'd') return true;
335    //only search txt files
336    if(substr($file, -4) != '.txt') return true;
337
338    //simple stringmatching
339    if (!empty($opts['query'])){
340        if(strpos($file, (string) $opts['query']) !== false){
341            //check ACL
342            $id = pathID($file);
343            if(auth_quickaclcheck($id) < AUTH_READ){
344                return false;
345            }
346            $data[]['id'] = $id;
347        }
348    }
349    return true;
350}
351
352/**
353 * Just lists all documents
354 *
355 * $opts['depth']   recursion level, 0 for all
356 * $opts['hash']    do md5 sum of content?
357 * $opts['skipacl'] list everything regardless of ACL
358 *
359 * @author  Andreas Gohr <andi@splitbrain.org>
360 *
361 * @param array $data
362 * @param string $base
363 * @param string $file
364 * @param string $type
365 * @param integer $lvl
366 * @param array $opts
367 *
368 * @return bool
369 */
370function search_allpages(&$data,$base,$file,$type,$lvl,$opts){
371    if(isset($opts['depth']) && $opts['depth']){
372        $parts = explode('/', ltrim($file, '/'));
373        if(($type == 'd' && count($parts) >= $opts['depth'])
374          || ($type != 'd' && count($parts) > $opts['depth'])){
375            return false; // depth reached
376        }
377    }
378
379    //we do nothing with directories
380    if($type == 'd'){
381        return true;
382    }
383
384    //only search txt files
385    if(substr($file, -4) != '.txt') return true;
386
387    $item = [];
388    $item['id']   = pathID($file);
389    if(empty($opts['skipacl']) && auth_quickaclcheck($item['id']) < AUTH_READ){
390        return false;
391    }
392
393    $item['rev']   = filemtime($base.'/'.$file);
394    $item['mtime'] = $item['rev'];
395    $item['size']  = filesize($base.'/'.$file);
396    if(!empty($opts['hash'])){
397        $item['hash'] = md5(trim(rawWiki($item['id'])));
398    }
399
400    $data[] = $item;
401    return true;
402}
403
404/* ------------- helper functions below -------------- */
405
406/**
407 * fulltext sort
408 *
409 * Callback sort function for use with usort to sort the data
410 * structure created by search_fulltext. Sorts descending by count
411 *
412 * @author  Andreas Gohr <andi@splitbrain.org>
413 *
414 * @param array $a
415 * @param array $b
416 *
417 * @return int
418 */
419function sort_search_fulltext($a,$b){
420    if($a['count'] > $b['count']){
421        return -1;
422    }elseif($a['count'] < $b['count']){
423        return 1;
424    }else{
425        return Sort::strcmp($a['id'], $b['id']);
426    }
427}
428
429/**
430 * translates a document path to an ID
431 *
432 * @author  Andreas Gohr <andi@splitbrain.org>
433 * @todo    move to pageutils
434 *
435 * @param string $path
436 * @param bool $keeptxt
437 *
438 * @return mixed|string
439 */
440function pathID($path,$keeptxt=false){
441    $id = utf8_decodeFN($path);
442    $id = str_replace('/', ':', $id);
443    if(!$keeptxt) $id = preg_replace('#\.txt$#', '', $id);
444    $id = trim($id, ':');
445    return $id;
446}
447
448
449/**
450 * This is a very universal callback for the search() function, replacing
451 * many of the former individual functions at the cost of a more complex
452 * setup.
453 *
454 * How the function behaves, depends on the options passed in the $opts
455 * array, where the following settings can be used.
456 *
457 * depth      int     recursion depth. 0 for unlimited                       (default: 0)
458 * keeptxt    bool    keep .txt extension for IDs                            (default: false)
459 * listfiles  bool    include files in listing                               (default: false)
460 * listdirs   bool    include namespaces in listing                          (default: false)
461 * pagesonly  bool    restrict files to pages                                (default: false)
462 * skipacl    bool    do not check for READ permission                       (default: false)
463 * sneakyacl  bool    don't recurse into nonreadable dirs                    (default: false)
464 * hash       bool    create MD5 hash for files                              (default: false)
465 * meta       bool    return file metadata                                   (default: false)
466 * filematch  string  match files against this regexp                        (default: '', so accept everything)
467 * idmatch    string  match full ID against this regexp                      (default: '', so accept everything)
468 * dirmatch   string  match directory against this regexp when adding        (default: '', so accept everything)
469 * nsmatch    string  match namespace against this regexp when adding        (default: '', so accept everything)
470 * recmatch   string  match directory against this regexp when recursing     (default: '', so accept everything)
471 * showmsg    bool    warn about non-ID files                                (default: false)
472 * showhidden bool    show hidden files(e.g. by hidepages config) too        (default: false)
473 * firsthead  bool    return first heading for pages                         (default: false)
474 *
475 * @param array &$data  - Reference to the result data structure
476 * @param string $base  - Base usually $conf['datadir']
477 * @param string $file  - current file or directory relative to $base
478 * @param string $type  - Type either 'd' for directory or 'f' for file
479 * @param int    $lvl   - Current recursion depht
480 * @param array  $opts  - option array as given to search()
481 * @return bool if this directory should be traversed (true) or not (false)
482 *              return value is ignored for files
483 *
484 * @author Andreas Gohr <gohr@cosmocode.de>
485 */
486function search_universal(&$data,$base,$file,$type,$lvl,$opts){
487    $item   = [];
488    $return = true;
489
490    // get ID and check if it is a valid one
491    $item['id'] = pathID($file, ($type == 'd' || !empty($opts['keeptxt'])));
492    if($item['id'] != cleanID($item['id'])){
493        if(!empty($opts['showmsg'])){
494            msg(hsc($item['id']).' is not a valid file name for DokuWiki - skipped', -1);
495        }
496        return false; // skip non-valid files
497    }
498    $item['ns']  = getNS($item['id']);
499
500    if($type == 'd') {
501        // decide if to recursion into this directory is wanted
502        if(empty($opts['depth'])){
503            $return = true; // recurse forever
504        }else{
505            $depth = substr_count($file, '/');
506            if($depth >= $opts['depth']){
507                $return = false; // depth reached
508            }else{
509                $return = true;
510            }
511        }
512
513        if ($return) {
514            $match = empty($opts['recmatch']) || preg_match('/'.$opts['recmatch'].'/', $file);
515            if (!$match) {
516                return false; // doesn't match
517            }
518        }
519    }
520
521    // check ACL
522    if(empty($opts['skipacl'])){
523        if($type == 'd'){
524            $item['perm'] = auth_quickaclcheck($item['id'].':*');
525        }else{
526            $item['perm'] = auth_quickaclcheck($item['id']); //FIXME check namespace for media files
527        }
528    }else{
529        $item['perm'] = AUTH_DELETE;
530    }
531
532    // are we done here maybe?
533    if($type == 'd'){
534        if(empty($opts['listdirs'])) return $return;
535        //neither list nor recurse forbidden items:
536        if(empty($opts['skipacl']) && !empty($opts['sneakyacl']) && $item['perm'] < AUTH_READ) return false;
537        if(!empty($opts['dirmatch']) && !preg_match('/'.$opts['dirmatch'].'/', $file)) return $return;
538        if(!empty($opts['nsmatch']) && !preg_match('/'.$opts['nsmatch'].'/', $item['ns'])) return $return;
539    }else{
540        if(empty($opts['listfiles'])) return $return;
541        if(empty($opts['skipacl']) && $item['perm'] < AUTH_READ) return $return;
542        if(!empty($opts['pagesonly']) && (substr($file, -4) != '.txt')) return $return;
543        if(empty($opts['showhidden']) && isHiddenPage($item['id'])) return $return;
544        if(!empty($opts['filematch']) && !preg_match('/'.$opts['filematch'].'/', $file)) return $return;
545        if(!empty($opts['idmatch']) && !preg_match('/'.$opts['idmatch'].'/', $item['id'])) return $return;
546    }
547
548    // still here? prepare the item
549    $item['type']  = $type;
550    $item['level'] = $lvl;
551    $item['open']  = $return;
552
553    if(!empty($opts['meta'])){
554        $item['file']       = PhpString::basename($file);
555        $item['size']       = filesize($base.'/'.$file);
556        $item['mtime']      = filemtime($base.'/'.$file);
557        $item['rev']        = $item['mtime'];
558        $item['writable']   = is_writable($base.'/'.$file);
559        $item['executable'] = is_executable($base.'/'.$file);
560    }
561
562    if($type == 'f'){
563        if(!empty($opts['hash'])) $item['hash'] = md5(io_readFile($base.'/'.$file, false));
564        if(!empty($opts['firsthead'])) $item['title'] = p_get_first_heading($item['id'], METADATA_DONT_RENDER);
565    }
566
567    // finally add the item
568    $data[] = $item;
569    return $return;
570}
571
572//Setup VIM: ex: et ts=4 :
573