xref: /dokuwiki/inc/search.php (revision 15f699ac4db38c7098b4ae4cd0782dff13d46637) 
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 */
8
9use dokuwiki\Utf8\Sort;
10
11/**
12 * Recurse directory
13 *
14 * This function recurses into a given base directory
15 * and calls the supplied function for each file and directory
16 *
17 * @param   array    &$data The results of the search are stored here
18 * @param   string    $base Where to start the search
19 * @param   callback  $func Callback (function name or array with object,method)
20 * @param   array     $opts option array will be given to the Callback
21 * @param   string    $dir  Current directory beyond $base
22 * @param   int       $lvl  Recursion Level
23 * @param   mixed     $sort 'natural' to use natural order sorting (default);
24 *                          'date' to sort by filemtime; leave empty to skip sorting.
25 * @author  Andreas Gohr <andi@splitbrain.org>
26 */
27function search(&$data, $base, $func, $opts, $dir='', $lvl=1, $sort='natural')
28{
29    $dirs   = array();
30    $files  = array();
31    $filepaths = array();
32
33    // safeguard against runaways #1452
34    if ($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, array(&$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, array(&$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{
111    $opts = array(
112            'idmatch'   => '(^|:)'.preg_quote($opts['query'],'/').'/',
113            'listfiles' => true,
114            'pagesonly' => true,
115            );
116    return search_universal($data, $base, $file, $type, $lvl, $opts);
117}
118
119/**
120 * Build the browsable index of pages
121 *
122 * $opts['ns'] is the currently viewed namespace
123 *
124 * @author  Andreas Gohr <andi@splitbrain.org>
125 *
126 * @param array $data
127 * @param string $base
128 * @param string $file
129 * @param string $type
130 * @param integer $lvl
131 * @param array $opts
132 *
133 * @return bool
134 */
135function search_index(&$data, $base, $file, $type, $lvl, $opts)
136{
137    global $conf;
138    $opts = array(
139        'pagesonly' => true,
140        'listdirs' => true,
141        'listfiles' => empty($opts['nofiles']),
142        'sneakyacl' => $conf['sneaky_index'],
143        // Hacky, should rather use recmatch
144        'depth' => preg_match('#^'.preg_quote($file, '#').'(/|$)#','/'.$opts['ns']) ? 0 : -1
145    );
146
147    return search_universal($data, $base, $file, $type, $lvl, $opts);
148}
149
150/**
151 * List all namespaces
152 *
153 * @author  Andreas Gohr <andi@splitbrain.org>
154 *
155 * @param array $data
156 * @param string $base
157 * @param string $file
158 * @param string $type
159 * @param integer $lvl
160 * @param array $opts
161 *
162 * @return bool
163 */
164function search_namespaces(&$data, $base, $file, $type, $lvl, $opts)
165{
166    $opts = array(
167            'listdirs' => true,
168            );
169    return search_universal($data, $base, $file, $type, $lvl, $opts);
170}
171
172/**
173 * List all mediafiles in a namespace
174 *   $opts['depth']     recursion level, 0 for all
175 *   $opts['showmsg']   shows message if invalid media id is used
176 *   $opts['skipacl']   skip acl checking
177 *   $opts['pattern']   check given pattern
178 *   $opts['hash']      add hashes to result list
179 *
180 * @author  Andreas Gohr <andi@splitbrain.org>
181 *
182 * @param array $data
183 * @param string $base
184 * @param string $file
185 * @param string $type
186 * @param integer $lvl
187 * @param array $opts
188 *
189 * @return bool
190 */
191function search_media(&$data, $base, $file, $type, $lvl, $opts)
192{
193    //we do nothing with directories
194    if ($type == 'd') {
195        if (empty($opts['depth'])) return true; // recurse forever
196        $depth = substr_count($file,'/');
197        if ($depth >= $opts['depth']) return false; // depth reached
198        return true;
199    }
200
201    $info         = array();
202    $info['id']   = pathID($file, true);
203    if ($info['id'] != cleanID($info['id'])) {
204        if ($opts['showmsg'])
205            msg(hsc($info['id']).' is not a valid file name for DokuWiki - skipped',-1);
206        return false; // skip non-valid files
207    }
208
209    //check ACL for namespace (we have no ACL for mediafiles)
210    $info['perm'] = auth_quickaclcheck(getNS($info['id']).':*');
211    if (empty($opts['skipacl']) && $info['perm'] < AUTH_READ) {
212        return false;
213    }
214
215    //check pattern filter
216    if (!empty($opts['pattern']) && !@preg_match($opts['pattern'], $info['id'])) {
217        return false;
218    }
219
220    $info['file']     = \dokuwiki\Utf8\PhpString::basename($file);
221    $info['size']     = filesize($base.'/'.$file);
222    $info['mtime']    = filemtime($base.'/'.$file);
223    $info['writable'] = is_writable($base.'/'.$file);
224    if (preg_match("/\.(jpe?g|gif|png)$/", $file)) {
225        $info['isimg'] = true;
226        $info['meta']  = new JpegMeta($base.'/'.$file);
227    } else {
228        $info['isimg'] = false;
229    }
230    if (!empty($opts['hash'])) {
231        $info['hash'] = md5(io_readFile(mediaFN($info['id']), false));
232    }
233
234    $data[] = $info;
235
236    return false;
237}
238
239/**
240 * This function just lists documents (for RSS namespace export)
241 *
242 * @author  Andreas Gohr <andi@splitbrain.org>
243 *
244 * @param array $data
245 * @param string $base
246 * @param string $file
247 * @param string $type
248 * @param integer $lvl
249 * @param array $opts
250 *
251 * @return bool
252 */
253function search_list(&$data, $base, $file, $type, $lvl, $opts)
254{
255    //we do nothing with directories
256    if ($type == 'd') return false;
257    //only search txt files
258    if (substr($file, -4) == '.txt') {
259        //check ACL
260        $id = pathID($file);
261        if (auth_quickaclcheck($id) < AUTH_READ) {
262            return false;
263        }
264        $data[]['id'] = $id;
265    }
266    return false;
267}
268
269/**
270 * Quicksearch for searching matching pagenames
271 *
272 * $opts['query'] is the search query
273 *
274 * @author  Andreas Gohr <andi@splitbrain.org>
275 *
276 * @param array $data
277 * @param string $base
278 * @param string $file
279 * @param string $type
280 * @param integer $lvl
281 * @param array $opts
282 *
283 * @return bool
284 */
285function search_pagename(&$data, $base, $file, $type, $lvl, $opts)
286{
287    //we do nothing with directories
288    if ($type == 'd') return true;
289    //only search txt files
290    if (substr($file, -4) != '.txt') return true;
291
292    //simple stringmatching
293    if (!empty($opts['query'])) {
294        if (strpos($file, $opts['query']) !== false) {
295            //check ACL
296            $id = pathID($file);
297            if (auth_quickaclcheck($id) < AUTH_READ) {
298                return false;
299            }
300            $data[]['id'] = $id;
301        }
302    }
303    return true;
304}
305
306/**
307 * Just lists all documents
308 *
309 * $opts['depth']   recursion level, 0 for all
310 * $opts['hash']    do md5 sum of content?
311 * $opts['skipacl'] list everything regardless of ACL
312 *
313 * @author  Andreas Gohr <andi@splitbrain.org>
314 *
315 * @param array $data
316 * @param string $base
317 * @param string $file
318 * @param string $type
319 * @param integer $lvl
320 * @param array $opts
321 *
322 * @return bool
323 */
324function search_allpages(&$data, $base, $file, $type, $lvl, $opts)
325{
326    if (isset($opts['depth']) && $opts['depth']) {
327        $parts = explode('/',ltrim($file, '/'));
328        if (($type == 'd' && count($parts) >= $opts['depth'])
329            || ($type != 'd' && count($parts) > $opts['depth'])
330        ){
331            return false; // depth reached
332        }
333    }
334
335    //we do nothing with directories
336    if ($type == 'd') {
337        return true;
338    }
339
340    //only search txt files
341    if (substr($file, -4) != '.txt') return true;
342
343    $item = array();
344    $item['id']   = pathID($file);
345    if (empty($opts['skipacl']) && auth_quickaclcheck($item['id']) < AUTH_READ) {
346        return false;
347    }
348
349    $item['rev']   = filemtime($base.'/'.$file);
350    $item['mtime'] = $item['rev'];
351    $item['size']  = filesize($base.'/'.$file);
352    if (!empty($opts['hash'])) {
353        $item['hash'] = md5(trim(rawWiki($item['id'])));
354    }
355
356    $data[] = $item;
357    return true;
358}
359
360/* ------------- helper functions below -------------- */
361
362/**
363 * fulltext sort
364 *
365 * Callback sort function for use with usort to sort the data
366 * structure created by search_fulltext. Sorts descending by count
367 *
368 * @author  Andreas Gohr <andi@splitbrain.org>
369 *
370 * @param array $a
371 * @param array $b
372 *
373 * @return int
374 */
375function sort_search_fulltext($a, $b)
376{
377    if ($a['count'] > $b['count']) {
378        return -1;
379    } elseif ($a['count'] < $b['count']) {
380        return 1;
381    }else{
382        return Sort::strcmp($a['id'],$b['id']);
383    }
384}
385
386/**
387 * translates a document path to an ID
388 *
389 * @author  Andreas Gohr <andi@splitbrain.org>
390 * @todo    move to pageutils
391 *
392 * @param string $path
393 * @param bool $keeptxt
394 *
395 * @return mixed|string
396 */
397function pathID($path, $keeptxt=false)
398{
399    $id = utf8_decodeFN($path);
400    $id = str_replace('/', ':', $id);
401    if (!$keeptxt) $id = preg_replace('#\.txt$#', '', $id);
402    $id = trim($id, ':');
403    return $id;
404}
405
406
407/**
408 * This is a very universal callback for the search() function, replacing
409 * many of the former individual functions at the cost of a more complex
410 * setup.
411 *
412 * How the function behaves, depends on the options passed in the $opts
413 * array, where the following settings can be used.
414 *
415 * depth      int     recursion depth. 0 for unlimited                       (default: 0)
416 * keeptxt    bool    keep .txt extension for IDs                            (default: false)
417 * listfiles  bool    include files in listing                               (default: false)
418 * listdirs   bool    include namespaces in listing                          (default: false)
419 * pagesonly  bool    restrict files to pages                                (default: false)
420 * skipacl    bool    do not check for READ permission                       (default: false)
421 * sneakyacl  bool    don't recurse into nonreadable dirs                    (default: false)
422 * hash       bool    create MD5 hash for files                              (default: false)
423 * meta       bool    return file metadata                                   (default: false)
424 * filematch  string  match files against this regexp                        (default: '', so accept everything)
425 * idmatch    string  match full ID against this regexp                      (default: '', so accept everything)
426 * dirmatch   string  match directory against this regexp when adding        (default: '', so accept everything)
427 * nsmatch    string  match namespace against this regexp when adding        (default: '', so accept everything)
428 * recmatch   string  match directory against this regexp when recursing     (default: '', so accept everything)
429 * showmsg    bool    warn about non-ID files                                (default: false)
430 * showhidden bool    show hidden files(e.g. by hidepages config) too        (default: false)
431 * firsthead  bool    return first heading for pages                         (default: false)
432 *
433 * @param array &$data  - Reference to the result data structure
434 * @param string $base  - Base usually $conf['datadir']
435 * @param string $file  - current file or directory relative to $base
436 * @param string $type  - Type either 'd' for directory or 'f' for file
437 * @param int    $lvl   - Current recursion depht
438 * @param array  $opts  - option array as given to search()
439 * @return bool if this directory should be traversed (true) or not (false)
440 *              return value is ignored for files
441 *
442 * @author Andreas Gohr <gohr@cosmocode.de>
443 */
444function search_universal(&$data, $base, $file, $type, $lvl, $opts)
445{
446    $item   = array();
447    $return = true;
448
449    // get ID and check if it is a valid one
450    $item['id'] = pathID($file, ($type == 'd' || !empty($opts['keeptxt'])));
451    if ($item['id'] != cleanID($item['id'])){
452        if (!empty($opts['showmsg'])) {
453            msg(hsc($item['id']).' is not a valid file name for DokuWiki - skipped',-1);
454        }
455        return false; // skip non-valid files
456    }
457    $item['ns']  = getNS($item['id']);
458
459    if ($type == 'd') {
460        // decide if to recursion into this directory is wanted
461        if (empty($opts['depth'])) {
462            $return = true; // recurse forever
463        } else {
464            $depth = substr_count($file,'/');
465            if ($depth >= $opts['depth']) {
466                $return = false; // depth reached
467            } else {
468                $return = true;
469            }
470        }
471
472        if ($return) {
473            $match = empty($opts['recmatch']) || preg_match('/'.$opts['recmatch'].'/', $file);
474            if (!$match) {
475                return false; // doesn't match
476            }
477        }
478    }
479
480    // check ACL
481    if (empty($opts['skipacl'])) {
482        if ($type == 'd') {
483            $item['perm'] = auth_quickaclcheck($item['id'].':*');
484        } else {
485            $item['perm'] = auth_quickaclcheck($item['id']); //FIXME check namespace for media files
486        }
487    } else {
488        $item['perm'] = AUTH_DELETE;
489    }
490
491    // are we done here maybe?
492    if ($type == 'd') {
493        if (empty($opts['listdirs'])) return $return;
494        //neither list nor recurse forbidden items:
495        if (empty($opts['skipacl']) && !empty($opts['sneakyacl']) && $item['perm'] < AUTH_READ) return false;
496        if (!empty($opts['dirmatch']) && !preg_match('/'.$opts['dirmatch'].'/', $file)) return $return;
497        if (!empty($opts['nsmatch']) && !preg_match('/'.$opts['nsmatch'].'/', $item['ns'])) return $return;
498    } else {
499        if (empty($opts['listfiles'])) return $return;
500        if (empty($opts['skipacl']) && $item['perm'] < AUTH_READ) return $return;
501        if (!empty($opts['pagesonly']) && (substr($file, -4) != '.txt')) return $return;
502        if (empty($opts['showhidden']) && isHiddenPage($item['id'])) return $return;
503        if (!empty($opts['filematch']) && !preg_match('/'.$opts['filematch'].'/', $file)) return $return;
504        if (!empty($opts['idmatch']) && !preg_match('/'.$opts['idmatch'].'/', $item['id'])) return $return;
505    }
506
507    // still here? prepare the item
508    $item['type']  = $type;
509    $item['level'] = $lvl;
510    $item['open']  = $return;
511
512    if (!empty($opts['meta'])) {
513        $item['file']       = \dokuwiki\Utf8\PhpString::basename($file);
514        $item['size']       = filesize($base.'/'.$file);
515        $item['mtime']      = filemtime($base.'/'.$file);
516        $item['rev']        = $item['mtime'];
517        $item['writable']   = is_writable($base.'/'.$file);
518        $item['executable'] = is_executable($base.'/'.$file);
519    }
520
521    if ($type == 'f') {
522        if (!empty($opts['hash'])) $item['hash'] = md5(io_readFile($base.'/'.$file, false));
523        if (!empty($opts['firsthead'])) {
524            $item['title'] = p_get_first_heading($item['id'], METADATA_DONT_RENDER);
525        }
526    }
527
528    // finally add the item
529    $data[] = $item;
530    return $return;
531}
532
533//Setup VIM: ex: et ts=4 :
534