[Sommaire] [Imprimer]

   1  <?php
   2  /*****************************************************
   3  * This file is part of Agora, web based content management system.
   4  *
   5  * Agora is free software; you can redistribute it and/or modify
   6  * it under the terms of the GNU General Public License as published by
   7  * the Free Software Foundation; version 2 of the License.
   8  *
   9  * Agora is distributed in the hope that it will be useful,
  10  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  11  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  12  * GNU General Public License for more details (file "COPYING").
  13  *
  14  * Copyright © Arnaud Martin, Antoine Pitrou et Philippe Rivière.
  15  * List of authors detailed in "copyright_fr.html" file.
  16  * E-mail : agora@sig.premier-ministre.gouv.fr
  17  * Web site : http://www.agora.gouv.fr
  18  *****************************************************/
  19  // Base class for Indexer
  20  // $Id$
  21  
  22  require_once dirname(__FILE__). "/../bd/inc_index_dico_factory.php";
  23  
  24  require_once dirname(__FILE__). "/Indexer_Utils.php";
  25  
  26  class Indexer {
  27  
  28      // {{{ properties
  29  
  30      /**
  31       * type of Indexation.
  32       * @var     String
  33       * @access  private
  34       */
  35      var $_type;
  36  
  37      /**
  38       * $mots contains (hash, mot).
  39       * @var     Array
  40       * @access  private
  41       */
  42      var $_mots;
  43  
  44      /**
  45       * $index contains (hash, poids).
  46       * @var     Array
  47       * @access  private
  48       */
  49      var $_index;
  50  
  51      /**
  52       * $poids contains the definition of weight.
  53       * @var     Array
  54       * @access  private
  55       */
  56      var $_poids;
  57  
  58      /**
  59       * $full is a flag to index more data.
  60       * @var     Array
  61       * @access  private
  62       */
  63      var $_full = true;
  64  
  65      // }}}
  66  
  67      // {{{ getFull()
  68  
  69  	function getFull () {
  70          return $this->_full;
  71      }
  72  
  73      // }}}
  74  
  75      // {{{ setFull($full)
  76  
  77  	function setFull ($full) {
  78          $this->_full = $full;
  79      }
  80  
  81      // }}}
  82  
  83      // {{{ factory($typeIndexer)
  84  
  85      /**
  86       * Indexer Factory.
  87       *
  88       * @access public
  89       */
  90      function &factory ($typeIndexer) {
  91          if (!file_exists(dirname(
  92                               __FILE__). '/' . $GLOBALS['type_indexation']. '/Indexer_' . $typeIndexer . '_' . $GLOBALS['type_indexation']. '.php'))
  93              {
  94              include_once (dirname(__FILE__). '/common/Indexer_' . $typeIndexer . '_common.php');
  95              $classname = 'Indexer_' . $typeIndexer . '_common';
  96          }
  97          else {
  98              include_once (dirname(
  99                                __FILE__). '/' . $GLOBALS['type_indexation']. '/Indexer_' . $typeIndexer . '_' . $GLOBALS['type_indexation']. '.php');
 100              $classname = 'Indexer_' . $typeIndexer . "_" . $GLOBALS['type_indexation'];
 101          }
 102  
 103          if (!class_exists($classname)) {
 104              return PEAR::raiseError("Cannot instanciate class $classname", null, null, null, null, null, false);
 105          }
 106  
 107          $obj = &new $classname();
 108  
 109          return $obj;
 110      }
 111  
 112      // }}}
 113  
 114      // {{{ constructor
 115  
 116      /**
 117       * Indexer constructor.
 118       *
 119       * @access public
 120       */
 121  
 122  	function Indexer () { }
 123  
 124      // }}}
 125  
 126      // {{{ index($id_objet, $forcer_reset = true, $full = true)
 127  
 128      /**
 129       * index an object
 130       * @param $id_objet
 131       * @param $forcer_reset
 132       * @param $full
 133       * @access public
 134       */
 135  
 136  	function index ($id_objet, $forcer_reset = true) {
 137          if (!$id_objet OR (!$forcer_reset AND $this->alreadyIndex($id_objet)))
 138              return;
 139  
 140          if (function_exists(spip_log)) {
 141              spip_log ("indexation " . $this->_type . " $id_objet");
 142          }
 143  
 144          $this->_index = '';
 145  
 146          $indexDicoMetier = &recuperer_instance_index_dico();
 147  
 148          $this->_indexData($id_objet);
 149  
 150          if ($this->_index) {
 151              $indexDicoMetier->addDico($this->_mots);
 152  
 153              reset ($this->_index);
 154  
 155              $this->_createIndex($id_objet);
 156          }
 157      }
 158  
 159      // }}}
 160  
 161      // {{{ _indexer_chaine($texte, $val = 1, $min_long = 3)
 162  
 163  	function _indexer_chaine ($texte, $val = 1, $min_long = 3) {
 164          if ($val == 0) { // pas d'indexation pour un poids de zéro
 165              return;
 166          }
 167  
 168          // Nettoyer les tags, entites HTML, signes diacritiques...
 169          $texte = ' ' . ereg_replace("<[^>]*>", " ", $texte). ' ';
 170          $texte = Indexer_Utils::nettoyer_chaine_indexation($texte);
 171  
 172          // Nettoyer les caracteres non-alphanumeriques
 173          $regs = Indexer_Utils::separateurs_indexation();
 174          $texte = strtr($texte, $regs, ereg_replace('.', ' ', $regs));
 175  
 176          // Cas particulier : sigles d'au moins deux lettres
 177          $texte = ereg_replace(" ([A-Z][0-9A-Z]{1," . ($min_long - 1). "}) ", ' \\1___ ', $texte);
 178          $texte = strtolower($texte);
 179  
 180          // Separer les mots
 181          $table = Indexer_Utils::spip_split(" +", $texte);
 182  
 183          while (list(, $mot) = each($table)) {
 184              if (strlen($mot) > $min_long) {
 185                  $h = substr(md5($mot), 0, 16);
 186                  $this->_index[$h] += $val;
 187                  $this->_mots[] = array("hash" => "0x$h", "dico" => $mot);
 188              }
 189          }
 190      }
 191  
 192      // }}}
 193  
 194      // {{{ alreadyIndex($id_objet)
 195  
 196      /**
 197       * check id an object is indexed.
 198       * @param $id_objet
 199       * @access public
 200       */
 201  
 202  	function alreadyIndex ($id_objet) { }
 203  
 204      // {{{ _indexData($id_objet)
 205  
 206      /**
 207       * this method initialses this attributes 'mots' and 'index' 
 208       * @param $id_objet
 209       * @access private
 210       */
 211  
 212  	function _indexData ($id_object) { }
 213  
 214      // }}}
 215  
 216      // {{{ _createIndex($id_objet)
 217  
 218      /**
 219       * this method create indexation data 
 220       * @param $id_objet
 221       * @access private
 222       */
 223  
 224  	function _createIndex ($id_object) { }
 225  
 226  // }}}
 227  
 228  }
 229  ?>