Code Coverage |
||||||||||
Lines |
Functions and Methods |
Classes and Traits |
||||||||
| Total | |
0.00% |
0 / 45 |
|
0.00% |
0 / 6 |
CRAP | |
0.00% |
0 / 1 |
| EntitySearchUtils | |
0.00% |
0 / 45 |
|
0.00% |
0 / 6 |
272 | |
0.00% |
0 / 1 |
| makeConstScoreQuery | |
0.00% |
0 / 8 |
|
0.00% |
0 / 1 |
6 | |||
| normalizeId | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
| normalizeIdFromSearchQuery | |
0.00% |
0 / 9 |
|
0.00% |
0 / 1 |
30 | |||
| parseOrNull | |
0.00% |
0 / 4 |
|
0.00% |
0 / 1 |
6 | |||
| entityIdParserNormalizer | |
0.00% |
0 / 6 |
|
0.00% |
0 / 1 |
6 | |||
| findTermForDisplay | |
0.00% |
0 / 17 |
|
0.00% |
0 / 1 |
20 | |||
| 1 | <?php |
| 2 | namespace Wikibase\Search\Elastic; |
| 3 | |
| 4 | use Elastica\Query\ConstantScore; |
| 5 | use Elastica\Query\MatchQuery; |
| 6 | use Wikibase\DataModel\Entity\EntityId; |
| 7 | use Wikibase\DataModel\Entity\EntityIdParser; |
| 8 | use Wikibase\DataModel\Entity\EntityIdParsingException; |
| 9 | use Wikibase\DataModel\Term\Term; |
| 10 | use Wikibase\Lib\TermLanguageFallbackChain; |
| 11 | |
| 12 | /** |
| 13 | * Utilities useful for entity searches. |
| 14 | */ |
| 15 | final class EntitySearchUtils { |
| 16 | |
| 17 | /** |
| 18 | * Create constant score query for a field. |
| 19 | * @param string $field |
| 20 | * @param string|double $boost |
| 21 | * @param string $text |
| 22 | * @param string $matchOperator |
| 23 | * @return ConstantScore |
| 24 | */ |
| 25 | public static function makeConstScoreQuery( $field, $boost, $text, $matchOperator = MatchQuery::OPERATOR_OR ) { |
| 26 | if ( $matchOperator === MatchQuery::OPERATOR_AND ) { |
| 27 | $filter = new MatchQuery( $field, [ 'query' => $text ] ); |
| 28 | $filter->setFieldOperator( $field, $matchOperator ); |
| 29 | } else { |
| 30 | $filter = new MatchQuery( $field, $text ); |
| 31 | } |
| 32 | |
| 33 | $csquery = new ConstantScore(); |
| 34 | $csquery->setFilter( $filter ); |
| 35 | $csquery->setBoost( $boost ); |
| 36 | return $csquery; |
| 37 | } |
| 38 | |
| 39 | /** |
| 40 | * If the text looks like ID, normalize it to ID title |
| 41 | * Cases handled: |
| 42 | * - q42 |
| 43 | * - (q42) |
| 44 | * - leading/trailing spaces |
| 45 | * - http://www.wikidata.org/entity/Q42 |
| 46 | * @param string $text |
| 47 | * @param EntityIdParser $idParser |
| 48 | * @return string Normalized ID or original string |
| 49 | */ |
| 50 | public static function normalizeId( $text, EntityIdParser $idParser ) { |
| 51 | return self::normalizeIdFromSearchQuery( $text, self::entityIdParserNormalizer( $idParser ) ); |
| 52 | } |
| 53 | |
| 54 | /** |
| 55 | * If the text looks like ID, normalize it to ID title |
| 56 | * Cases handled: |
| 57 | * - q42 |
| 58 | * - (q42) |
| 59 | * - leading/trailing spaces |
| 60 | * - http://www.wikidata.org/entity/Q42 |
| 61 | * @param string $text |
| 62 | * @param callable(string):(string|null) $idNormalizer |
| 63 | * @return string Normalized ID or original string |
| 64 | */ |
| 65 | public static function normalizeIdFromSearchQuery( $text, callable $idNormalizer ) { |
| 66 | // TODO: this is a bit hacky, better way would be to make the field case-insensitive |
| 67 | // or add new subfiled which is case-insensitive |
| 68 | $text = strtoupper( str_replace( [ '(', ')' ], '', trim( $text ) ) ); |
| 69 | $id = $idNormalizer( $text ); |
| 70 | if ( $id !== null ) { |
| 71 | return $id; |
| 72 | } |
| 73 | if ( preg_match( '/\b(\w+)$/', $text, $matches ) && $matches[1] ) { |
| 74 | $id = $idNormalizer( $matches[1] ); |
| 75 | if ( $id !== null ) { |
| 76 | return $id; |
| 77 | } |
| 78 | } |
| 79 | return $text; |
| 80 | } |
| 81 | |
| 82 | /** |
| 83 | * Parse entity ID or return null |
| 84 | * @param string $text |
| 85 | * @param EntityIdParser $idParser |
| 86 | * @return ?EntityId |
| 87 | */ |
| 88 | public static function parseOrNull( $text, EntityIdParser $idParser ): ?EntityId { |
| 89 | try { |
| 90 | $id = $idParser->parse( $text ); |
| 91 | } catch ( EntityIdParsingException ) { |
| 92 | return null; |
| 93 | } |
| 94 | return $id; |
| 95 | } |
| 96 | |
| 97 | /** |
| 98 | * An id normalizer based on a given EntityIdParser. |
| 99 | * |
| 100 | * @param EntityIdParser $idParser |
| 101 | * @return callable(string):(string|null) |
| 102 | */ |
| 103 | public static function entityIdParserNormalizer( EntityIdParser $idParser ): callable { |
| 104 | return static function ( string $text ) use ( $idParser ): ?string { |
| 105 | $id = EntitySearchUtils::parseOrNull( $text, $idParser ); |
| 106 | if ( $id === null ) { |
| 107 | return null; |
| 108 | } |
| 109 | return $id->getSerialization(); |
| 110 | }; |
| 111 | } |
| 112 | |
| 113 | /** |
| 114 | * Locate label for display among the source data, basing on fallback chain. |
| 115 | * @param array $sourceData |
| 116 | * @param string $field |
| 117 | * @param TermLanguageFallbackChain $termFallbackChain |
| 118 | * @return null|Term |
| 119 | */ |
| 120 | public static function findTermForDisplay( $sourceData, $field, TermLanguageFallbackChain $termFallbackChain ) { |
| 121 | if ( empty( $sourceData[$field] ) ) { |
| 122 | return null; |
| 123 | } |
| 124 | |
| 125 | $data = $sourceData[$field]; |
| 126 | $first = reset( $data ); |
| 127 | if ( is_array( $first ) ) { |
| 128 | // If we have multiple, like for labels, extract the first one |
| 129 | $labels_data = array_map( |
| 130 | static function ( $data ) { |
| 131 | return $data[0] ?? null; |
| 132 | }, |
| 133 | $data |
| 134 | ); |
| 135 | } else { |
| 136 | $labels_data = $data; |
| 137 | } |
| 138 | // Drop empty ones |
| 139 | $labels_data = array_filter( $labels_data ); |
| 140 | |
| 141 | $preferredValue = $termFallbackChain->extractPreferredValueOrAny( $labels_data ); |
| 142 | if ( $preferredValue ) { |
| 143 | return new Term( $preferredValue['language'], $preferredValue['value'] ); |
| 144 | } |
| 145 | |
| 146 | return null; |
| 147 | } |
| 148 | |
| 149 | } |