ILIAS Release_3_9_x_branch .rev 46835: classes/class.ilLinkChecker.php Source File

Go to the documentation of this file.
00001 <?php
00002 /*
00003         +-----------------------------------------------------------------------------+
00004         | ILIAS open source                                                           |
00005         +-----------------------------------------------------------------------------+
00006         | Copyright (c) 1998-2001 ILIAS open source, University of Cologne            |
00007         |                                                                             |
00008         | This program is free software; you can redistribute it and/or               |
00009         | modify it under the terms of the GNU General Public License                 |
00010         | as published by the Free Software Foundation; either version 2              |
00011         | of the License, or (at your option) any later version.                      |
00012         |                                                                             |
00013         | This program is distributed in the hope that it will be useful,             |
00014         | but WITHOUT ANY WARRANTY; without even the implied warranty of              |
00015         | MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the               |
00016         | GNU General Public License for more details.                                |
00017         |                                                                             |
00018         | You should have received a copy of the GNU General Public License           |
00019         | along with this program; if not, write to the Free Software                 |
00020         | Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA  02111-1307, USA. |
00021         +-----------------------------------------------------------------------------+
00022 */
00023 
00032 class ilLinkChecker
00033 {
00034         var $db = null;
00035         var $log_messages = array();
00036         var $invalid_links = array();
00037 
00038         var $validate_all = true;
00039         var $mail_status = false;
00040         var $page_id = 0;
00041 
00042 
00043         function ilLinkChecker(&$db,$a_validate_all = true)
00044         {
00045                 global $ilDB;
00046 
00047                 define('DEBUG',1);
00048                 define('SOCKET_TIMEOUT',5);
00049 
00050                 $this->db =& $db;
00051 
00052                 // SET GLOBAL DB HANDLER FOR STATIC METHODS OTHER CLASSES
00053                 $ilDB =& $db;
00054 
00055                 $this->validate_all = $a_validate_all;
00056         }
00057 
00058         function setCheckPeriod($a_period)
00059         {
00060                 $this->period = $a_period;
00061         }
00062         function getCheckPeriod()
00063         {
00064                 return $this->period;
00065         }
00066 
00067         function setMailStatus($a_status)
00068         {
00069                 $this->mail_status = (bool) $a_status;
00070         }
00071         function getMailStatus()
00072         {
00073                 return (bool) $this->mail_status;
00074         }
00075 
00076         function __setType($a_type)
00077         {
00078                 $this->type = $a_type;
00079         }
00080         function __getType()
00081         {
00082                 return $this->type;
00083         }
00084 
00085         function setObjId($a_page_id)
00086         {
00087                 return $this->page_id = $a_page_id;
00088         }
00089         function getObjId()
00090         {
00091                 return $this->page_id;
00092         }
00093 
00094         function getValidateAll()
00095         {
00096                 return $this->validate_all ? true : false;
00097         }
00098 
00099         function getLogMessages()
00100         {
00101                 return $this->log_messages ? $this->log_messages : array();
00102         }
00103 
00104         function getInvalidLinks()
00105         {
00106                 return $this->invalid_links ? $this->invalid_links : array();
00107         }
00108 
00109         function getInvalidLinksFromDB()
00110         {
00111                 global $ilDB;
00112                 
00113                 $query = "SELECT * FROM link_check ".
00114                         "WHERE obj_id = ".$ilDB->quote($this->getObjId())." ";
00115 
00116                 $res = $this->db->query($query);
00117                 while($row = $res->fetchRow(DB_FETCHMODE_OBJECT))
00118                 {
00119                         $invalid[] = array('page_id' => $row->page_id,
00120                                                            'url'         => $row->url);
00121                 }
00122 
00123                 return $invalid ? $invalid : array();
00124         }
00125 
00126         function getLastCheckTimestamp()
00127         {
00128                 global $ilDB;           
00129                 
00130                 if($this->getValidateAll())
00131                 {
00132                         $query = "SELECT MAX(last_check) as last_check FROM link_check ";
00133                 }
00134                 else
00135                 {
00136                         $query = "SELECT MAX(last_check) as last_check FROM link_check ".
00137                                 "WHERE obj_id = ".$ilDB->quote($this->getObjId())." ";
00138                 }
00139                 $row = $this->db->getRow($query,DB_FETCHMODE_OBJECT);
00140 
00141                 return $row->last_check ? $row->last_check : 0;
00142         }
00143 
00144         function checkWebResourceLinks()
00145         {
00146                 $pages = array();
00147 
00148                 $this->__setType('webr');
00149                 $this->__clearLogMessages();
00150                 $this->__clearInvalidLinks();
00151                 $this->__appendLogMessage('LinkChecker: Start checkLinks()');
00152 
00153                 if(count($invalid = $this->__validateLinks($this->__getWebResourceLinks())))
00154                 {
00155                         foreach($invalid as $invalid_item)
00156                         {
00157                                 $this->__appendLogMessage('LinkChecker: found invalid link: '.$invalid_item['complete']);
00158                                 $this->__appendInvalidLink($invalid_item);
00159                         }
00160                 }
00161                 
00162                 $this->__appendLogMessage('LinkChecker: End checkLinks()');
00163                 $this->__saveInDB();
00164                 
00165                 $this->__sendMail();
00166                 
00167                 return $this->getInvalidLinks();
00168         }
00169         
00170         function checkLinks()
00171         {
00172                 global $ilDB;
00173                 
00174                 $pages = array();
00175 
00176                 $this->__setType('lm');
00177                 $this->__clearLogMessages();
00178                 $this->__clearInvalidLinks();
00179                 $this->__appendLogMessage('LinkChecker: Start checkLinks()');
00180 
00181                 if(!$this->getValidateAll() and !$this->getObjId())
00182                 {
00183                         echo "ilLinkChecker::checkLinks() No Page id given";
00184 
00185                         return false;
00186                 }
00187                 elseif(!$this->getValidateAll() and $this->getObjId())
00188                 {
00189                         $query = "SELECT * FROM page_object ".
00190                                 "WHERE parent_id = ".$ilDB->quote($this->getObjId())." ".
00191                                 "AND parent_type = 'lm'";
00192 
00193                         $res = $this->db->query($query);
00194                         while($row = $res->fetchRow(DB_FETCHMODE_OBJECT))
00195                         {
00196                                 $pages[] = array('page_id' => $row->page_id,
00197                                                                  'content' => $row->content,
00198                                                                  'type'  => $row->parent_type);
00199                         }
00200                 }
00201                 elseif($this->getValidateAll())
00202                 {
00203                         $query = "SELECT * FROM page_object ".
00204                                 "WHERE parent_type = 'lm'";
00205 
00206                         $res = $this->db->query($query);
00207                         while($row = $res->fetchRow(DB_FETCHMODE_OBJECT))
00208                         {
00209                                 $pages[] = array('page_id' => $row->page_id,
00210                                                                  'content' => $row->content,
00211                                                                  'type'  => $row->parent_type);
00212                         }
00213                 }
00214 
00215                 // VALIDATE
00216                 foreach($pages as $page)
00217                 {
00218                         if(count($invalid = $this->__validateLinks($this->__getLinks($page))))
00219                         {
00220                                 foreach($invalid as $invalid_item)
00221                                 {
00222                                         $this->__appendLogMessage('LinkChecker: found invalid link: '.$invalid_item['complete']);
00223                                         $this->__appendInvalidLink($invalid_item);
00224                                 }
00225                         }
00226                 }
00227                 
00228                 $this->__appendLogMessage('LinkChecker: End checkLinks()');
00229                 $this->__saveInDB();
00230 
00231                 $this->__sendMail();
00232 
00233                 return $this->getInvalidLinks();
00234         }
00235 
00236         function checkPear()
00237         {
00238                 if(!@include_once('HTTP/Request.php'))
00239                 {
00240                         return false;
00241                 }
00242                 return true;
00243         }
00244                 
00245 
00246         // PRIVATE
00247         function __txt($language,$key,$module = 'common')
00248         {
00249                 global $ilDB;
00250                 
00251                 $query = "SELECT value FROM lng_data ".
00252                         "WHERE module = ".$ilDB->quote($module)." ".
00253                         "AND identifier = ".$ilDB->quote($key)." ".
00254                         "AND lang_key = ".$ilDB->quote($language)."";
00255 
00256                 $res = $this->db->query($query);
00257                 while($row = $res->fetchRow(DB_FETCHMODE_OBJECT))
00258                 {
00259                         $value = $row->value;
00260                 }
00261                 if(!$value)
00262                 {
00263                         $query = "SELECT value FROM lng_data ".
00264                                 "WHERE module = ".$ilDB->quote($module)." ".
00265                                 "AND identifier = ".$ilDB->quote($key)." ".
00266                                 "AND lang_key = 'en'";
00267                         
00268                         $res = $this->db->query($query);
00269                         while($row = $res->fetchRow(DB_FETCHMODE_OBJECT))
00270                         {
00271                                 $value = $row->value;
00272                         }
00273                 }
00274                 return $value ? $value : '-'.$key.'-';
00275         }
00276 
00277         function __fetchUserData($a_usr_id)
00278         {
00279                 global $ilDB;
00280                 
00281                 $query = "SELECT email FROM usr_data WHERE usr_id = ".$ilDB->quote($a_usr_id)."";
00282 
00283                 $row = $this->db->getRow($query,DB_FETCHMODE_OBJECT);
00284 
00285                 $data['email'] = $row->email;
00286 
00287                 $query = "SELECT * FROM usr_pref ".
00288                         "WHERE usr_id = ".$ilDB->quote($a_usr_id)." ".
00289                         "AND keyword = 'language'";
00290 
00291                 $row = $this->db->getRow($query,DB_FETCHMODE_OBJECT);
00292 
00293                 $data['lang'] = $row->value;
00294 
00295                 return $data;
00296         }
00297 
00298         function __getTitle($a_lm_obj_id)
00299         {
00300                 global $ilDB;
00301                 
00302                 $query = "SELECT title FROM object_data ".
00303                         "WHERE obj_id = ".$ilDB->quote($a_lm_obj_id)." ";
00304 
00305                 $row = $this->db->getRow($query,DB_FETCHMODE_OBJECT);
00306 
00307                 return $row->title;
00308         }
00309 
00310         function __sendMail()
00311         {
00312                 global $ilUser;
00313 
00314 
00315                 if(!count($notify = $this->__getNotifyLinks()))
00316                 {
00317                         // Nothing to do
00318                         return true;
00319                 }
00320                 if(!$this->getMailStatus())
00321                 {
00322                         return true;
00323                 }
00324 
00325                 include_once './classes/class.ilLinkCheckNotify.php';
00326                 
00327                 foreach(ilLinkCheckNotify::_getAllNotifiers($this->db) as $usr_id => $obj_ids)
00328                 {
00329                         if(!is_object($tmp_user =& ilObjectFactory::getInstanceByObjId($usr_id,false)))
00330                         {
00331                                 $this->__appendLogMessage('LinkChecker: Cannot find user with id: '.$usr_id);
00332                                 continue;
00333                         }
00334 
00335                         $counter = 0;
00336                         foreach($obj_ids as $obj_id)
00337                         {
00338                                 if(!isset($notify[$obj_id]))
00339                                 {
00340                                         continue;
00341                                 }
00342                                 ++$counter;
00343 
00344                                 switch($this->__getType())
00345                                 {
00346                                         case 'webr':
00347                                                 $body .= $this->__txt($tmp_user->getLanguage(),'obj_webr');
00348                                                 break;
00349 
00350                                         case 'lm':
00351                                         default:
00352                                                 $body .= $this->__txt($tmp_user->getLanguage(),'lo');
00353                                                 break;
00354                                 }
00355                                                 
00356                                 $body .= ': ';
00357                                 $body .= $this->__getTitle($obj_id)."\r\n";
00358 
00359                                 // Print all invalid
00360                                 foreach($notify[$obj_id] as $data)
00361                                 {
00362                                         $body .= $data['url']."\r\n";
00363                                 }
00364                                 $body .= "\r\n";
00365                         }
00366                         if($counter)
00367                         {
00368                                 include_once "Services/Mail/classes/class.ilFormatMail.php";
00369                                 
00370                                 $umail = new ilFormatMail($tmp_user->getId());
00371                                 $subject = $this->__txt($tmp_user->getLanguage(),'link_check_subject');
00372 
00373                                 $umail->sendMail($tmp_user->getLogin(),"","",$subject,$body,array(),array("normal"));
00374                                 $this->__appendLogMessage('LinkChecker: Sent mail to '.$tmp_user->getEmail());
00375                         }
00376 
00377                 }
00378                                 
00379 
00380         }
00381 
00382         function __getNotifyLinks()
00383         {
00384                 return $this->notify ? $this->notify : array();
00385         }
00386 
00387 
00388         function __clearInvalidLinks()
00389         {
00390                 $this->invalid_links = array();
00391         }
00392         function __appendInvalidLink($a_link)
00393         {
00394                 $this->invalid_links[] = $a_link;
00395         }
00396                                         
00397 
00398         function __appendLogMessage($a_string)
00399         {
00400                 $this->log_messages[] = $a_string;
00401         }
00402         function __clearLogMessages()
00403         {
00404                 return $this->log_messages = array();
00405         }
00406 
00407         function __getLinks($a_page)
00408         {
00409                 $matches = array();
00410 
00411                 $pattern_complete = '/<ExtLink Href="([^"]*)">/';
00412                 if(preg_match_all($pattern_complete,$a_page['content'],$matches))
00413                 {
00414                         for($i = 0;$i < count($matches[0]); ++$i)
00415                         {
00416                                 $url_data = @parse_url($matches[1][$i]);
00417                                 // continue if mailto link
00418                                 if($url_data['scheme'] == 'mailto')
00419                                 {
00420                                         continue;
00421                                 }
00422                                 
00423                                 // PUH, HTTP_REQUEST needs a beginning http://
00424                                 if(!$url_data['scheme'])
00425                                 {
00426                                         $matches[1][$i] = 'http://'.$matches[1][$i];
00427                                 }
00428 
00429                                 $lm_id = $this->__getObjIdByPageId($a_page['page_id']);
00430                                 $link[] = array('page_id'  => $a_page['page_id'],
00431                                                                 'obj_id'   => $lm_id,
00432                                                                 'type'     => $a_page['type'],
00433                                                                 'complete' => $matches[1][$i],
00434                                                                 'scheme'   => isset($url_data['scheme']) ? $url_data['scheme'] : 'http',
00435                                                                 'host'     => isset($url_data['host']) ? $url_data['host'] : $url_data['path']);
00436                         }
00437                 }
00438 
00439                 return $link ? $link : array();
00440         }
00441 
00442         function __getWebResourceLinks()
00443         {
00444                 include_once 'Modules/WebResource/classes/class.ilLinkResourceItems.php';
00445 
00446                 $link_res_obj = new ilLinkResourceItems($this->getObjId());
00447 
00448                 foreach($check_links = $link_res_obj->getCheckItems($this->getCheckPeriod()) as $item_data)
00449                 {
00450                                 $url_data = @parse_url($item_data['target']);
00451                                 
00452                                 // PUH, HTTP_REQUEST needs a beginning http://
00453                                 if(!$url_data['scheme'])
00454                                 {
00455                                         $item_data['target'] = 'http://'.$item_data['target'];
00456                                 }
00457 
00458                                 $link[] = array('page_id'  => $item_data['link_id'],
00459                                                                 'obj_id'   => $this->getObjId(),
00460                                                                 'type'     => 'webr',
00461                                                                 'complete' => $item_data['target'],
00462                                                                 'scheme'   => isset($url_data['scheme']) ? $url_data['scheme'] : 'http',
00463                                                                 'host'     => isset($url_data['host']) ? $url_data['host'] : $url_data['path']);
00464                 }
00465                 return $link ? $link : array();
00466         }                       
00467 
00468                 
00469 
00470         function __validateLinks($a_links)
00471         {
00472                 if(!@include_once('HTTP/Request.php'))
00473                 {
00474                         $this->__appendLogMessage('LinkChecker: Pear HTTP_Request is not installed. Aborting');
00475 
00476                         return array();
00477                 }
00478 
00479                 foreach($a_links as $link)
00480                 {
00481                         if(gethostbyname($link['host']) == $link['host'])
00482                         {
00483                                 $invalid[] = $link;
00484                                 continue;
00485                         }
00486 
00487                         if($link['scheme'] !== 'http' and $link['scheme'] !== 'https')
00488                         {
00489                                 continue;
00490                         }
00491                         $req =& new HTTP_Request($link['complete']);
00492                         $req->sendRequest();
00493 
00494                         switch($req->getResponseCode())
00495                         {
00496                                 // EVERYTHING OK
00497                                 case '200':
00498                                         // In the moment 301 will be handled as ok
00499                                 case '301':
00500                                 case '302':
00501                                         break;
00502 
00503                                 default:
00504                                         $link['http_status_code'] = $req->getResponseCode();
00505                                         $invalid[] = $link;
00506                                         break;
00507                         }
00508                 }
00509                 return $invalid ? $invalid : array();
00510         }
00511 
00512         function __getObjIdByPageId($a_page_id)
00513         {
00514                 $query = "SELECT lm_id FROM lm_data ".
00515                         "WHERE obj_id = '".$a_page_id."'";
00516 
00517                 $row = $this->db->getRow($query,DB_FETCHMODE_OBJECT);
00518 
00519                 return $row->lm_id ? $row->lm_id : 0;
00520         }
00521 
00522         function __isInvalid($a_page_id, $a_url)
00523         {
00524                 foreach($this->getInvalidLinks() as $link)
00525                 {
00526                         if($link['page_id'] == $a_page_id and
00527                            substr($link['complete'],0,255) == $a_url)
00528                         {
00529                                 return true;
00530                         }
00531                 }
00532                 return false;
00533         }
00534 
00535         function __saveInDB()
00536         {
00537                 global $ilDB;
00538                 
00539                 if($this->getMailStatus())
00540                 {
00541                         $this->__checkNotify();
00542                 }
00543                 $this->__clearDBData();
00544 
00545 
00546                 foreach($this->getInvalidLinks() as $link)
00547                 {
00548 
00549                         $query = "INSERT INTO link_check ".
00550                                 "SET page_id = ".$ilDB->quote($link['page_id']).", ".
00551                                 "obj_id = ".$ilDB->quote($link['obj_id']).", ".
00552                                 "url = ".$ilDB->quote((substr($link['complete'],0,255))).", ".
00553                                 "parent_type = ".$ilDB->quote(($link['type'])).", ".
00554                                 "http_status_code = ".$ilDB->quote(($link['http_status_code'])).", ".
00555                                 "last_check = '".time()."'";
00556 
00557 
00558                         $res = $this->db->query($query);
00559                 }
00560 
00561                 // delete old values
00562                 
00563         }
00564 
00565         function __checkNotify()
00566         {
00567                 global $ilDB;
00568                 
00569                 foreach($this->getInvalidLinks() as $link)
00570                 {
00571                         $query = "SELECT * FROM link_check ".
00572                                 "WHERE page_id = ".$ilDB->quote($link['page_id'])." ".
00573                                 "AND url = ".$ilDB->quote((substr($link['complete'],0,255)))." ";
00574                         $res = $this->db->query($query);
00575                         
00576                         if(!$res->numRows())
00577                         {
00578                                 $this->notify["$link[obj_id]"][] = array('page_id' => $link['page_id'],
00579                                                                                                                  'url'     => $link['complete']);
00580                         }
00581                 }
00582         }
00583 
00584 
00585         function __clearDBData()
00586         {
00587                 global $ilDB;
00588                 
00589                 if($this->getValidateAll())
00590                 {
00591                         $query = "DELETE FROM link_check";
00592                 }
00593                 else
00594                 {
00595                         $query = "DELETE FROM link_check ".
00596                                 "WHERE obj_id = ".$ilDB->quote($this->getObjId())." ";
00597                 }
00598 
00599                 $this->db->query($query);
00600 
00601                 return true;
00602         }
00603 }
00604 ?>