tools/scripts/solr/createSolrXML.php

   1 <?php
   2 /*
   3  +--------------------------------------------------------------------+
   4  | CiviCRM version 4.5                                                |
   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.                                       |
  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 along with this program; if not, contact CiviCRM LLC       |
  21  | at info[AT]civicrm[DOT]org. If you have questions about the        |
  22  | GNU Affero General Public License or the licensing of CiviCRM,     |
  23  | see the CiviCRM license FAQ at http://civicrm.org/licensing        |
  24  +--------------------------------------------------------------------+
  25 */
  26
  27 /**
  28  * Create a xml file for a set of contact ID's in a format digestible
  29  * by Solr
  30  */
  31
  32 require_once '../../civicrm.config.php';
  33 require_once 'CRM/Core/Config.php';
  34
  35 define('CHUNK_SIZE', 128);
  36
  37 /**
  38  * Split a large array of contactIDs into more manageable smaller chunks
  39  */
  40 function &splitContactIDs(&$contactIDs) {
  41   // contactIDs could be a real large array, so we split it up into
  42   // smaller chunks and then general xml for each chunk
  43   $chunks           = array();
  44   $current          = 0;
  45   $chunks[$current] = array();
  46   $count            = 0;
  47
  48   foreach ($contactIDs as $cid) {
  49     $chunks[$current][] = $cid;
  50     $count++;
  51
  52     if ($count == CHUNK_SIZE) {
  53       $current++;
  54       $chunks[$current] = array();
  55       $count = 0;
  56     }
  57   }
  58
  59   if (empty($chunks[$current])) {
  60     unset($chunks[$current]);
  61   }
  62
  63   return $chunks;
  64 }
  65
  66 /**
  67  * Given an array of values, generate the XML in the Solr format
  68  */
  69 function &generateSolrXML($values) {
  70   $result = "<add>\n";
  71   foreach ($values as $cid => $tokens) {
  72     if (empty($tokens)) {
  73       continue;
  74     }
  75
  76     $result .= <<<EOT
  77   <doc>
  78     <field name="id">$cid</field>\n
  79 EOT;
  80
  81     foreach ($tokens as $t) {
  82       $result .= <<<EOT
  83     <field name="$t[0]">$t[1]</field>\n
  84 EOT;
  85     }
  86
  87     $result .= "  </doc>\n";
  88   }
  89   $result .= "</add>\n";
  90
  91
  92   return $result;
  93 }
  94
  95 /**
  96  * Given a set of contact IDs get the values
  97  */
  98 function getValues(&$contactIDs, &$values) {
  99   $values = array();
 100
 101   foreach ($contactIDs as $cid) {
 102     $values[$cid] = array();
 103   }
 104
 105   getContactInfo($contactIDs, $values);
 106   getLocationInfo($contactIDs, $values);
 107
 108   return $values;
 109 }
 110
 111 /**
 112  * @param $contactIDs
 113  * @param $values
 114  * @param $tableName
 115  * @param $fields
 116  * @param $whereField
 117  * @param null $additionalWhereCond
 118  */
 119 function getTableInfo(&$contactIDs, &$values, $tableName, &$fields, $whereField, $additionalWhereCond = NULL) {
 120   $selectString = implode(',', array_keys($fields));
 121   $idString = implode(',', $contactIDs);
 122
 123   $sql = "
 124 SELECT $selectString, $whereField as contact_id
 125   FROM $tableName
 126  WHERE $whereField IN ( $idString )
 127 ";
 128
 129   if ($additionalWhereCond) {
 130     $sql .= " AND $additionalWhereCond";
 131   }
 132
 133   $dao = &CRM_Core_DAO::executeQuery($sql, CRM_Core_DAO::$_nullArray);
 134   while ($dao->fetch()) {
 135     foreach ($fields as $fld => $name) {
 136       if (empty($dao->$fld)) {
 137         continue;
 138       }
 139       if (!$name) {
 140         $name = $fld;
 141       }
 142       $values[$dao->contact_id][] = array($name, $dao->$fld);
 143     }
 144   }
 145 }
 146
 147 /**
 148  * @param $contactIDs
 149  * @param $values
 150  */
 151 function getContactInfo(&$contactIDs, &$values) {
 152   $fields = array('sort_name' => NULL,
 153     'display_name' => NULL,
 154     'contact_type' => NULL,
 155     'legal_identifier' => NULL,
 156     'external_identifier' => NULL,
 157     'source' => 'contact_source',
 158   );
 159   getTableInfo($contactIDs, $values, 'civicrm_contact', $fields, 'id');
 160
 161   $fields = array('first_name' => NULL,
 162     'last_name' => NULL,
 163     'middle_name' => NULL,
 164     'job_title' => NULL,
 165   );
 166   getTableInfo($contactIDs, $values, 'civicrm_individual', $fields, 'contact_id');
 167
 168   $fields = array('household_name' => NULL);
 169   getTableInfo($contactIDs, $values, 'civicrm_household', $fields, 'contact_id');
 170
 171   $fields = array('organization_name' => NULL,
 172     'legal_name' => NULL,
 173     'sic_code' => NULL,
 174   );
 175   getTableInfo($contactIDs, $values, 'civicrm_organization', $fields, 'contact_id');
 176
 177   $fields = array('note' => 'note_body',
 178     'subject' => 'note_subject',
 179   );
 180   getTableInfo($contactIDs, $values, 'civicrm_note', $fields, 'entity_id', "entity_table = 'civicrm_contact'");
 181 }
 182
 183 /**
 184  * @param $contactIDs
 185  * @param $values
 186  */
 187 function getLocationInfo(&$contactIDs, &$values) {
 188   $ids = implode(',', $contactIDs);
 189
 190   $sql = "
 191 SELECT
 192   l.entity_id as contact_id, l.name as location_name,
 193   a.street_address, a.supplemental_address_1, a.supplemental_address_2,
 194   a.city, a.postal_code,
 195   co.name as county, s.name as state, c.name as country,
 196   e.email, p.phone, i.name as im
 197 FROM
 198   civicrm_location l
 199 LEFT JOIN civicrm_address        a  ON a.location_id       = l.id
 200 LEFT JOIN civicrm_email          e  ON e.location_id       = l.id
 201 LEFT JOIN civicrm_phone          p  ON p.location_id       = l.id
 202 LEFT JOIN civicrm_im             i  ON i.location_id       = l.id
 203 LEFT JOIN civicrm_state_province s  ON a.state_province_id = s.id
 204 LEFT JOIN civicrm_country        c  ON a.country_id        = c.id
 205 LEFT JOIN civicrm_county         co ON a.county_id         = co.id
 206 WHERE l.entity_table = 'civicrm_contact'
 207   AND l.entity_id IN ( $ids )
 208 ";
 209
 210   $fields = array('location_name', 'street_address', 'supplemental_address_1',
 211     'supplemental_address_2', 'city', 'postal_code', 'county', 'state',
 212     'country', 'email', 'phone', 'im',
 213   );
 214   $dao = &CRM_Core_DAO::executeQuery($sql, CRM_Core_DAO::$_nullArray);
 215   while ($dao->fetch()) {
 216     foreach ($fields as $fld) {
 217       if (empty($dao->$fld)) {
 218         continue;
 219       }
 220       $values[$dao->contact_id][] = array($fld, $dao->$fld);
 221     }
 222   }
 223 }
 224
 225 /**
 226  * @param $contactIDs
 227  */
 228 function run(&$contactIDs) {
 229   $chunks = &splitContactIDs($contactIDs);
 230
 231   foreach ($chunks as $chunk) {
 232     $values = array();
 233     getValues($chunk, $values);
 234     $xml = &generateSolrXML($values);
 235     echo $xml;
 236   }
 237 }
 238
 239 $config = &CRM_Core_Config::singleton();
 240 $config->userFramework = 'Soap';
 241 $config->userFrameworkClass = 'CRM_Utils_System_Soap';
 242 $config->userHookClass = 'CRM_Utils_Hook_Soap';
 243
 244 $sql = <<<EOT
 245 SELECT id
 246 FROM civicrm_contact
 247 EOT;
 248 $dao = &CRM_Core_DAO::executeQuery($sql, CRM_Core_DAO::$_nullArray);
 249
 250 $contactIDs = array();
 251 while ($dao->fetch()) {
 252   $contactIDs[] = $dao->id;
 253 }
 254
 255 run($contactIDs);
 256