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