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 $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 $opts 105 * 106 * @return bool 107 */ 108function search_qsearch(&$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 $opts 130 * 131 * @return bool 132 */ 133function search_index(&$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 $opts 159 * 160 * @return bool 161 */ 162function search_namespaces(&$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 $opts 185 * 186 * @return bool 187 */ 188function search_media(&$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($opts['showmsg']) 202 msg(hsc($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 * This function just lists documents (for RSS namespace export) 238 * 239 * @author Andreas Gohr <andi@splitbrain.org> 240 * 241 * @param array $data 242 * @param string $base 243 * @param string $file 244 * @param string $type 245 * @param integer $lvl 246 * @param array $opts 247 * 248 * @return bool 249 */ 250function search_list(&$data,$base,$file,$type,$lvl,$opts){ 251 //we do nothing with directories 252 if($type == 'd') return false; 253 //only search txt files 254 if(substr($file,-4) == '.txt'){ 255 //check ACL 256 $id = pathID($file); 257 if(auth_quickaclcheck($id) < AUTH_READ){ 258 return false; 259 } 260 $data[]['id'] = $id; 261 } 262 return false; 263} 264 265/** 266 * Quicksearch for searching matching pagenames 267 * 268 * $opts['query'] is the search query 269 * 270 * @author Andreas Gohr <andi@splitbrain.org> 271 * 272 * @param array $data 273 * @param string $base 274 * @param string $file 275 * @param string $type 276 * @param integer $lvl 277 * @param array $opts 278 * 279 * @return bool 280 */ 281function search_pagename(&$data,$base,$file,$type,$lvl,$opts){ 282 //we do nothing with directories 283 if($type == 'd') return true; 284 //only search txt files 285 if(substr($file,-4) != '.txt') return true; 286 287 //simple stringmatching 288 if (!empty($opts['query'])){ 289 if(strpos($file,$opts['query']) !== false){ 290 //check ACL 291 $id = pathID($file); 292 if(auth_quickaclcheck($id) < AUTH_READ){ 293 return false; 294 } 295 $data[]['id'] = $id; 296 } 297 } 298 return true; 299} 300 301/** 302 * Just lists all documents 303 * 304 * $opts['depth'] recursion level, 0 for all 305 * $opts['hash'] do md5 sum of content? 306 * $opts['skipacl'] list everything regardless of ACL 307 * 308 * @author Andreas Gohr <andi@splitbrain.org> 309 * 310 * @param array $data 311 * @param string $base 312 * @param string $file 313 * @param string $type 314 * @param integer $lvl 315 * @param array $opts 316 * 317 * @return bool 318 */ 319function search_allpages(&$data,$base,$file,$type,$lvl,$opts){ 320 if(isset($opts['depth']) && $opts['depth']){ 321 $parts = explode('/',ltrim($file,'/')); 322 if(($type == 'd' && count($parts) >= $opts['depth']) 323 || ($type != 'd' && count($parts) > $opts['depth'])){ 324 return false; // depth reached 325 } 326 } 327 328 //we do nothing with directories 329 if($type == 'd'){ 330 return true; 331 } 332 333 //only search txt files 334 if(substr($file,-4) != '.txt') return true; 335 336 $item = array(); 337 $item['id'] = pathID($file); 338 if(empty($opts['skipacl']) && auth_quickaclcheck($item['id']) < AUTH_READ){ 339 return false; 340 } 341 342 $item['rev'] = filemtime($base.'/'.$file); 343 $item['mtime'] = $item['rev']; 344 $item['size'] = filesize($base.'/'.$file); 345 if(!empty($opts['hash'])){ 346 $item['hash'] = md5(trim(rawWiki($item['id']))); 347 } 348 349 $data[] = $item; 350 return true; 351} 352 353/* ------------- helper functions below -------------- */ 354 355/** 356 * fulltext sort 357 * 358 * Callback sort function for use with usort to sort the data 359 * structure created by search_fulltext. Sorts descending by count 360 * 361 * @author Andreas Gohr <andi@splitbrain.org> 362 * 363 * @param array $a 364 * @param array $b 365 * 366 * @return int 367 */ 368function sort_search_fulltext($a,$b){ 369 if($a['count'] > $b['count']){ 370 return -1; 371 }elseif($a['count'] < $b['count']){ 372 return 1; 373 }else{ 374 return Sort::strcmp($a['id'],$b['id']); 375 } 376} 377 378/** 379 * translates a document path to an ID 380 * 381 * @author Andreas Gohr <andi@splitbrain.org> 382 * @todo move to pageutils 383 * 384 * @param string $path 385 * @param bool $keeptxt 386 * 387 * @return mixed|string 388 */ 389function pathID($path,$keeptxt=false){ 390 $id = utf8_decodeFN($path); 391 $id = str_replace('/',':',$id); 392 if(!$keeptxt) $id = preg_replace('#\.txt$#','',$id); 393 $id = trim($id, ':'); 394 return $id; 395} 396 397 398/** 399 * This is a very universal callback for the search() function, replacing 400 * many of the former individual functions at the cost of a more complex 401 * setup. 402 * 403 * How the function behaves, depends on the options passed in the $opts 404 * array, where the following settings can be used. 405 * 406 * depth int recursion depth. 0 for unlimited (default: 0) 407 * keeptxt bool keep .txt extension for IDs (default: false) 408 * listfiles bool include files in listing (default: false) 409 * listdirs bool include namespaces in listing (default: false) 410 * pagesonly bool restrict files to pages (default: false) 411 * skipacl bool do not check for READ permission (default: false) 412 * sneakyacl bool don't recurse into nonreadable dirs (default: false) 413 * hash bool create MD5 hash for files (default: false) 414 * meta bool return file metadata (default: false) 415 * filematch string match files against this regexp (default: '', so accept everything) 416 * idmatch string match full ID against this regexp (default: '', so accept everything) 417 * dirmatch string match directory against this regexp when adding (default: '', so accept everything) 418 * nsmatch string match namespace against this regexp when adding (default: '', so accept everything) 419 * recmatch string match directory against this regexp when recursing (default: '', so accept everything) 420 * showmsg bool warn about non-ID files (default: false) 421 * showhidden bool show hidden files(e.g. by hidepages config) too (default: false) 422 * firsthead bool return first heading for pages (default: false) 423 * 424 * @param array &$data - Reference to the result data structure 425 * @param string $base - Base usually $conf['datadir'] 426 * @param string $file - current file or directory relative to $base 427 * @param string $type - Type either 'd' for directory or 'f' for file 428 * @param int $lvl - Current recursion depht 429 * @param array $opts - option array as given to search() 430 * @return bool if this directory should be traversed (true) or not (false) 431 * return value is ignored for files 432 * 433 * @author Andreas Gohr <gohr@cosmocode.de> 434 */ 435function search_universal(&$data,$base,$file,$type,$lvl,$opts){ 436 $item = array(); 437 $return = true; 438 439 // get ID and check if it is a valid one 440 $item['id'] = pathID($file,($type == 'd' || !empty($opts['keeptxt']))); 441 if($item['id'] != cleanID($item['id'])){ 442 if(!empty($opts['showmsg'])){ 443 msg(hsc($item['id']).' is not a valid file name for DokuWiki - skipped',-1); 444 } 445 return false; // skip non-valid files 446 } 447 $item['ns'] = getNS($item['id']); 448 449 if($type == 'd') { 450 // decide if to recursion into this directory is wanted 451 if(empty($opts['depth'])){ 452 $return = true; // recurse forever 453 }else{ 454 $depth = substr_count($file,'/'); 455 if($depth >= $opts['depth']){ 456 $return = false; // depth reached 457 }else{ 458 $return = true; 459 } 460 } 461 462 if ($return) { 463 $match = empty($opts['recmatch']) || preg_match('/'.$opts['recmatch'].'/',$file); 464 if (!$match) { 465 return false; // doesn't match 466 } 467 } 468 } 469 470 // check ACL 471 if(empty($opts['skipacl'])){ 472 if($type == 'd'){ 473 $item['perm'] = auth_quickaclcheck($item['id'].':*'); 474 }else{ 475 $item['perm'] = auth_quickaclcheck($item['id']); //FIXME check namespace for media files 476 } 477 }else{ 478 $item['perm'] = AUTH_DELETE; 479 } 480 481 // are we done here maybe? 482 if($type == 'd'){ 483 if(empty($opts['listdirs'])) return $return; 484 //neither list nor recurse forbidden items: 485 if(empty($opts['skipacl']) && !empty($opts['sneakyacl']) && $item['perm'] < AUTH_READ) return false; 486 if(!empty($opts['dirmatch']) && !preg_match('/'.$opts['dirmatch'].'/',$file)) return $return; 487 if(!empty($opts['nsmatch']) && !preg_match('/'.$opts['nsmatch'].'/',$item['ns'])) return $return; 488 }else{ 489 if(empty($opts['listfiles'])) return $return; 490 if(empty($opts['skipacl']) && $item['perm'] < AUTH_READ) return $return; 491 if(!empty($opts['pagesonly']) && (substr($file,-4) != '.txt')) return $return; 492 if(empty($opts['showhidden']) && isHiddenPage($item['id'])) return $return; 493 if(!empty($opts['filematch']) && !preg_match('/'.$opts['filematch'].'/',$file)) return $return; 494 if(!empty($opts['idmatch']) && !preg_match('/'.$opts['idmatch'].'/',$item['id'])) return $return; 495 } 496 497 // still here? prepare the item 498 $item['type'] = $type; 499 $item['level'] = $lvl; 500 $item['open'] = $return; 501 502 if(!empty($opts['meta'])){ 503 $item['file'] = \dokuwiki\Utf8\PhpString::basename($file); 504 $item['size'] = filesize($base.'/'.$file); 505 $item['mtime'] = filemtime($base.'/'.$file); 506 $item['rev'] = $item['mtime']; 507 $item['writable'] = is_writable($base.'/'.$file); 508 $item['executable'] = is_executable($base.'/'.$file); 509 } 510 511 if($type == 'f'){ 512 if(!empty($opts['hash'])) $item['hash'] = md5(io_readFile($base.'/'.$file,false)); 513 if(!empty($opts['firsthead'])) $item['title'] = p_get_first_heading($item['id'],METADATA_DONT_RENDER); 514 } 515 516 // finally add the item 517 $data[] = $item; 518 return $return; 519} 520 521//Setup VIM: ex: et ts=4 : 522