|
1
|
|
|
<?php |
|
2
|
|
|
|
|
3
|
|
|
namespace PPP\Wikidata\TreeSimplifier; |
|
4
|
|
|
|
|
5
|
|
|
use DataValues\TimeValue; |
|
6
|
|
|
use DateInterval; |
|
7
|
|
|
use DateTime; |
|
8
|
|
|
use InvalidArgumentException; |
|
9
|
|
|
use PPP\DataModel\AbstractNode; |
|
10
|
|
|
use PPP\DataModel\IntersectionNode; |
|
11
|
|
|
use PPP\DataModel\JsonLdResourceNode; |
|
12
|
|
|
use PPP\DataModel\MissingNode; |
|
13
|
|
|
use PPP\DataModel\ResourceListNode; |
|
14
|
|
|
use PPP\DataModel\ResourceNode; |
|
15
|
|
|
use PPP\DataModel\StringResourceNode; |
|
16
|
|
|
use PPP\DataModel\TripleNode; |
|
17
|
|
|
use PPP\DataModel\UnionNode; |
|
18
|
|
|
use PPP\Module\TreeSimplifier\NodeSimplifier; |
|
19
|
|
|
use PPP\Wikidata\ValueParsers\ResourceListNodeParser; |
|
20
|
|
|
use PPP\Wikidata\WikibaseResourceNode; |
|
21
|
|
|
use stdClass; |
|
22
|
|
|
use Wikibase\DataModel\Entity\EntityIdValue; |
|
23
|
|
|
use Wikibase\DataModel\Entity\ItemId; |
|
24
|
|
|
use Wikibase\DataModel\Entity\PropertyId; |
|
25
|
|
|
|
|
26
|
|
|
/** |
|
27
|
|
|
* Do some actions for specific use case: |
|
28
|
|
|
* - if a predicate is not useful like "name" or "identity" cast subjects to wikibase items |
|
29
|
|
|
* - if the predicte is son or daughter use "child" with an intersection with the relevant sex |
|
30
|
|
|
* |
|
31
|
|
|
* @licence GPLv2+ |
|
32
|
|
|
* @author Thomas Pellissier Tanon |
|
33
|
|
|
* |
|
34
|
|
|
* |
|
35
|
|
|
* TODO: remove hardcoded values? |
|
36
|
|
|
*/ |
|
37
|
|
|
class SpecificTripleNodeSimplifier implements NodeSimplifier { |
|
38
|
|
|
|
|
39
|
|
|
private static $MEANINGLESS_PREDICATES = array( |
|
40
|
|
|
'name', |
|
41
|
|
|
'identity', |
|
42
|
|
|
'definition' |
|
43
|
|
|
); |
|
44
|
|
|
|
|
45
|
|
|
const PROPERTY_CHILD = 'P40'; |
|
46
|
|
|
const PROPERTY_SEX = 'P21'; |
|
47
|
|
|
const PROPERTY_BIRTH_DATE = 'P569'; |
|
48
|
|
|
const ITEM_MALE = 'Q6581097'; |
|
49
|
|
|
const ITEM_FEMALE = 'Q6581072'; |
|
50
|
|
|
|
|
51
|
|
|
/** |
|
52
|
|
|
* @var ResourceListNodeParser |
|
53
|
|
|
*/ |
|
54
|
|
|
private $resourceListNodeParser; |
|
55
|
|
|
|
|
56
|
|
|
/** |
|
57
|
|
|
* @var ResourceListForEntityProperty |
|
58
|
|
|
*/ |
|
59
|
|
|
private $resourceListForEntityProperty; |
|
60
|
|
|
|
|
61
|
|
|
/** |
|
62
|
|
|
* @var DateTime |
|
63
|
|
|
*/ |
|
64
|
|
|
private $now; |
|
65
|
|
|
|
|
66
|
|
|
/** |
|
67
|
|
|
* @param ResourceListNodeParser $resourceListNodeParser |
|
68
|
|
|
* @param ResourceListForEntityProperty $resourceListForEntityProperty |
|
69
|
|
|
* @param DateTime $now |
|
70
|
|
|
*/ |
|
71
|
11 |
|
public function __construct(ResourceListNodeParser $resourceListNodeParser, ResourceListForEntityProperty $resourceListForEntityProperty, DateTime $now) { |
|
72
|
11 |
|
$this->resourceListNodeParser = $resourceListNodeParser; |
|
73
|
11 |
|
$this->resourceListForEntityProperty = $resourceListForEntityProperty; |
|
74
|
11 |
|
$this->now = $now; |
|
75
|
11 |
|
} |
|
76
|
|
|
|
|
77
|
|
|
/** |
|
78
|
|
|
* @see NodeSimplifier::isSimplifierFor |
|
79
|
|
|
*/ |
|
80
|
10 |
|
public function isSimplifierFor(AbstractNode $node) { |
|
81
|
10 |
|
return $node instanceof TripleNode && |
|
82
|
10 |
|
$node->getSubject() instanceof ResourceListNode && |
|
83
|
10 |
|
$node->getPredicate() instanceof ResourceListNode && |
|
84
|
10 |
|
$node->getObject() instanceof MissingNode; |
|
85
|
|
|
} |
|
86
|
|
|
|
|
87
|
|
|
/** |
|
88
|
|
|
* @see NodeSimplifier::doSimplification |
|
89
|
|
|
*/ |
|
90
|
8 |
|
public function simplify(AbstractNode $node) { |
|
91
|
8 |
|
if(!$this->isSimplifierFor($node)) { |
|
92
|
1 |
|
throw new InvalidArgumentException('SpecificTripleNodeSimplifier can only clean TripleNode objects'); |
|
93
|
|
|
} |
|
94
|
|
|
|
|
95
|
7 |
|
return $this->doSimplification($node); |
|
|
|
|
|
|
96
|
|
|
} |
|
97
|
|
|
|
|
98
|
7 |
|
public function doSimplification(TripleNode $node) { |
|
99
|
7 |
|
$additionalNodes = array(); |
|
100
|
7 |
|
$otherPredicates = array(); |
|
101
|
|
|
|
|
102
|
|
|
/** @var ResourceNode $predicate */ |
|
103
|
7 |
|
foreach($node->getPredicate() as $predicate) { |
|
|
|
|
|
|
104
|
7 |
|
if(in_array($predicate->getValue(), self::$MEANINGLESS_PREDICATES)) { |
|
105
|
3 |
|
$additionalNodes[] = $this->resourceListNodeParser->parse($node->getSubject(), 'wikibase-item'); |
|
|
|
|
|
|
106
|
7 |
|
} else if($predicate->equals(new StringResourceNode('son'))) { |
|
107
|
1 |
|
$additionalNodes[] = $this->buildSonNode($node); |
|
108
|
5 |
|
} else if($predicate->equals(new StringResourceNode('daughter'))) { |
|
109
|
1 |
|
$additionalNodes[] = $this->buildDaughterNode($node); |
|
110
|
4 |
|
} else if($predicate->equals(new StringResourceNode('age'))) { |
|
111
|
1 |
|
$additionalNodes[] = $this->buildAgeNode($node); |
|
112
|
1 |
|
} else { |
|
113
|
2 |
|
$otherPredicates[] = $predicate; |
|
114
|
|
|
} |
|
115
|
7 |
|
} |
|
116
|
|
|
|
|
117
|
7 |
|
if(!empty($otherPredicates)) { |
|
118
|
2 |
|
$additionalNodes[] = new TripleNode($node->getSubject(), new ResourceListNode($otherPredicates), $node->getObject()); |
|
119
|
2 |
|
} |
|
120
|
|
|
|
|
121
|
7 |
|
if(count($additionalNodes) === 1) { |
|
122
|
6 |
|
return $additionalNodes[0]; |
|
123
|
|
|
} |
|
124
|
|
|
|
|
125
|
1 |
|
return new UnionNode($additionalNodes); |
|
126
|
|
|
} |
|
127
|
|
|
|
|
128
|
1 |
|
private function buildSonNode(TripleNode $node) { |
|
129
|
1 |
|
return new IntersectionNode(array( |
|
130
|
1 |
|
new TripleNode( |
|
131
|
1 |
|
$node->getSubject(), |
|
132
|
1 |
|
new ResourceListNode(array(new WikibaseResourceNode('', new EntityIdValue(new PropertyId(self::PROPERTY_CHILD))))), |
|
133
|
1 |
|
$node->getObject() |
|
134
|
1 |
|
), |
|
135
|
1 |
|
new TripleNode( |
|
136
|
1 |
|
$node->getObject(), |
|
137
|
1 |
|
new ResourceListNode(array(new WikibaseResourceNode('', new EntityIdValue(new PropertyId(self::PROPERTY_SEX))))), |
|
138
|
1 |
|
new ResourceListNode(array(new WikibaseResourceNode('', new EntityIdValue(new ItemId(self::ITEM_MALE))))) |
|
139
|
1 |
|
), |
|
140
|
1 |
|
)); |
|
141
|
|
|
} |
|
142
|
|
|
|
|
143
|
1 |
|
private function buildDaughterNode(TripleNode $node) { |
|
144
|
1 |
|
return new IntersectionNode(array( |
|
145
|
1 |
|
new TripleNode( |
|
146
|
1 |
|
$node->getSubject(), |
|
147
|
1 |
|
new ResourceListNode(array(new WikibaseResourceNode('', new EntityIdValue(new PropertyId(self::PROPERTY_CHILD))))), |
|
148
|
1 |
|
$node->getObject() |
|
149
|
1 |
|
), |
|
150
|
1 |
|
new TripleNode( |
|
151
|
1 |
|
$node->getObject(), |
|
152
|
1 |
|
new ResourceListNode(array(new WikibaseResourceNode('', new EntityIdValue(new PropertyId(self::PROPERTY_SEX))))), |
|
153
|
1 |
|
new ResourceListNode(array(new WikibaseResourceNode('', new EntityIdValue(new ItemId(self::ITEM_FEMALE))))) |
|
154
|
1 |
|
), |
|
155
|
1 |
|
)); |
|
156
|
|
|
} |
|
157
|
|
|
|
|
158
|
1 |
|
private function buildAgeNode(TripleNode $node) { |
|
159
|
1 |
|
$subjectNodes = $this->resourceListNodeParser->parse($node->getSubject(), 'wikibase-item'); |
|
|
|
|
|
|
160
|
1 |
|
$ageValues = array(); |
|
161
|
|
|
|
|
162
|
|
|
/** @var WikibaseResourceNode $subjectNode */ |
|
163
|
1 |
|
foreach($subjectNodes as $subjectNode) { |
|
164
|
1 |
|
$birthDateObjects = $this->resourceListForEntityProperty->getForEntityProperty( |
|
165
|
1 |
|
$subjectNode->getDataValue()->getEntityId(), |
|
166
|
1 |
|
new PropertyId(self::PROPERTY_BIRTH_DATE) |
|
167
|
1 |
|
); |
|
168
|
|
|
|
|
169
|
|
|
/** @var WikibaseResourceNode $birthDateObject */ |
|
170
|
1 |
|
foreach($birthDateObjects as $birthDateObject) { |
|
171
|
1 |
|
$ageValues[] = $this->formatAgeValue($this->computeAge($birthDateObject->getDataValue())); |
|
172
|
1 |
|
} |
|
173
|
1 |
|
} |
|
174
|
|
|
|
|
175
|
1 |
|
return new ResourceListNode($ageValues); |
|
176
|
|
|
} |
|
177
|
|
|
|
|
178
|
1 |
|
private function computeAge(TimeValue $birthDateValue) { |
|
179
|
1 |
|
$birthDate = new DateTime(preg_replace('/^\+0*/', '', $birthDateValue->getTime())); |
|
180
|
1 |
|
return $this->now->diff($birthDate); |
|
181
|
|
|
} |
|
182
|
|
|
|
|
183
|
|
|
//TODO: implement a DurationValue? |
|
184
|
1 |
|
private function formatAgeValue(DateInterval $age) { |
|
185
|
1 |
|
$formattedAge = $age->format('%y'); |
|
186
|
|
|
|
|
187
|
1 |
|
$literal = new stdClass(); |
|
188
|
1 |
|
$literal->{'@type'} = 'Duration'; |
|
189
|
1 |
|
$literal->{'@value'} = $age->format('P%yY%mM%dD'); |
|
190
|
|
|
|
|
191
|
1 |
|
$resource = new stdClass(); |
|
192
|
1 |
|
$resource->{'@context'} = 'http://schema.org'; |
|
193
|
1 |
|
$resource->{'@type'} = 'Duration'; |
|
194
|
1 |
|
$resource->{'name'} = $formattedAge; |
|
195
|
1 |
|
$resource->{'http://www.w3.org/1999/02/22-rdf-syntax-ns#value'} = $literal; |
|
196
|
|
|
|
|
197
|
1 |
|
return new JsonLdResourceNode($formattedAge, $resource); |
|
198
|
|
|
} |
|
199
|
|
|
} |
|
200
|
|
|
|
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.