|
1
|
|
|
<?php |
|
2
|
|
|
|
|
3
|
|
|
namespace Wikibase\DataModel\Services\Lookup; |
|
4
|
|
|
|
|
5
|
|
|
use Wikibase\DataModel\Entity\EntityId; |
|
6
|
|
|
use Wikibase\DataModel\Entity\EntityIdValue; |
|
7
|
|
|
use Wikibase\DataModel\Entity\PropertyId; |
|
8
|
|
|
use Wikibase\DataModel\Services\Entity\EntityPrefetcher; |
|
9
|
|
|
use Wikibase\DataModel\Snak\PropertyValueSnak; |
|
10
|
|
|
use Wikibase\DataModel\Snak\Snak; |
|
11
|
|
|
use Wikibase\DataModel\Statement\StatementListProvider; |
|
12
|
|
|
|
|
13
|
|
|
/** |
|
14
|
|
|
* Service for getting the closest entity (out of a specified set), |
|
15
|
|
|
* from a given starting entity. The starting entity, and the target entities |
|
16
|
|
|
* are (potentially indirectly, via intermediate entities) linked by statements |
|
17
|
|
|
* with a given property ID, pointing from the starting entity to one of the |
|
18
|
|
|
* target entities. |
|
19
|
|
|
* |
|
20
|
|
|
* @since 3.10 |
|
21
|
|
|
* |
|
22
|
|
|
* @license GPL-2.0-or-later |
|
23
|
|
|
* @author Marius Hoch |
|
24
|
|
|
*/ |
|
25
|
|
|
class EntityRetrievingClosestReferencedEntityIdLookup implements ReferencedEntityIdLookup { |
|
26
|
|
|
|
|
27
|
|
|
/** |
|
28
|
|
|
* @var EntityLookup |
|
29
|
|
|
*/ |
|
30
|
|
|
private $entityLookup; |
|
31
|
|
|
|
|
32
|
|
|
/** |
|
33
|
|
|
* @var EntityPrefetcher |
|
34
|
|
|
*/ |
|
35
|
|
|
private $entityPrefetcher; |
|
36
|
|
|
|
|
37
|
|
|
/** |
|
38
|
|
|
* @var int Maximum search depth: Maximum number of intermediate entities to search through. |
|
39
|
|
|
* For example 0 means that only the entities immediately referenced will be found. |
|
40
|
|
|
*/ |
|
41
|
|
|
private $maxDepth; |
|
42
|
|
|
|
|
43
|
|
|
/** |
|
44
|
|
|
* @var int Maximum number of entities to retrieve. |
|
45
|
|
|
*/ |
|
46
|
|
|
private $maxEntityVisits; |
|
47
|
|
|
|
|
48
|
|
|
/** |
|
49
|
|
|
* Map (entity id => true) of already visited entities. |
|
50
|
|
|
* |
|
51
|
|
|
* @var bool[] |
|
52
|
|
|
*/ |
|
53
|
|
|
private $alreadyVisited = []; |
|
54
|
|
|
|
|
55
|
|
|
/** |
|
56
|
|
|
* @param EntityLookup $entityLookup |
|
57
|
|
|
* @param EntityPrefetcher $entityPrefetcher |
|
58
|
|
|
* @param int $maxDepth Maximum search depth: Maximum number of intermediate entities to search through. |
|
59
|
|
|
* For example if 0 is given, only the entities immediately referenced will be found. |
|
60
|
|
|
* If this limit gets exhausted, a MaxReferenceDepthExhaustedException is thrown. |
|
61
|
|
|
* @param int $maxEntityVisits Maximum number of entities to retrieve during a lookup. |
|
62
|
|
|
* If this limit gets exhausted, a MaxReferencedEntityVisitsExhaustedException is thrown. |
|
63
|
|
|
*/ |
|
64
|
|
|
public function __construct( |
|
65
|
|
|
EntityLookup $entityLookup, |
|
66
|
|
|
EntityPrefetcher $entityPrefetcher, |
|
67
|
|
|
$maxDepth, |
|
68
|
|
|
$maxEntityVisits |
|
69
|
|
|
) { |
|
70
|
|
|
$this->entityLookup = $entityLookup; |
|
71
|
|
|
$this->entityPrefetcher = $entityPrefetcher; |
|
72
|
|
|
$this->maxDepth = $maxDepth; |
|
73
|
|
|
$this->maxEntityVisits = $maxEntityVisits; |
|
74
|
|
|
} |
|
75
|
|
|
|
|
76
|
|
|
/** |
|
77
|
|
|
* Get the closest entity (out of $toIds), from a given entity. The starting entity, and |
|
78
|
|
|
* the target entities are (potentially indirectly, via intermediate entities) linked by |
|
79
|
|
|
* statements with the given property ID, pointing from the starting entity to one of the |
|
80
|
|
|
* target entities. |
|
81
|
|
|
* |
|
82
|
|
|
* @since 3.10 |
|
83
|
|
|
* |
|
84
|
|
|
* @param EntityId $fromId |
|
85
|
|
|
* @param PropertyId $propertyId |
|
86
|
|
|
* @param EntityId[] $toIds |
|
87
|
|
|
* |
|
88
|
|
|
* @return EntityId|null Returns null in case none of the target entities are referenced. |
|
89
|
|
|
* @throws ReferencedEntityIdLookupException |
|
90
|
|
|
*/ |
|
91
|
|
|
public function getReferencedEntityId( EntityId $fromId, PropertyId $propertyId, array $toIds ) { |
|
92
|
|
|
if ( !$toIds ) { |
|
|
|
|
|
|
93
|
|
|
return null; |
|
94
|
|
|
} |
|
95
|
|
|
|
|
96
|
|
|
$this->alreadyVisited = []; |
|
97
|
|
|
|
|
98
|
|
|
$steps = $this->maxDepth + 1; // Add one as checking $fromId already is a step |
|
99
|
|
|
$toVisit = [ $fromId ]; |
|
100
|
|
|
|
|
101
|
|
|
while ( $steps-- ) { |
|
102
|
|
|
$this->entityPrefetcher->prefetch( $toVisit ); |
|
103
|
|
|
$toVisitNext = []; |
|
104
|
|
|
|
|
105
|
|
|
foreach ( $toVisit as $curId ) { |
|
106
|
|
|
$result = $this->processEntityById( $curId, $fromId, $propertyId, $toIds, $toVisitNext ); |
|
|
|
|
|
|
107
|
|
|
if ( $result ) { |
|
108
|
|
|
return $result; |
|
109
|
|
|
} |
|
110
|
|
|
} |
|
111
|
|
|
// Remove already visited entities |
|
112
|
|
|
$toVisit = array_unique( |
|
113
|
|
|
array_diff( $toVisitNext, array_keys( $this->alreadyVisited ) ) |
|
114
|
|
|
); |
|
115
|
|
|
|
|
116
|
|
|
if ( !$toVisit ) { |
|
|
|
|
|
|
117
|
|
|
return null; |
|
118
|
|
|
} |
|
119
|
|
|
} |
|
120
|
|
|
|
|
121
|
|
|
// Exhausted the max. depth without finding anything. |
|
122
|
|
|
throw new MaxReferenceDepthExhaustedException( |
|
123
|
|
|
$fromId, |
|
124
|
|
|
$propertyId, |
|
125
|
|
|
$toIds, |
|
126
|
|
|
$this->maxDepth |
|
127
|
|
|
); |
|
128
|
|
|
} |
|
129
|
|
|
|
|
130
|
|
|
/** |
|
131
|
|
|
* Find out whether an entity (directly) references one of the target ids. |
|
132
|
|
|
* |
|
133
|
|
|
* @param EntityId $id Id of the entity to process |
|
134
|
|
|
* @param EntityId $fromId Id this lookup started from |
|
135
|
|
|
* @param PropertyId $propertyId |
|
136
|
|
|
* @param EntityId[] $toIds |
|
137
|
|
|
* @param EntityId[] &$toVisit List of entities that still need to be checked |
|
138
|
|
|
* @return EntityId|null Target id the entity refers to, null if none. |
|
139
|
|
|
*/ |
|
140
|
|
|
private function processEntityById( |
|
141
|
|
|
EntityId $id, |
|
142
|
|
|
EntityId $fromId, |
|
143
|
|
|
PropertyId $propertyId, |
|
144
|
|
|
array $toIds, |
|
145
|
|
|
array &$toVisit |
|
146
|
|
|
) { |
|
147
|
|
|
$entity = $this->getEntity( $id, $fromId, $propertyId, $toIds ); |
|
148
|
|
|
if ( !$entity ) { |
|
149
|
|
|
return null; |
|
150
|
|
|
} |
|
151
|
|
|
|
|
152
|
|
|
$mainSnaks = $this->getMainSnaks( $entity, $propertyId ); |
|
153
|
|
|
|
|
154
|
|
|
foreach ( $mainSnaks as $mainSnak ) { |
|
155
|
|
|
$result = $this->processSnak( $mainSnak, $toVisit, $toIds ); |
|
|
|
|
|
|
156
|
|
|
if ( $result ) { |
|
157
|
|
|
return $result; |
|
158
|
|
|
} |
|
159
|
|
|
} |
|
160
|
|
|
|
|
161
|
|
|
return null; |
|
162
|
|
|
} |
|
163
|
|
|
|
|
164
|
|
|
/** |
|
165
|
|
|
* @param EntityId $id Id of the entity to get |
|
166
|
|
|
* @param EntityId $fromId Id this lookup started from |
|
167
|
|
|
* @param PropertyId $propertyId |
|
168
|
|
|
* @param EntityId[] $toIds |
|
169
|
|
|
* |
|
170
|
|
|
* @return StatementListProvider|null Null if not applicable. |
|
171
|
|
|
*/ |
|
172
|
|
|
private function getEntity( EntityId $id, EntityId $fromId, PropertyId $propertyId, array $toIds ) { |
|
173
|
|
|
if ( isset( $this->alreadyVisited[$id->getSerialization()] ) ) { |
|
174
|
|
|
trigger_error( |
|
175
|
|
|
'Entity ' . $id->getSerialization() . ' already visited.', |
|
176
|
|
|
E_USER_WARNING |
|
177
|
|
|
); |
|
178
|
|
|
|
|
179
|
|
|
return null; |
|
180
|
|
|
} |
|
181
|
|
|
|
|
182
|
|
|
$this->alreadyVisited[$id->getSerialization()] = true; |
|
183
|
|
|
|
|
184
|
|
|
if ( count( $this->alreadyVisited ) > $this->maxEntityVisits ) { |
|
185
|
|
|
throw new MaxReferencedEntityVisitsExhaustedException( |
|
186
|
|
|
$fromId, |
|
187
|
|
|
$propertyId, |
|
188
|
|
|
$toIds, |
|
189
|
|
|
$this->maxEntityVisits |
|
190
|
|
|
); |
|
191
|
|
|
} |
|
192
|
|
|
|
|
193
|
|
|
try { |
|
194
|
|
|
$entity = $this->entityLookup->getEntity( $id ); |
|
195
|
|
|
} catch ( EntityLookupException $ex ) { |
|
196
|
|
|
throw new ReferencedEntityIdLookupException( $fromId, $propertyId, $toIds, null, $ex ); |
|
197
|
|
|
} |
|
198
|
|
|
|
|
199
|
|
|
if ( !( $entity instanceof StatementListProvider ) ) { |
|
200
|
|
|
return null; |
|
201
|
|
|
} |
|
202
|
|
|
|
|
203
|
|
|
return $entity; |
|
204
|
|
|
} |
|
205
|
|
|
|
|
206
|
|
|
/** |
|
207
|
|
|
* Decide whether a single Snak is pointing to one of the target ids. |
|
208
|
|
|
* |
|
209
|
|
|
* @param Snak $snak |
|
210
|
|
|
* @param EntityId[] &$toVisit List of entities that still need to be checked |
|
211
|
|
|
* @param EntityId[] $toIds |
|
212
|
|
|
* @return EntityId|null Target id the Snak refers to, null if none. |
|
213
|
|
|
*/ |
|
214
|
|
|
private function processSnak( Snak $snak, array &$toVisit, array $toIds ) { |
|
215
|
|
|
if ( ! ( $snak instanceof PropertyValueSnak ) ) { |
|
216
|
|
|
return null; |
|
217
|
|
|
} |
|
218
|
|
|
$dataValue = $snak->getDataValue(); |
|
219
|
|
|
if ( ! ( $dataValue instanceof EntityIdValue ) ) { |
|
220
|
|
|
return null; |
|
221
|
|
|
} |
|
222
|
|
|
|
|
223
|
|
|
$entityId = $dataValue->getEntityId(); |
|
224
|
|
|
if ( in_array( $entityId, $toIds, false ) ) { |
|
225
|
|
|
return $entityId; |
|
226
|
|
|
} |
|
227
|
|
|
|
|
228
|
|
|
$toVisit[] = $entityId; |
|
229
|
|
|
|
|
230
|
|
|
return null; |
|
231
|
|
|
} |
|
232
|
|
|
|
|
233
|
|
|
/** |
|
234
|
|
|
* @param StatementListProvider $statementListProvider |
|
235
|
|
|
* @param PropertyId $propertyId |
|
236
|
|
|
* @return Snak[] |
|
237
|
|
|
*/ |
|
238
|
|
|
private function getMainSnaks( |
|
239
|
|
|
StatementListProvider $statementListProvider, |
|
240
|
|
|
PropertyId $propertyId |
|
241
|
|
|
) { |
|
242
|
|
|
return $statementListProvider |
|
243
|
|
|
->getStatements() |
|
244
|
|
|
->getByPropertyId( $propertyId ) |
|
245
|
|
|
->getBestStatements() |
|
246
|
|
|
->getMainSnaks(); |
|
247
|
|
|
} |
|
248
|
|
|
|
|
249
|
|
|
} |
|
250
|
|
|
|
This check marks implicit conversions of arrays to boolean values in a comparison. While in PHP an empty array is considered to be equal (but not identical) to false, this is not always apparent.
Consider making the comparison explicit by using
empty(..)or! empty(...)instead.