| 1 | <?php |
| 2 | /* |
| 3 | +--------------------------------------------------------------------+ |
| 4 | | Copyright CiviCRM LLC. All rights reserved. | |
| 5 | | | |
| 6 | | This work is published under the GNU AGPLv3 license with some | |
| 7 | | permitted exceptions and without any warranty. For full license | |
| 8 | | and copyright information, see https://civicrm.org/licensing | |
| 9 | +--------------------------------------------------------------------+ |
| 10 | */ |
| 11 | |
| 12 | /** |
| 13 | * |
| 14 | * @package CRM |
| 15 | * @copyright CiviCRM LLC https://civicrm.org/licensing |
| 16 | */ |
| 17 | |
| 18 | /** |
| 19 | * The CiviCRM duplicate discovery engine is based on an |
| 20 | * algorithm designed by David Strauss <david@fourkitchens.com>. |
| 21 | */ |
| 22 | class CRM_Dedupe_Finder { |
| 23 | |
| 24 | /** |
| 25 | * Return a contact_id-keyed array of arrays of possible dupes |
| 26 | * (of the key contact_id) - limited to dupes of $cids if provided. |
| 27 | * |
| 28 | * @param int $rgid |
| 29 | * Rule group id. |
| 30 | * @param array $cids |
| 31 | * Contact ids to limit the search to. |
| 32 | * |
| 33 | * @param bool $checkPermissions |
| 34 | * Respect logged in user permissions. |
| 35 | * |
| 36 | * @return array |
| 37 | * Array of (cid1, cid2, weight) dupe triples |
| 38 | * |
| 39 | * @throws \CRM_Core_Exception |
| 40 | */ |
| 41 | public static function dupes($rgid, $cids = [], $checkPermissions = TRUE) { |
| 42 | $rgBao = new CRM_Dedupe_BAO_RuleGroup(); |
| 43 | $rgBao->id = $rgid; |
| 44 | $rgBao->contactIds = $cids; |
| 45 | if (!$rgBao->find(TRUE)) { |
| 46 | throw new CRM_Core_Exception('Dedupe rule not found for selected contacts'); |
| 47 | } |
| 48 | |
| 49 | $rgBao->fillTable(); |
| 50 | $dao = new CRM_Core_DAO(); |
| 51 | $dao->query($rgBao->thresholdQuery($checkPermissions)); |
| 52 | $dupes = []; |
| 53 | while ($dao->fetch()) { |
| 54 | $dupes[] = [$dao->id1, $dao->id2, $dao->weight]; |
| 55 | } |
| 56 | $dao->query($rgBao->tableDropQuery()); |
| 57 | |
| 58 | return $dupes; |
| 59 | } |
| 60 | |
| 61 | /** |
| 62 | * Return an array of possible dupes, based on the provided array of |
| 63 | * params, using the default rule group for the given contact type and |
| 64 | * usage. |
| 65 | * |
| 66 | * check_permission is a boolean flag to indicate if permission should be considered. |
| 67 | * default is to always check permissioning but public pages for example might not want |
| 68 | * permission to be checked for anonymous users. Refer CRM-6211. We might be breaking |
| 69 | * Multi-Site dedupe for public pages. |
| 70 | * |
| 71 | * @param array $params |
| 72 | * Array of params of the form $params[$table][$field] == $value. |
| 73 | * @param string $ctype |
| 74 | * Contact type to match against. |
| 75 | * @param string $used |
| 76 | * Dedupe rule group usage ('Unsupervised' or 'Supervised' or 'General'). |
| 77 | * @param array $except |
| 78 | * Array of contacts that shouldn't be considered dupes. |
| 79 | * @param int $ruleGroupID |
| 80 | * The id of the dedupe rule we should be using. |
| 81 | * |
| 82 | * @return array |
| 83 | * matching contact ids |
| 84 | * @throws \CRM_Core_Exception |
| 85 | */ |
| 86 | public static function dupesByParams( |
| 87 | $params, |
| 88 | $ctype, |
| 89 | $used = 'Unsupervised', |
| 90 | $except = [], |
| 91 | $ruleGroupID = NULL |
| 92 | ) { |
| 93 | // If $params is empty there is zero reason to proceed. |
| 94 | if (!$params) { |
| 95 | return []; |
| 96 | } |
| 97 | $checkPermission = CRM_Utils_Array::value('check_permission', $params, TRUE); |
| 98 | // This may no longer be required - see https://github.com/civicrm/civicrm-core/pull/13176 |
| 99 | $params = array_filter($params); |
| 100 | |
| 101 | $foundByID = FALSE; |
| 102 | if ($ruleGroupID) { |
| 103 | $rgBao = new CRM_Dedupe_BAO_RuleGroup(); |
| 104 | $rgBao->id = $ruleGroupID; |
| 105 | $rgBao->contact_type = $ctype; |
| 106 | if ($rgBao->find(TRUE)) { |
| 107 | $foundByID = TRUE; |
| 108 | } |
| 109 | } |
| 110 | |
| 111 | if (!$foundByID) { |
| 112 | $rgBao = new CRM_Dedupe_BAO_RuleGroup(); |
| 113 | $rgBao->contact_type = $ctype; |
| 114 | $rgBao->used = $used; |
| 115 | if (!$rgBao->find(TRUE)) { |
| 116 | throw new CRM_Core_Exception("$used rule for $ctype does not exist"); |
| 117 | } |
| 118 | } |
| 119 | |
| 120 | if (isset($params['civicrm_phone']['phone_numeric'])) { |
| 121 | $orig = $params['civicrm_phone']['phone_numeric']; |
| 122 | $params['civicrm_phone']['phone_numeric'] = preg_replace('/[^\d]/', '', $orig); |
| 123 | } |
| 124 | $rgBao->params = $params; |
| 125 | $rgBao->fillTable(); |
| 126 | $dao = new CRM_Core_DAO(); |
| 127 | $dao->query($rgBao->thresholdQuery($checkPermission)); |
| 128 | $dupes = []; |
| 129 | while ($dao->fetch()) { |
| 130 | if (isset($dao->id) && $dao->id) { |
| 131 | $dupes[] = $dao->id; |
| 132 | } |
| 133 | } |
| 134 | $dao->query($rgBao->tableDropQuery()); |
| 135 | return array_diff($dupes, $except); |
| 136 | } |
| 137 | |
| 138 | /** |
| 139 | * Return a contact_id-keyed array of arrays of possible dupes in the given group. |
| 140 | * |
| 141 | * @param int $rgid |
| 142 | * Rule group id. |
| 143 | * @param int $gid |
| 144 | * Contact group id. |
| 145 | * |
| 146 | * @param int $searchLimit |
| 147 | * Limit for the number of contacts to be used for comparison. |
| 148 | * The search methodology finds all matches for the searchedContacts so this limits |
| 149 | * the number of searched contacts, not the matches found. |
| 150 | * |
| 151 | * @return array |
| 152 | * array of (cid1, cid2, weight) dupe triples |
| 153 | * |
| 154 | * @throws \CRM_Core_Exception |
| 155 | */ |
| 156 | public static function dupesInGroup($rgid, $gid, $searchLimit = 0) { |
| 157 | $cids = array_keys(CRM_Contact_BAO_Group::getMember($gid, TRUE, $searchLimit)); |
| 158 | if (!empty($cids)) { |
| 159 | return self::dupes($rgid, $cids); |
| 160 | } |
| 161 | return []; |
| 162 | } |
| 163 | |
| 164 | /** |
| 165 | * A hackish function needed to massage CRM_Contact_Form_$ctype::formRule() |
| 166 | * object into a valid $params array for dedupe |
| 167 | * |
| 168 | * @param array $fields |
| 169 | * Contact structure from formRule(). |
| 170 | * @param string $ctype |
| 171 | * Contact type of the given contact. |
| 172 | * |
| 173 | * @return array |
| 174 | * valid $params array for dedupe |
| 175 | * @throws \CRM_Core_Exception |
| 176 | */ |
| 177 | public static function formatParams($fields, $ctype) { |
| 178 | $flat = []; |
| 179 | CRM_Utils_Array::flatten($fields, $flat); |
| 180 | |
| 181 | // FIXME: This may no longer be necessary - check inputs |
| 182 | $replace_these = [ |
| 183 | 'individual_prefix' => 'prefix_id', |
| 184 | 'individual_suffix' => 'suffix_id', |
| 185 | 'gender' => 'gender_id', |
| 186 | ]; |
| 187 | foreach (['individual_suffix', 'individual_prefix', 'gender'] as $name) { |
| 188 | if (!empty($fields[$name])) { |
| 189 | $flat[$replace_these[$name]] = $flat[$name]; |
| 190 | unset($flat[$name]); |
| 191 | } |
| 192 | } |
| 193 | |
| 194 | // handle {birth,deceased}_date |
| 195 | foreach ([ |
| 196 | 'birth_date', |
| 197 | 'deceased_date', |
| 198 | ] as $date) { |
| 199 | if (!empty($fields[$date])) { |
| 200 | $flat[$date] = $fields[$date]; |
| 201 | if (is_array($flat[$date])) { |
| 202 | $flat[$date] = CRM_Utils_Date::format($flat[$date]); |
| 203 | } |
| 204 | $flat[$date] = CRM_Utils_Date::processDate($flat[$date]); |
| 205 | } |
| 206 | } |
| 207 | |
| 208 | if (!empty($flat['contact_source'])) { |
| 209 | $flat['source'] = $flat['contact_source']; |
| 210 | unset($flat['contact_source']); |
| 211 | } |
| 212 | |
| 213 | // handle preferred_communication_method |
| 214 | if (!empty($fields['preferred_communication_method'])) { |
| 215 | $methods = array_intersect($fields['preferred_communication_method'], ['1']); |
| 216 | $methods = array_keys($methods); |
| 217 | sort($methods); |
| 218 | if ($methods) { |
| 219 | $flat['preferred_communication_method'] = CRM_Core_DAO::VALUE_SEPARATOR . implode(CRM_Core_DAO::VALUE_SEPARATOR, $methods) . CRM_Core_DAO::VALUE_SEPARATOR; |
| 220 | } |
| 221 | } |
| 222 | |
| 223 | // handle custom data |
| 224 | $tree = CRM_Core_BAO_CustomGroup::getTree($ctype, NULL, NULL, -1); |
| 225 | CRM_Core_BAO_CustomGroup::postProcess($tree, $fields, TRUE); |
| 226 | foreach ($tree as $key => $cg) { |
| 227 | if (!is_int($key)) { |
| 228 | continue; |
| 229 | } |
| 230 | foreach ($cg['fields'] as $cf) { |
| 231 | $flat[$cf['column_name']] = $cf['customValue']['data'] ?? NULL; |
| 232 | } |
| 233 | } |
| 234 | |
| 235 | // if the key is dotted, keep just the last part of it |
| 236 | foreach ($flat as $key => $value) { |
| 237 | if (substr_count($key, '.')) { |
| 238 | $last = explode('.', $key); |
| 239 | $last = array_pop($last); |
| 240 | // make sure the first occurrence is kept, not the last |
| 241 | if (!isset($flat[$last])) { |
| 242 | $flat[$last] = $value; |
| 243 | } |
| 244 | unset($flat[$key]); |
| 245 | } |
| 246 | } |
| 247 | |
| 248 | // drop the -digit (and -Primary, for CRM-3902) postfixes (so event registration's $flat['email-5'] becomes $flat['email']) |
| 249 | // FIXME: CRM-5026 should be fixed here; the below clobbers all address info; we should split off address fields and match |
| 250 | // the -digit to civicrm_address.location_type_id and -Primary to civicrm_address.is_primary |
| 251 | foreach ($flat as $key => $value) { |
| 252 | $matches = []; |
| 253 | if (preg_match('/(.*)-(Primary-[\d+])$|(.*)-(\d+-\d+)$|(.*)-(\d+|Primary)$/', $key, $matches)) { |
| 254 | $return = array_values(array_filter($matches)); |
| 255 | // make sure the first occurrence is kept, not the last |
| 256 | $flat[$return[1]] = empty($flat[$return[1]]) ? $value : $flat[$return[1]]; |
| 257 | unset($flat[$key]); |
| 258 | } |
| 259 | } |
| 260 | |
| 261 | $params = []; |
| 262 | $supportedFields = CRM_Dedupe_BAO_RuleGroup::supportedFields($ctype); |
| 263 | if (is_array($supportedFields)) { |
| 264 | foreach ($supportedFields as $table => $fields) { |
| 265 | if ($table === 'civicrm_address') { |
| 266 | // for matching on civicrm_address fields, we also need the location_type_id |
| 267 | $fields['location_type_id'] = ''; |
| 268 | // FIXME: we also need to do some hacking for id and name fields, see CRM-3902’s comments |
| 269 | $fixes = [ |
| 270 | 'address_name' => 'name', |
| 271 | 'country' => 'country_id', |
| 272 | 'state_province' => 'state_province_id', |
| 273 | 'county' => 'county_id', |
| 274 | ]; |
| 275 | foreach ($fixes as $orig => $target) { |
| 276 | if (!empty($flat[$orig])) { |
| 277 | $params[$table][$target] = $flat[$orig]; |
| 278 | } |
| 279 | } |
| 280 | } |
| 281 | if ($table === 'civicrm_phone') { |
| 282 | $fixes = [ |
| 283 | 'phone' => 'phone_numeric', |
| 284 | ]; |
| 285 | foreach ($fixes as $orig => $target) { |
| 286 | if (!empty($flat[$orig])) { |
| 287 | $params[$table][$target] = $flat[$orig]; |
| 288 | } |
| 289 | } |
| 290 | } |
| 291 | foreach ($fields as $field => $title) { |
| 292 | if (!empty($flat[$field])) { |
| 293 | $params[$table][$field] = $flat[$field]; |
| 294 | } |
| 295 | } |
| 296 | } |
| 297 | } |
| 298 | return $params; |
| 299 | } |
| 300 | |
| 301 | /** |
| 302 | * Parse duplicate pairs into a standardised array and store in the prev_next_cache. |
| 303 | * |
| 304 | * @param array $foundDupes |
| 305 | * @param string $cacheKeyString |
| 306 | * |
| 307 | * @return array |
| 308 | * Dupe pairs with the keys |
| 309 | * -srcID |
| 310 | * -srcName |
| 311 | * -dstID |
| 312 | * -dstName |
| 313 | * -weight |
| 314 | * -canMerge |
| 315 | */ |
| 316 | public static function parseAndStoreDupePairs($foundDupes, $cacheKeyString) { |
| 317 | $cids = []; |
| 318 | foreach ($foundDupes as $dupe) { |
| 319 | $cids[$dupe[0]] = 1; |
| 320 | $cids[$dupe[1]] = 1; |
| 321 | } |
| 322 | $cidString = implode(', ', array_keys($cids)); |
| 323 | |
| 324 | $dao = CRM_Core_DAO::executeQuery("SELECT id, display_name FROM civicrm_contact WHERE id IN ($cidString) ORDER BY sort_name"); |
| 325 | $displayNames = []; |
| 326 | while ($dao->fetch()) { |
| 327 | $displayNames[$dao->id] = $dao->display_name; |
| 328 | } |
| 329 | |
| 330 | $userId = CRM_Core_Session::getLoggedInContactID(); |
| 331 | foreach ($foundDupes as $dupes) { |
| 332 | $srcID = $dupes[1]; |
| 333 | $dstID = $dupes[0]; |
| 334 | // The logged in user should never be the src (ie. the contact to be removed). |
| 335 | if ($srcID == $userId) { |
| 336 | $srcID = $dstID; |
| 337 | $dstID = $userId; |
| 338 | } |
| 339 | |
| 340 | $mainContacts[] = $row = [ |
| 341 | 'dstID' => (int) $dstID, |
| 342 | 'dstName' => $displayNames[$dstID], |
| 343 | 'srcID' => (int) $srcID, |
| 344 | 'srcName' => $displayNames[$srcID], |
| 345 | 'weight' => $dupes[2], |
| 346 | 'canMerge' => TRUE, |
| 347 | ]; |
| 348 | |
| 349 | CRM_Core_DAO::executeQuery("INSERT INTO civicrm_prevnext_cache (entity_table, entity_id1, entity_id2, cacheKey, data) VALUES |
| 350 | ('civicrm_contact', %1, %2, %3, %4)", [ |
| 351 | 1 => [$dstID, 'Integer'], |
| 352 | 2 => [$srcID, 'Integer'], |
| 353 | 3 => [$cacheKeyString, 'String'], |
| 354 | 4 => [serialize($row), 'String'], |
| 355 | ] |
| 356 | ); |
| 357 | } |
| 358 | return $mainContacts; |
| 359 | } |
| 360 | |
| 361 | } |