CRM/Dedupe/BAO/Rule.php

   1 <?php
   2 /*
   3  +--------------------------------------------------------------------+
   4  | CiviCRM version 4.6                                                |
   5  +--------------------------------------------------------------------+
   6  | Copyright CiviCRM LLC (c) 2004-2014                                |
   7  +--------------------------------------------------------------------+
   8  | This file is a part of CiviCRM.                                    |
   9  |                                                                    |
  10  | CiviCRM is free software; you can copy, modify, and distribute it  |
  11  | under the terms of the GNU Affero General Public License           |
  12  | Version 3, 19 November 2007 and the CiviCRM Licensing Exception.   |
  13  |                                                                    |
  14  | CiviCRM is distributed in the hope that it will be useful, but     |
  15  | WITHOUT ANY WARRANTY; without even the implied warranty of         |
  16  | MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.               |
  17  | See the GNU Affero General Public License for more details.        |
  18  |                                                                    |
  19  | You should have received a copy of the GNU Affero General Public   |
  20  | License and the CiviCRM Licensing Exception along                  |
  21  | with this program; if not, contact CiviCRM LLC                     |
  22  | at info[AT]civicrm[DOT]org. If you have questions about the        |
  23  | GNU Affero General Public License or the licensing of CiviCRM,     |
  24  | see the CiviCRM license FAQ at http://civicrm.org/licensing        |
  25  +--------------------------------------------------------------------+
  26 */
  27
  28 /**
  29  *
  30  * @package CRM
  31  * @copyright CiviCRM LLC (c) 2004-2014
  32  * $Id$
  33  *
  34  */
  35
  36 /**
  37  * The CiviCRM duplicate discovery engine is based on an
  38  * algorithm designed by David Strauss <david@fourkitchens.com>.
  39  */
  40 class CRM_Dedupe_BAO_Rule extends CRM_Dedupe_DAO_Rule {
  41
  42   /**
  43    * Ids of the contacts to limit the SQL queries (whole-database queries otherwise)
  44    */
  45   var $contactIds = array();
  46
  47   /**
  48    * Params to dedupe against (queries against the whole contact set otherwise)
  49    */
  50   var $params = array();
  51
  52   /**
  53    * Return the SQL query for the given rule - either for finding matching
  54    * pairs of contacts, or for matching against the $params variable (if set).
  55    *
  56    * @return string
  57    *   SQL query performing the search
  58    */
  59   public function sql() {
  60     if ($this->params &&
  61       (!array_key_exists($this->rule_table, $this->params) ||
  62         !array_key_exists($this->rule_field, $this->params[$this->rule_table])
  63       )
  64     ) {
  65       // if params is present and doesn't have an entry for a field, don't construct the clause.
  66       return NULL;
  67     }
  68
  69     // we need to initialise WHERE, ON and USING here, as some table types
  70     // extend them; $where is an array of required conditions, $on and
  71     // $using are arrays of required field matchings (for substring and
  72     // full matches, respectively)
  73     $where = array();
  74     $on = array("SUBSTR(t1.{$this->rule_field}, 1, {$this->rule_length}) = SUBSTR(t2.{$this->rule_field}, 1, {$this->rule_length})");
  75     $using = array($this->rule_field);
  76
  77     switch ($this->rule_table) {
  78       case 'civicrm_contact':
  79         $id = 'id';
  80         //we should restrict by contact type in the first step
  81         $sql = "SELECT contact_type FROM civicrm_dedupe_rule_group WHERE id = {$this->dedupe_rule_group_id};";
  82         $ct = CRM_Core_DAO::singleValueQuery($sql);
  83         if ($this->params) {
  84           $where[] = "t1.contact_type = '{$ct}'";
  85         }
  86         else {
  87           $where[] = "t1.contact_type = '{$ct}'";
  88           $where[] = "t2.contact_type = '{$ct}'";
  89         }
  90         break;
  91
  92       case 'civicrm_address':
  93         $id = 'contact_id';
  94         $on[] = 't1.location_type_id = t2.location_type_id';
  95         $using[] = 'location_type_id';
  96         if ($this->params['civicrm_address']['location_type_id']) {
  97           $locTypeId = CRM_Utils_Type::escape($this->params['civicrm_address']['location_type_id'], 'Integer', FALSE);
  98           if ($locTypeId) {
  99             $where[] = "t1.location_type_id = $locTypeId";
 100           }
 101         }
 102         break;
 103
 104       case 'civicrm_email':
 105       case 'civicrm_im':
 106       case 'civicrm_openid':
 107       case 'civicrm_phone':
 108         $id = 'contact_id';
 109         break;
 110
 111       case 'civicrm_note':
 112         $id = 'entity_id';
 113         if ($this->params) {
 114           $where[] = "t1.entity_table = 'civicrm_contact'";
 115         }
 116         else {
 117           $where[] = "t1.entity_table = 'civicrm_contact'";
 118           $where[] = "t2.entity_table = 'civicrm_contact'";
 119         }
 120         break;
 121
 122       default:
 123         // custom data tables
 124         if (preg_match('/^civicrm_value_/', $this->rule_table) || preg_match('/^custom_value_/', $this->rule_table)) {
 125           $id = 'entity_id';
 126         }
 127         else {
 128           CRM_Core_Error::fatal("Unsupported rule_table for civicrm_dedupe_rule.id of {$this->id}");
 129         }
 130         break;
 131     }
 132
 133     // build SELECT based on the field names containing contact ids
 134     // if there are params provided, id1 should be 0
 135     if ($this->params) {
 136       $select = "t1.$id id1, {$this->rule_weight} weight";
 137     }
 138     else {
 139       $select = "t1.$id id1, t2.$id id2, {$this->rule_weight} weight";
 140     }
 141
 142     // build FROM (and WHERE, if it's a parametrised search)
 143     // based on whether the rule is about substrings or not
 144     if ($this->params) {
 145       $from = "{$this->rule_table} t1";
 146       $str = 'NULL';
 147       if (isset($this->params[$this->rule_table][$this->rule_field])) {
 148         $str = CRM_Utils_Type::escape($this->params[$this->rule_table][$this->rule_field], 'String');
 149       }
 150       if ($this->rule_length) {
 151         $where[] = "SUBSTR(t1.{$this->rule_field}, 1, {$this->rule_length}) = SUBSTR('$str', 1, {$this->rule_length})";
 152         $where[] = "t1.{$this->rule_field} IS NOT NULL";
 153       }
 154       else {
 155         $where[] = "t1.{$this->rule_field} = '$str'";
 156       }
 157     }
 158     else {
 159       if ($this->rule_length) {
 160         $from = "{$this->rule_table} t1 JOIN {$this->rule_table} t2 ON (" . implode(' AND ', $on) . ")";
 161       }
 162       else {
 163         $from = "{$this->rule_table} t1 JOIN {$this->rule_table} t2 USING (" . implode(', ', $using) . ")";
 164       }
 165     }
 166
 167     // finish building WHERE, also limit the results if requested
 168     if (!$this->params) {
 169       $where[] = "t1.$id < t2.$id";
 170       $where[] = "t1.{$this->rule_field} IS NOT NULL";
 171     }
 172     if ($this->contactIds) {
 173       $cids = array();
 174       foreach ($this->contactIds as $cid) {
 175         $cids[] = CRM_Utils_Type::escape($cid, 'Integer');
 176       }
 177       if (count($cids) == 1) {
 178         $where[] = "(t1.$id = {$cids[0]} OR t2.$id = {$cids[0]})";
 179       }
 180       else {
 181         $where[] = "(t1.$id IN (" . implode(',', $cids) . ") OR t2.$id IN (" . implode(',', $cids) . "))";
 182       }
 183     }
 184
 185     return "SELECT $select FROM $from WHERE " . implode(' AND ', $where);
 186   }
 187
 188   /**
 189    * find fields related to a rule group.
 190    *
 191    * @param array contains the rule group property to identify rule group
 192    *
 193    * @return array
 194    *   rule fields array associated to rule group
 195    */
 196   public static function dedupeRuleFields($params) {
 197     $rgBao = new CRM_Dedupe_BAO_RuleGroup();
 198     $rgBao->used = $params['used'];
 199     $rgBao->contact_type = $params['contact_type'];
 200     $rgBao->find(TRUE);
 201
 202     $ruleBao = new CRM_Dedupe_BAO_Rule();
 203     $ruleBao->dedupe_rule_group_id = $rgBao->id;
 204     $ruleBao->find();
 205     $ruleFields = array();
 206     while ($ruleBao->fetch()) {
 207       $ruleFields[] = $ruleBao->rule_field;
 208     }
 209     return $ruleFields;
 210   }
 211
 212   /**
 213    * @param int $cid
 214    * @param int $oid
 215    *
 216    * @return bool
 217    */
 218   public static function validateContacts($cid, $oid) {
 219     if (!$cid || !$oid) {
 220       return;
 221     }
 222     $exception = new CRM_Dedupe_DAO_Exception();
 223     $exception->contact_id1 = $cid;
 224     $exception->contact_id2 = $oid;
 225     //make sure contact2 > contact1.
 226     if ($cid > $oid) {
 227       $exception->contact_id1 = $oid;
 228       $exception->contact_id2 = $cid;
 229     }
 230
 231     return $exception->find(TRUE) ? FALSE : TRUE;
 232   }
 233 }