|
1
|
|
|
<?php |
|
2
|
|
|
|
|
3
|
|
|
namespace PPP\Wikidata\TreeSimplifier; |
|
4
|
|
|
|
|
5
|
|
|
use InvalidArgumentException; |
|
6
|
|
|
use PPP\DataModel\AbstractNode; |
|
7
|
|
|
use PPP\DataModel\IntersectionNode; |
|
8
|
|
|
use PPP\DataModel\MissingNode; |
|
9
|
|
|
use PPP\DataModel\ResourceListNode; |
|
10
|
|
|
use PPP\DataModel\TripleNode; |
|
11
|
|
|
use PPP\Module\TreeSimplifier\NodeSimplifier; |
|
12
|
|
|
use PPP\Module\TreeSimplifier\NodeSimplifierFactory; |
|
13
|
|
|
use PPP\Wikidata\ValueParsers\ResourceListNodeParser; |
|
14
|
|
|
use PPP\Wikidata\WikibaseResourceNode; |
|
15
|
|
|
use Wikibase\DataModel\Entity\ItemId; |
|
16
|
|
|
use Wikibase\DataModel\Snak\PropertyValueSnak; |
|
17
|
|
|
use Wikibase\DataModel\Snak\SnakList; |
|
18
|
|
|
use Wikibase\DataModel\Statement\Statement; |
|
19
|
|
|
use Wikibase\EntityStore\EntityStore; |
|
20
|
|
|
|
|
21
|
|
|
/** |
|
22
|
|
|
* Simplifies cases like [Q42, Q43] ∩ (?, instanceof, human) without requests to WikidataQuery: just check that the items |
|
23
|
|
|
* have PropertyValueSnak(instanceof, human). |
|
24
|
|
|
* |
|
25
|
|
|
* @licence AGPLv3+ |
|
26
|
|
|
* @author Thomas Pellissier Tanon |
|
27
|
|
|
*/ |
|
28
|
|
|
class IntersectionWithFilterNodeSimplifier implements NodeSimplifier { |
|
29
|
|
|
|
|
30
|
|
|
/** |
|
31
|
|
|
* @var NodeSimplifierFactory |
|
32
|
|
|
*/ |
|
33
|
|
|
private $nodeSimplifierFactory; |
|
34
|
|
|
|
|
35
|
|
|
/** |
|
36
|
|
|
* @var EntityStore |
|
37
|
|
|
*/ |
|
38
|
|
|
private $entityStore; |
|
39
|
|
|
|
|
40
|
|
|
/** |
|
41
|
|
|
* @var ResourceListNodeParser |
|
42
|
|
|
*/ |
|
43
|
|
|
private $resourceListNodeParser; |
|
44
|
|
|
|
|
45
|
|
|
/** |
|
46
|
|
|
* @param NodeSimplifierFactory $nodeSimplifierFactory |
|
47
|
|
|
* @param EntityStore $entityStore |
|
48
|
|
|
* @param ResourceListNodeParser $resourceListNodeParser |
|
49
|
|
|
*/ |
|
50
|
|
|
public function __construct(NodeSimplifierFactory $nodeSimplifierFactory, EntityStore $entityStore, ResourceListNodeParser $resourceListNodeParser) { |
|
51
|
|
|
$this->nodeSimplifierFactory = $nodeSimplifierFactory; |
|
52
|
|
|
$this->entityStore = $entityStore; |
|
53
|
|
|
$this->resourceListNodeParser = $resourceListNodeParser; |
|
54
|
|
|
} |
|
55
|
|
|
|
|
56
|
|
|
/** |
|
57
|
|
|
* @see AbstractNode::isSimplifierFor |
|
58
|
|
|
*/ |
|
59
|
|
|
public function isSimplifierFor(AbstractNode $node) { |
|
60
|
|
|
return $node instanceof IntersectionNode; |
|
61
|
|
|
} |
|
62
|
|
|
|
|
63
|
|
|
/** |
|
64
|
|
|
* @see NodeSimplifier::doSimplification |
|
65
|
|
|
*/ |
|
66
|
|
|
public function simplify(AbstractNode $node) { |
|
67
|
|
|
if(!$this->isSimplifierFor($node)) { |
|
68
|
|
|
throw new InvalidArgumentException('IntersectionWithFilterNodeSimplifier can only simplify IntersectionNode'); |
|
69
|
|
|
} |
|
70
|
|
|
|
|
71
|
|
|
return $this->doSimplification($node); |
|
|
|
|
|
|
72
|
|
|
} |
|
73
|
|
|
|
|
74
|
|
|
private function doSimplification(IntersectionNode $node) { |
|
75
|
|
|
$triplesWithMissingSubjects = array(); |
|
76
|
|
|
$otherOperands = array(); |
|
77
|
|
|
|
|
78
|
|
|
foreach($node->getOperands() as $operand) { |
|
79
|
|
|
if($this->isTripleWithMissingSubject($operand)) { |
|
80
|
|
|
$triplesWithMissingSubjects[] = $operand; |
|
81
|
|
|
} else { |
|
82
|
|
|
$otherOperands[] = $operand; |
|
83
|
|
|
} |
|
84
|
|
|
} |
|
85
|
|
|
|
|
86
|
|
|
if(empty($otherOperands) || empty($triplesWithMissingSubjects)) { |
|
87
|
|
|
return $node; //case of the MissingSubjectNodeSimplifier |
|
88
|
|
|
} |
|
89
|
|
|
|
|
90
|
|
|
$baseList = $this->nodeSimplifierFactory->newNodeSimplifier()->simplify(new IntersectionNode($otherOperands)); |
|
91
|
|
|
|
|
92
|
|
|
if(!($baseList instanceof ResourceListNode)) { |
|
93
|
|
|
$triplesWithMissingSubjects[] = $baseList; |
|
94
|
|
|
return new IntersectionNode($triplesWithMissingSubjects); |
|
95
|
|
|
} |
|
96
|
|
|
$baseList = $this->resourceListNodeParser->parse($baseList, 'wikibase-item'); |
|
97
|
|
|
|
|
98
|
|
|
foreach($triplesWithMissingSubjects as $tripleWithMissingSubject) { |
|
99
|
|
|
$baseList = $this->applyTripleAsFilter($baseList, $tripleWithMissingSubject); |
|
100
|
|
|
} |
|
101
|
|
|
|
|
102
|
|
|
return $baseList; |
|
103
|
|
|
} |
|
104
|
|
|
|
|
105
|
|
|
private function isTripleWithMissingSubject(AbstractNode $node) { |
|
106
|
|
|
return $node instanceof TripleNode && |
|
107
|
|
|
$node->getSubject() instanceof MissingNode && |
|
108
|
|
|
$node->getPredicate() instanceof ResourceListNode && |
|
109
|
|
|
$node->getObject() instanceof ResourceListNode; |
|
110
|
|
|
} |
|
111
|
|
|
|
|
112
|
|
|
private function applyTripleAsFilter(ResourceListNode $baseList, TripleNode $triple) { |
|
113
|
|
|
$possibleSnaks = $this->buildPossibleSnaks($triple); |
|
114
|
|
|
$filteredList = array(); |
|
115
|
|
|
|
|
116
|
|
|
foreach($baseList as $resource) { |
|
117
|
|
|
if($this->isOneOfSnakInItem($resource->getDataValue()->getEntityId(), $possibleSnaks)) { |
|
118
|
|
|
$filteredList[] = $resource; |
|
119
|
|
|
} |
|
120
|
|
|
} |
|
121
|
|
|
|
|
122
|
|
|
return new ResourceListNode($filteredList); |
|
123
|
|
|
} |
|
124
|
|
|
|
|
125
|
|
|
private function isOneOfSnakInItem(ItemId $itemId, SnakList $snaks) { |
|
126
|
|
|
$item = $this->entityStore->getItemLookup()->getItemForId($itemId); |
|
127
|
|
|
|
|
128
|
|
|
/** @var Statement $statement */ |
|
129
|
|
|
foreach($item->getStatements() as $statement) { |
|
130
|
|
|
if($snaks->hasSnak($statement->getMainSnak())) { |
|
131
|
|
|
return true; |
|
132
|
|
|
} |
|
133
|
|
|
} |
|
134
|
|
|
|
|
135
|
|
|
return false; |
|
136
|
|
|
} |
|
137
|
|
|
|
|
138
|
|
|
private function buildPossibleSnaks(TripleNode $triple) { |
|
139
|
|
|
$propertyNodes = $this->resourceListNodeParser->parse($triple->getPredicate(), 'wikibase-property'); |
|
|
|
|
|
|
140
|
|
|
$possibleSnaks = new SnakList(); |
|
141
|
|
|
|
|
142
|
|
|
foreach($this->bagsPropertiesPerType($propertyNodes) as $objectType => $propertyNodes) { |
|
143
|
|
|
$objectNodes = $this->resourceListNodeParser->parse($triple->getObject(), $objectType); |
|
|
|
|
|
|
144
|
|
|
|
|
145
|
|
|
foreach($propertyNodes as $property) { |
|
146
|
|
|
foreach($objectNodes as $object) { |
|
147
|
|
|
$possibleSnaks->addSnak(new PropertyValueSnak($property->getDataValue()->getEntityId(), $object->getDataValue())); |
|
148
|
|
|
} |
|
149
|
|
|
} |
|
150
|
|
|
} |
|
151
|
|
|
|
|
152
|
|
|
return $possibleSnaks; |
|
153
|
|
|
} |
|
154
|
|
|
|
|
155
|
|
|
private function bagsPropertiesPerType($propertyNodes) { |
|
156
|
|
|
$propertyNodesPerType = array(); |
|
157
|
|
|
|
|
158
|
|
|
/** @var WikibaseResourceNode $propertyNode */ |
|
159
|
|
|
foreach($propertyNodes as $propertyNode) { |
|
160
|
|
|
$objectType = $this->entityStore->getPropertyLookup()->getPropertyForId( |
|
161
|
|
|
$propertyNode->getDataValue()->getEntityId() |
|
162
|
|
|
)->getDataTypeId(); |
|
163
|
|
|
$propertyNodesPerType[$objectType][] = $propertyNode; |
|
164
|
|
|
} |
|
165
|
|
|
|
|
166
|
|
|
return $propertyNodesPerType; |
|
167
|
|
|
} |
|
168
|
|
|
} |
|
169
|
|
|
|
ProjetPP /
PPP-Wikidata
Loading history...
This check looks for parameters that are defined as one type in their type hint or doc comment but seem to be used as a narrower type, i.e an implementation of an interface or a subclass.
Consider changing the type of the parameter or doing an instanceof check before assuming your parameter is of the expected type.