3 +--------------------------------------------------------------------+
5 +--------------------------------------------------------------------+
6 | Copyright CiviCRM LLC (c) 2004-2019 |
7 +--------------------------------------------------------------------+
8 | This file is a part of CiviCRM. |
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. |
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. |
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 +--------------------------------------------------------------------+
31 * @copyright CiviCRM LLC (c) 2004-2019
37 * The CiviCRM duplicate discovery engine is based on an
38 * algorithm designed by David Strauss <david@fourkitchens.com>.
40 class CRM_Dedupe_Finder
{
43 * Return a contact_id-keyed array of arrays of possible dupes
44 * (of the key contact_id) - limited to dupes of $cids if provided.
49 * Contact ids to limit the search to.
51 * @param bool $checkPermissions
52 * Respect logged in user permissions.
54 * @param int $searchLimit
55 * Limit for the number of contacts to be used for comparison.
56 * The search methodology finds all matches for the searchedContacts so this limits
57 * the number of searched contacts, not the matches found.
60 * Array of (cid1, cid2, weight) dupe triples
62 * @throws CiviCRM_API3_Exception
65 public static function dupes($rgid, $cids = [], $checkPermissions = TRUE, $searchLimit = 0) {
66 $rgBao = new CRM_Dedupe_BAO_RuleGroup();
68 $rgBao->contactIds
= $cids;
69 if (!$rgBao->find(TRUE)) {
70 CRM_Core_Error
::fatal("Dedupe rule not found for selected contacts");
72 if (empty($rgBao->contactIds
) && !empty($searchLimit)) {
73 $limitedContacts = civicrm_api3('Contact', 'get', [
75 'contact_type' => $rgBao->contact_type
,
76 'options' => ['limit' => $searchLimit],
78 $rgBao->contactIds
= array_keys($limitedContacts['values']);
82 $dao = new CRM_Core_DAO();
83 $dao->query($rgBao->thresholdQuery($checkPermissions));
85 while ($dao->fetch()) {
86 $dupes[] = [$dao->id1
, $dao->id2
, $dao->weight
];
88 $dao->query($rgBao->tableDropQuery());
94 * Return an array of possible dupes, based on the provided array of
95 * params, using the default rule group for the given contact type and
98 * check_permission is a boolean flag to indicate if permission should be considered.
99 * default is to always check permissioning but public pages for example might not want
100 * permission to be checked for anonymous users. Refer CRM-6211. We might be breaking
101 * Multi-Site dedupe for public pages.
103 * @param array $params
104 * Array of params of the form $params[$table][$field] == $value.
105 * @param string $ctype
106 * Contact type to match against.
107 * @param string $used
108 * Dedupe rule group usage ('Unsupervised' or 'Supervised' or 'General').
109 * @param array $except
110 * Array of contacts that shouldn't be considered dupes.
111 * @param int $ruleGroupID
112 * The id of the dedupe rule we should be using.
115 * matching contact ids
117 public static function dupesByParams(
120 $used = 'Unsupervised',
124 // If $params is empty there is zero reason to proceed.
128 $checkPermission = CRM_Utils_Array
::value('check_permission', $params, TRUE);
129 // This may no longer be required - see https://github.com/civicrm/civicrm-core/pull/13176
130 $params = array_filter($params);
134 $rgBao = new CRM_Dedupe_BAO_RuleGroup();
135 $rgBao->id
= $ruleGroupID;
136 $rgBao->contact_type
= $ctype;
137 if ($rgBao->find(TRUE)) {
143 $rgBao = new CRM_Dedupe_BAO_RuleGroup();
144 $rgBao->contact_type
= $ctype;
145 $rgBao->used
= $used;
146 if (!$rgBao->find(TRUE)) {
147 CRM_Core_Error
::fatal("$used rule for $ctype does not exist");
151 if (isset($params['civicrm_phone']['phone_numeric'])) {
152 $orig = $params['civicrm_phone']['phone_numeric'];
153 $params['civicrm_phone']['phone_numeric'] = preg_replace('/[^\d]/', '', $orig);
155 $rgBao->params
= $params;
157 $dao = new CRM_Core_DAO();
158 $dao->query($rgBao->thresholdQuery($checkPermission));
160 while ($dao->fetch()) {
161 if (isset($dao->id
) && $dao->id
) {
165 $dao->query($rgBao->tableDropQuery());
166 return array_diff($dupes, $except);
170 * Return a contact_id-keyed array of arrays of possible dupes in the given group.
177 * @param int $searchLimit
178 * Limit for the number of contacts to be used for comparison.
179 * The search methodology finds all matches for the searchedContacts so this limits
180 * the number of searched contacts, not the matches found.
183 * array of (cid1, cid2, weight) dupe triples
185 public static function dupesInGroup($rgid, $gid, $searchLimit = 0) {
186 $cids = array_keys(CRM_Contact_BAO_Group
::getMember($gid, TRUE, $searchLimit));
188 return self
::dupes($rgid, $cids);
194 * A hackish function needed to massage CRM_Contact_Form_$ctype::formRule()
195 * object into a valid $params array for dedupe
197 * @param array $fields
198 * Contact structure from formRule().
199 * @param string $ctype
200 * Contact type of the given contact.
203 * valid $params array for dedupe
205 public static function formatParams($fields, $ctype) {
207 CRM_Utils_Array
::flatten($fields, $flat);
209 // FIXME: This may no longer be necessary - check inputs
211 'individual_prefix' => 'prefix_id',
212 'individual_suffix' => 'suffix_id',
213 'gender' => 'gender_id',
215 foreach (['individual_suffix', 'individual_prefix', 'gender'] as $name) {
216 if (!empty($fields[$name])) {
217 $flat[$replace_these[$name]] = $flat[$name];
222 // handle {birth,deceased}_date
227 if (!empty($fields[$date])) {
228 $flat[$date] = $fields[$date];
229 if (is_array($flat[$date])) {
230 $flat[$date] = CRM_Utils_Date
::format($flat[$date]);
232 $flat[$date] = CRM_Utils_Date
::processDate($flat[$date]);
236 if (!empty($flat['contact_source'])) {
237 $flat['source'] = $flat['contact_source'];
238 unset($flat['contact_source']);
241 // handle preferred_communication_method
242 if (!empty($fields['preferred_communication_method'])) {
243 $methods = array_intersect($fields['preferred_communication_method'], ['1']);
244 $methods = array_keys($methods);
247 $flat['preferred_communication_method'] = CRM_Core_DAO
::VALUE_SEPARATOR
. implode(CRM_Core_DAO
::VALUE_SEPARATOR
, $methods) . CRM_Core_DAO
::VALUE_SEPARATOR
;
251 // handle custom data
252 $tree = CRM_Core_BAO_CustomGroup
::getTree($ctype, NULL, NULL, -1);
253 CRM_Core_BAO_CustomGroup
::postProcess($tree, $fields, TRUE);
254 foreach ($tree as $key => $cg) {
258 foreach ($cg['fields'] as $cf) {
259 $flat[$cf['column_name']] = CRM_Utils_Array
::value('data', $cf['customValue']);
263 // if the key is dotted, keep just the last part of it
264 foreach ($flat as $key => $value) {
265 if (substr_count($key, '.')) {
266 $last = explode('.', $key);
267 $last = array_pop($last);
268 // make sure the first occurrence is kept, not the last
269 if (!isset($flat[$last])) {
270 $flat[$last] = $value;
276 // drop the -digit (and -Primary, for CRM-3902) postfixes (so event registration's $flat['email-5'] becomes $flat['email'])
277 // FIXME: CRM-5026 should be fixed here; the below clobbers all address info; we should split off address fields and match
278 // the -digit to civicrm_address.location_type_id and -Primary to civicrm_address.is_primary
279 foreach ($flat as $key => $value) {
281 if (preg_match('/(.*)-(Primary-[\d+])$|(.*)-(\d+|Primary)$/', $key, $matches)) {
282 $return = array_values(array_filter($matches));
283 $flat[$return[1]] = $value;
289 $supportedFields = CRM_Dedupe_BAO_RuleGroup
::supportedFields($ctype);
290 if (is_array($supportedFields)) {
291 foreach ($supportedFields as $table => $fields) {
292 if ($table == 'civicrm_address') {
293 // for matching on civicrm_address fields, we also need the location_type_id
294 $fields['location_type_id'] = '';
295 // FIXME: we also need to do some hacking for id and name fields, see CRM-3902’s comments
297 'address_name' => 'name',
298 'country' => 'country_id',
299 'state_province' => 'state_province_id',
300 'county' => 'county_id',
302 foreach ($fixes as $orig => $target) {
303 if (!empty($flat[$orig])) {
304 $params[$table][$target] = $flat[$orig];
308 if ($table == 'civicrm_phone') {
310 'phone' => 'phone_numeric',
312 foreach ($fixes as $orig => $target) {
313 if (!empty($flat[$orig])) {
314 $params[$table][$target] = $flat[$orig];
318 foreach ($fields as $field => $title) {
319 if (!empty($flat[$field])) {
320 $params[$table][$field] = $flat[$field];
329 * Parse duplicate pairs into a standardised array and store in the prev_next_cache.
331 * @param array $foundDupes
332 * @param string $cacheKeyString
335 * Dupe pairs with the keys
343 * @throws CRM_Core_Exception
345 public static function parseAndStoreDupePairs($foundDupes, $cacheKeyString) {
347 foreach ($foundDupes as $dupe) {
351 $cidString = implode(', ', array_keys($cids));
353 $dao = CRM_Core_DAO
::executeQuery("SELECT id, display_name FROM civicrm_contact WHERE id IN ($cidString) ORDER BY sort_name");
355 while ($dao->fetch()) {
356 $displayNames[$dao->id
] = $dao->display_name
;
359 $userId = CRM_Core_Session
::getLoggedInContactID();
360 foreach ($foundDupes as $dupes) {
363 // The logged in user should never be the src (ie. the contact to be removed).
364 if ($srcID == $userId) {
369 $mainContacts[] = $row = [
371 'dstName' => $displayNames[$dstID],
373 'srcName' => $displayNames[$srcID],
374 'weight' => $dupes[2],
378 $data = CRM_Core_DAO
::escapeString(serialize($row));
379 CRM_Core_BAO_PrevNextCache
::setItem('civicrm_contact', $dstID, $srcID, $cacheKeyString, $data);
381 return $mainContacts;