Commit | Line | Data |
---|---|---|
bb4187d7 MD |
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_BAO_DedupeRule extends CRM_Dedupe_DAO_DedupeRule { | |
23 | ||
24 | /** | |
25 | * Ids of the contacts to limit the SQL queries (whole-database queries otherwise) | |
26 | * @var array | |
27 | */ | |
28 | public $contactIds = []; | |
29 | ||
30 | /** | |
31 | * Params to dedupe against (queries against the whole contact set otherwise) | |
32 | * @var array | |
33 | */ | |
34 | public $params = []; | |
35 | ||
36 | /** | |
37 | * Return the SQL query for the given rule - either for finding matching | |
38 | * pairs of contacts, or for matching against the $params variable (if set). | |
39 | * | |
40 | * @return string | |
41 | * SQL query performing the search | |
42 | * | |
43 | * @throws \CRM_Core_Exception | |
44 | * @throws \CiviCRM_API3_Exception | |
45 | */ | |
46 | public function sql() { | |
47 | if ($this->params && | |
48 | (!array_key_exists($this->rule_table, $this->params) || | |
49 | !array_key_exists($this->rule_field, $this->params[$this->rule_table]) | |
50 | ) | |
51 | ) { | |
52 | // if params is present and doesn't have an entry for a field, don't construct the clause. | |
53 | return NULL; | |
54 | } | |
55 | ||
56 | // we need to initialise WHERE, ON and USING here, as some table types | |
57 | // extend them; $where is an array of required conditions, $on and | |
58 | // $using are arrays of required field matchings (for substring and | |
59 | // full matches, respectively) | |
60 | $where = []; | |
61 | $on = ["SUBSTR(t1.{$this->rule_field}, 1, {$this->rule_length}) = SUBSTR(t2.{$this->rule_field}, 1, {$this->rule_length})"]; | |
62 | ||
63 | $innerJoinClauses = [ | |
64 | "t1.{$this->rule_field} IS NOT NULL", | |
65 | "t2.{$this->rule_field} IS NOT NULL", | |
66 | "t1.{$this->rule_field} = t2.{$this->rule_field}", | |
67 | ]; | |
68 | ||
69 | if (in_array($this->getFieldType($this->rule_field), CRM_Utils_Type::getTextTypes(), TRUE)) { | |
70 | $innerJoinClauses[] = "t1.{$this->rule_field} <> ''"; | |
71 | $innerJoinClauses[] = "t2.{$this->rule_field} <> ''"; | |
72 | } | |
73 | ||
74 | switch ($this->rule_table) { | |
75 | case 'civicrm_contact': | |
76 | $id = 'id'; | |
77 | //we should restrict by contact type in the first step | |
78 | $sql = "SELECT contact_type FROM civicrm_dedupe_rule_group WHERE id = {$this->dedupe_rule_group_id};"; | |
79 | $ct = CRM_Core_DAO::singleValueQuery($sql); | |
80 | if ($this->params) { | |
81 | $where[] = "t1.contact_type = '{$ct}'"; | |
82 | } | |
83 | else { | |
84 | $where[] = "t1.contact_type = '{$ct}'"; | |
85 | $where[] = "t2.contact_type = '{$ct}'"; | |
86 | } | |
87 | break; | |
88 | ||
89 | case 'civicrm_address': | |
90 | case 'civicrm_email': | |
91 | case 'civicrm_im': | |
92 | case 'civicrm_openid': | |
93 | case 'civicrm_phone': | |
94 | $id = 'contact_id'; | |
95 | break; | |
96 | ||
97 | case 'civicrm_note': | |
98 | $id = 'entity_id'; | |
99 | if ($this->params) { | |
100 | $where[] = "t1.entity_table = 'civicrm_contact'"; | |
101 | } | |
102 | else { | |
103 | $where[] = "t1.entity_table = 'civicrm_contact'"; | |
104 | $where[] = "t2.entity_table = 'civicrm_contact'"; | |
105 | } | |
106 | break; | |
107 | ||
108 | default: | |
109 | // custom data tables | |
110 | if (preg_match('/^civicrm_value_/', $this->rule_table) || preg_match('/^custom_value_/', $this->rule_table)) { | |
111 | $id = 'entity_id'; | |
112 | } | |
113 | else { | |
114 | throw new CRM_Core_Exception("Unsupported rule_table for civicrm_dedupe_rule.id of {$this->id}"); | |
115 | } | |
116 | break; | |
117 | } | |
118 | ||
119 | // build SELECT based on the field names containing contact ids | |
120 | // if there are params provided, id1 should be 0 | |
121 | if ($this->params) { | |
122 | $select = "t1.$id id1, {$this->rule_weight} weight"; | |
123 | $subSelect = 'id1, weight'; | |
124 | } | |
125 | else { | |
126 | $select = "t1.$id id1, t2.$id id2, {$this->rule_weight} weight"; | |
127 | $subSelect = 'id1, id2, weight'; | |
128 | } | |
129 | ||
130 | // build FROM (and WHERE, if it's a parametrised search) | |
131 | // based on whether the rule is about substrings or not | |
132 | if ($this->params) { | |
133 | $from = "{$this->rule_table} t1"; | |
134 | $str = 'NULL'; | |
135 | if (isset($this->params[$this->rule_table][$this->rule_field])) { | |
136 | $str = trim(CRM_Utils_Type::escape($this->params[$this->rule_table][$this->rule_field], 'String')); | |
137 | } | |
138 | if ($this->rule_length) { | |
139 | $where[] = "SUBSTR(t1.{$this->rule_field}, 1, {$this->rule_length}) = SUBSTR('$str', 1, {$this->rule_length})"; | |
140 | $where[] = "t1.{$this->rule_field} IS NOT NULL"; | |
141 | } | |
142 | else { | |
143 | $where[] = "t1.{$this->rule_field} = '$str'"; | |
144 | } | |
145 | } | |
146 | else { | |
147 | if ($this->rule_length) { | |
148 | $from = "{$this->rule_table} t1 JOIN {$this->rule_table} t2 ON (" . implode(' AND ', $on) . ")"; | |
149 | } | |
150 | else { | |
151 | $from = "{$this->rule_table} t1 INNER JOIN {$this->rule_table} t2 ON (" . implode(' AND ', $innerJoinClauses) . ")"; | |
152 | } | |
153 | } | |
154 | ||
155 | // finish building WHERE, also limit the results if requested | |
156 | if (!$this->params) { | |
157 | $where[] = "t1.$id < t2.$id"; | |
158 | } | |
159 | $query = "SELECT $select FROM $from WHERE " . implode(' AND ', $where); | |
160 | if ($this->contactIds) { | |
161 | $cids = []; | |
162 | foreach ($this->contactIds as $cid) { | |
163 | $cids[] = CRM_Utils_Type::escape($cid, 'Integer'); | |
164 | } | |
165 | if (count($cids) == 1) { | |
166 | $query .= " AND (t1.$id = {$cids[0]}) UNION $query AND t2.$id = {$cids[0]}"; | |
167 | } | |
168 | else { | |
169 | $query .= " AND t1.$id IN (" . implode(',', $cids) . ") | |
170 | UNION $query AND t2.$id IN (" . implode(',', $cids) . ")"; | |
171 | } | |
172 | // The `weight` is ambiguous in the context of the union; put the whole | |
173 | // thing in a subquery. | |
174 | $query = "SELECT $subSelect FROM ($query) subunion"; | |
175 | } | |
176 | ||
177 | return $query; | |
178 | } | |
179 | ||
180 | /** | |
181 | * find fields related to a rule group. | |
182 | * | |
183 | * @param array $params contains the rule group property to identify rule group | |
184 | * | |
185 | * @return array | |
186 | * rule fields array associated to rule group | |
187 | */ | |
188 | public static function dedupeRuleFields($params) { | |
189 | $rgBao = new CRM_Dedupe_BAO_RuleGroup(); | |
190 | $rgBao->used = $params['used']; | |
191 | $rgBao->contact_type = $params['contact_type']; | |
192 | $rgBao->find(TRUE); | |
193 | ||
194 | $ruleBao = new CRM_Dedupe_BAO_Rule(); | |
195 | $ruleBao->dedupe_rule_group_id = $rgBao->id; | |
196 | $ruleBao->find(); | |
197 | $ruleFields = []; | |
198 | while ($ruleBao->fetch()) { | |
199 | $field_name = $ruleBao->rule_field; | |
200 | if ($field_name == 'phone_numeric') { | |
201 | $field_name = 'phone'; | |
202 | } | |
203 | $ruleFields[] = $field_name; | |
204 | } | |
205 | return $ruleFields; | |
206 | } | |
207 | ||
208 | /** | |
209 | * @param int $cid | |
210 | * @param int $oid | |
211 | * | |
212 | * @return bool | |
213 | */ | |
214 | public static function validateContacts($cid, $oid) { | |
215 | if (!$cid || !$oid) { | |
216 | return NULL; | |
217 | } | |
218 | $exception = new CRM_Dedupe_DAO_Exception(); | |
219 | $exception->contact_id1 = $cid; | |
220 | $exception->contact_id2 = $oid; | |
221 | //make sure contact2 > contact1. | |
222 | if ($cid > $oid) { | |
223 | $exception->contact_id1 = $oid; | |
224 | $exception->contact_id2 = $cid; | |
225 | } | |
226 | ||
227 | return !$exception->find(TRUE); | |
228 | } | |
229 | ||
230 | /** | |
231 | * Get the specification for the given field. | |
232 | * | |
233 | * @param string $fieldName | |
234 | * | |
235 | * @return array | |
236 | * @throws \CiviCRM_API3_Exception | |
237 | */ | |
238 | public function getFieldType($fieldName) { | |
239 | $entity = CRM_Core_DAO_AllCoreTables::getBriefName(CRM_Core_DAO_AllCoreTables::getClassForTable($this->rule_table)); | |
240 | if (!$entity) { | |
241 | // This means we have stored a custom field rather than an entity name in rule_table, figure out the entity. | |
242 | $entity = civicrm_api3('CustomGroup', 'getvalue', ['table_name' => $this->rule_table, 'return' => 'extends']); | |
243 | if (in_array($entity, ['Individual', 'Household', 'Organization'])) { | |
244 | $entity = 'Contact'; | |
245 | } | |
246 | $fieldName = 'custom_' . civicrm_api3('CustomField', 'getvalue', ['column_name' => $fieldName, 'return' => 'id']); | |
247 | } | |
248 | $fields = civicrm_api3($entity, 'getfields', ['action' => 'create'])['values']; | |
249 | return $fields[$fieldName]['type']; | |
250 | } | |
251 | ||
252 | } |