3 +--------------------------------------------------------------------+
4 | Copyright CiviCRM LLC. All rights reserved. |
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 +--------------------------------------------------------------------+
15 * @copyright CiviCRM LLC https://civicrm.org/licensing
21 * The CiviCRM duplicate discovery engine is based on an
22 * algorithm designed by David Strauss <david@fourkitchens.com>.
24 class CRM_Dedupe_Finder
{
27 * Return a contact_id-keyed array of arrays of possible dupes
28 * (of the key contact_id) - limited to dupes of $cids if provided.
33 * Contact ids to limit the search to.
35 * @param bool $checkPermissions
36 * Respect logged in user permissions.
39 * Array of (cid1, cid2, weight) dupe triples
41 * @throws \CRM_Core_Exception
43 public static function dupes($rgid, $cids = [], $checkPermissions = TRUE) {
44 $rgBao = new CRM_Dedupe_BAO_RuleGroup();
46 $rgBao->contactIds
= $cids;
47 if (!$rgBao->find(TRUE)) {
48 throw new CRM_Core_Exception('Dedupe rule not found for selected contacts');
52 $dao = new CRM_Core_DAO();
53 $dao->query($rgBao->thresholdQuery($checkPermissions));
55 while ($dao->fetch()) {
56 $dupes[] = [$dao->id1
, $dao->id2
, $dao->weight
];
58 $dao->query($rgBao->tableDropQuery());
64 * Return an array of possible dupes, based on the provided array of
65 * params, using the default rule group for the given contact type and
68 * check_permission is a boolean flag to indicate if permission should be considered.
69 * default is to always check permissioning but public pages for example might not want
70 * permission to be checked for anonymous users. Refer CRM-6211. We might be breaking
71 * Multi-Site dedupe for public pages.
73 * @param array $params
74 * Array of params of the form $params[$table][$field] == $value.
75 * @param string $ctype
76 * Contact type to match against.
78 * Dedupe rule group usage ('Unsupervised' or 'Supervised' or 'General').
79 * @param array $except
80 * Array of contacts that shouldn't be considered dupes.
81 * @param int $ruleGroupID
82 * The id of the dedupe rule we should be using.
85 * matching contact ids
86 * @throws \CRM_Core_Exception
88 public static function dupesByParams(
91 $used = 'Unsupervised',
95 // If $params is empty there is zero reason to proceed.
99 $checkPermission = CRM_Utils_Array
::value('check_permission', $params, TRUE);
100 // This may no longer be required - see https://github.com/civicrm/civicrm-core/pull/13176
101 $params = array_filter($params);
105 $rgBao = new CRM_Dedupe_BAO_RuleGroup();
106 $rgBao->id
= $ruleGroupID;
107 $rgBao->contact_type
= $ctype;
108 if ($rgBao->find(TRUE)) {
114 $rgBao = new CRM_Dedupe_BAO_RuleGroup();
115 $rgBao->contact_type
= $ctype;
116 $rgBao->used
= $used;
117 if (!$rgBao->find(TRUE)) {
118 throw new CRM_Core_Exception("$used rule for $ctype does not exist");
122 if (isset($params['civicrm_phone']['phone_numeric'])) {
123 $orig = $params['civicrm_phone']['phone_numeric'];
124 $params['civicrm_phone']['phone_numeric'] = preg_replace('/[^\d]/', '', $orig);
126 $rgBao->params
= $params;
128 $dao = new CRM_Core_DAO();
129 $dao->query($rgBao->thresholdQuery($checkPermission));
131 while ($dao->fetch()) {
132 if (isset($dao->id
) && $dao->id
) {
136 $dao->query($rgBao->tableDropQuery());
137 return array_diff($dupes, $except);
141 * Return a contact_id-keyed array of arrays of possible dupes in the given group.
148 * @param int $searchLimit
149 * Limit for the number of contacts to be used for comparison.
150 * The search methodology finds all matches for the searchedContacts so this limits
151 * the number of searched contacts, not the matches found.
154 * array of (cid1, cid2, weight) dupe triples
156 * @throws \CRM_Core_Exception
158 public static function dupesInGroup($rgid, $gid, $searchLimit = 0) {
159 $cids = array_keys(CRM_Contact_BAO_Group
::getMember($gid, TRUE, $searchLimit));
161 return self
::dupes($rgid, $cids);
167 * A hackish function needed to massage CRM_Contact_Form_$ctype::formRule()
168 * object into a valid $params array for dedupe
170 * @param array $fields
171 * Contact structure from formRule().
172 * @param string $ctype
173 * Contact type of the given contact.
176 * valid $params array for dedupe
177 * @throws \CRM_Core_Exception
179 public static function formatParams($fields, $ctype) {
181 CRM_Utils_Array
::flatten($fields, $flat);
183 // FIXME: This may no longer be necessary - check inputs
185 'individual_prefix' => 'prefix_id',
186 'individual_suffix' => 'suffix_id',
187 'gender' => 'gender_id',
189 foreach (['individual_suffix', 'individual_prefix', 'gender'] as $name) {
190 if (!empty($fields[$name])) {
191 $flat[$replace_these[$name]] = $flat[$name];
196 // handle {birth,deceased}_date
201 if (!empty($fields[$date])) {
202 $flat[$date] = $fields[$date];
203 if (is_array($flat[$date])) {
204 $flat[$date] = CRM_Utils_Date
::format($flat[$date]);
206 $flat[$date] = CRM_Utils_Date
::processDate($flat[$date]);
210 if (!empty($flat['contact_source'])) {
211 $flat['source'] = $flat['contact_source'];
212 unset($flat['contact_source']);
215 // handle preferred_communication_method
216 if (!empty($fields['preferred_communication_method'])) {
217 $methods = array_intersect($fields['preferred_communication_method'], ['1']);
218 $methods = array_keys($methods);
221 $flat['preferred_communication_method'] = CRM_Core_DAO
::VALUE_SEPARATOR
. implode(CRM_Core_DAO
::VALUE_SEPARATOR
, $methods) . CRM_Core_DAO
::VALUE_SEPARATOR
;
225 // handle custom data
226 $tree = CRM_Core_BAO_CustomGroup
::getTree($ctype, NULL, NULL, -1);
227 CRM_Core_BAO_CustomGroup
::postProcess($tree, $fields, TRUE);
228 foreach ($tree as $key => $cg) {
232 foreach ($cg['fields'] as $cf) {
233 $flat[$cf['column_name']] = CRM_Utils_Array
::value('data', $cf['customValue']);
237 // if the key is dotted, keep just the last part of it
238 foreach ($flat as $key => $value) {
239 if (substr_count($key, '.')) {
240 $last = explode('.', $key);
241 $last = array_pop($last);
242 // make sure the first occurrence is kept, not the last
243 if (!isset($flat[$last])) {
244 $flat[$last] = $value;
250 // drop the -digit (and -Primary, for CRM-3902) postfixes (so event registration's $flat['email-5'] becomes $flat['email'])
251 // FIXME: CRM-5026 should be fixed here; the below clobbers all address info; we should split off address fields and match
252 // the -digit to civicrm_address.location_type_id and -Primary to civicrm_address.is_primary
253 foreach ($flat as $key => $value) {
255 if (preg_match('/(.*)-(Primary-[\d+])$|(.*)-(\d+|Primary)$/', $key, $matches)) {
256 $return = array_values(array_filter($matches));
257 // make sure the first occurrence is kept, not the last
258 $flat[$return[1]] = empty($flat[$return[1]]) ?
$value : $flat[$return[1]];
264 $supportedFields = CRM_Dedupe_BAO_RuleGroup
::supportedFields($ctype);
265 if (is_array($supportedFields)) {
266 foreach ($supportedFields as $table => $fields) {
267 if ($table === 'civicrm_address') {
268 // for matching on civicrm_address fields, we also need the location_type_id
269 $fields['location_type_id'] = '';
270 // FIXME: we also need to do some hacking for id and name fields, see CRM-3902’s comments
272 'address_name' => 'name',
273 'country' => 'country_id',
274 'state_province' => 'state_province_id',
275 'county' => 'county_id',
277 foreach ($fixes as $orig => $target) {
278 if (!empty($flat[$orig])) {
279 $params[$table][$target] = $flat[$orig];
283 if ($table === 'civicrm_phone') {
285 'phone' => 'phone_numeric',
287 foreach ($fixes as $orig => $target) {
288 if (!empty($flat[$orig])) {
289 $params[$table][$target] = $flat[$orig];
293 foreach ($fields as $field => $title) {
294 if (!empty($flat[$field])) {
295 $params[$table][$field] = $flat[$field];
304 * Parse duplicate pairs into a standardised array and store in the prev_next_cache.
306 * @param array $foundDupes
307 * @param string $cacheKeyString
310 * Dupe pairs with the keys
318 public static function parseAndStoreDupePairs($foundDupes, $cacheKeyString) {
320 foreach ($foundDupes as $dupe) {
324 $cidString = implode(', ', array_keys($cids));
326 $dao = CRM_Core_DAO
::executeQuery("SELECT id, display_name FROM civicrm_contact WHERE id IN ($cidString) ORDER BY sort_name");
328 while ($dao->fetch()) {
329 $displayNames[$dao->id
] = $dao->display_name
;
332 $userId = CRM_Core_Session
::getLoggedInContactID();
333 foreach ($foundDupes as $dupes) {
336 // The logged in user should never be the src (ie. the contact to be removed).
337 if ($srcID == $userId) {
342 $mainContacts[] = $row = [
344 'dstName' => $displayNames[$dstID],
346 'srcName' => $displayNames[$srcID],
347 'weight' => $dupes[2],
351 $data = CRM_Core_DAO
::escapeString(serialize($row));
352 CRM_Core_BAO_PrevNextCache
::setItem('civicrm_contact', $dstID, $srcID, $cacheKeyString, $data);
354 return $mainContacts;