Code Coverage |
||||||||||
Lines |
Functions and Methods |
Classes and Traits |
||||||||
Total | |
70.91% |
39 / 55 |
|
20.00% |
1 / 5 |
CRAP | |
0.00% |
0 / 1 |
MoreLikeTrait | |
70.91% |
39 / 55 |
|
20.00% |
1 / 5 |
25.98 | |
0.00% |
0 / 1 |
doExpand | |
71.43% |
5 / 7 |
|
0.00% |
0 / 1 |
3.21 | |||
collectTitles | |
75.00% |
3 / 4 |
|
0.00% |
0 / 1 |
2.06 | |||
collectTitlesFromElastic | |
0.00% |
0 / 7 |
|
0.00% |
0 / 1 |
12 | |||
collectTitlesFromDB | |
73.91% |
17 / 23 |
|
0.00% |
0 / 1 |
9.14 | |||
buildMoreLikeQuery | |
100.00% |
14 / 14 |
|
100.00% |
1 / 1 |
2 | |||
getConfig | n/a |
0 / 0 |
n/a |
0 / 0 |
0 |
1 | <?php |
2 | |
3 | namespace CirrusSearch\Query; |
4 | |
5 | use CirrusSearch\Hooks; |
6 | use CirrusSearch\SearchConfig; |
7 | use CirrusSearch\WarningCollector; |
8 | use Elastica\Query\MoreLikeThis; |
9 | use MediaWiki\MediaWikiServices; |
10 | use MediaWiki\Title\Title; |
11 | |
12 | trait MoreLikeTrait { |
13 | /** |
14 | * @param string $key |
15 | * @param string $term |
16 | * @param WarningCollector $warningCollector |
17 | * @return Title[] |
18 | */ |
19 | protected function doExpand( $key, $term, WarningCollector $warningCollector ) { |
20 | // If no fields have been set we return no results. This can happen if |
21 | // the user override this setting with field names that are not allowed |
22 | // in $this->getConfig()->get( 'CirrusSearchMoreLikeThisAllowedFields' ) |
23 | // (see Hooks.php) |
24 | if ( !$this->getConfig()->get( 'CirrusSearchMoreLikeThisFields' ) ) { |
25 | $warningCollector->addWarning( "cirrussearch-mlt-not-configured", $key ); |
26 | return []; |
27 | } |
28 | $titles = $this->collectTitles( $term ); |
29 | if ( $titles === [] ) { |
30 | $warningCollector->addWarning( "cirrussearch-mlt-feature-no-valid-titles", $key ); |
31 | } |
32 | return $titles; |
33 | } |
34 | |
35 | /** |
36 | * @param string $term |
37 | * @return Title[] |
38 | */ |
39 | private function collectTitles( $term ) { |
40 | if ( $this->getConfig()->getElement( 'CirrusSearchDevelOptions', |
41 | 'morelike_collect_titles_from_elastic' ) |
42 | ) { |
43 | return $this->collectTitlesFromElastic( $term ); |
44 | } else { |
45 | return $this->collectTitlesFromDB( $term ); |
46 | } |
47 | } |
48 | |
49 | /** |
50 | * Use for devel purpose only |
51 | * @param string $terms |
52 | * @return Title[] |
53 | */ |
54 | private function collectTitlesFromElastic( $terms ) { |
55 | $titles = []; |
56 | foreach ( explode( '|', $terms ) as $term ) { |
57 | $title = null; |
58 | Hooks::handleSearchGetNearMatch( $term, $title ); |
59 | if ( $title != null ) { |
60 | $titles[] = $title; |
61 | } |
62 | } |
63 | return $titles; |
64 | } |
65 | |
66 | /** |
67 | * @param string $term |
68 | * @return Title[] |
69 | */ |
70 | private function collectTitlesFromDB( $term ) { |
71 | $titles = []; |
72 | $found = []; |
73 | $titleFactory = MediaWikiServices::getInstance()->getTitleFactory(); |
74 | $wikiPageFactory = MediaWikiServices::getInstance()->getWikiPageFactory(); |
75 | foreach ( explode( '|', $term ) as $title ) { |
76 | $title = $titleFactory->newFromText( trim( $title ) ); |
77 | while ( true ) { |
78 | if ( !$title ) { |
79 | continue 2; |
80 | } |
81 | $titleText = $title->getFullText(); |
82 | if ( isset( $found[$titleText] ) ) { |
83 | continue 2; |
84 | } |
85 | $found[$titleText] = true; |
86 | if ( !$title->exists() ) { |
87 | continue 2; |
88 | } |
89 | if ( !$title->isRedirect() ) { |
90 | break; |
91 | } |
92 | // If the page was a redirect loop the while( true ) again. |
93 | $page = $wikiPageFactory->newFromTitle( $title ); |
94 | if ( !$page->exists() ) { |
95 | continue 2; |
96 | } |
97 | $title = $page->getRedirectTarget(); |
98 | } |
99 | $titles[] = $title; |
100 | } |
101 | |
102 | return $titles; |
103 | } |
104 | |
105 | /** |
106 | * Builds a more like this query for the specified titles. Take care that |
107 | * this outputs a stable result, regardless of order of configuration |
108 | * parameters and input titles. The result of this is hashed to generate an |
109 | * application side cache key. If the result is unstable we will see a |
110 | * reduced hit rate, and waste cache storage space. |
111 | * |
112 | * @param Title[] $titles |
113 | * @return MoreLikeThis |
114 | */ |
115 | protected function buildMoreLikeQuery( array $titles ) { |
116 | sort( $titles, SORT_STRING ); |
117 | $docIds = []; |
118 | $likeDocs = []; |
119 | foreach ( $titles as $title ) { |
120 | $docId = $this->getConfig()->makeId( $title->getArticleID() ); |
121 | $docIds[] = $docId; |
122 | $likeDocs[] = [ '_id' => $docId ]; |
123 | } |
124 | |
125 | $moreLikeThisFields = $this->getConfig()->get( 'CirrusSearchMoreLikeThisFields' ); |
126 | sort( $moreLikeThisFields ); |
127 | $query = new MoreLikeThis(); |
128 | $query->setParams( $this->getConfig()->get( 'CirrusSearchMoreLikeThisConfig' ) ); |
129 | $query->setFields( $moreLikeThisFields ); |
130 | |
131 | /** @phan-suppress-next-line PhanTypeMismatchArgumentProbablyReal library is mis-annotated */ |
132 | $query->setLike( $likeDocs ); |
133 | |
134 | return $query; |
135 | } |
136 | |
137 | /** |
138 | * @return SearchConfig |
139 | */ |
140 | abstract public function getConfig(): SearchConfig; |
141 | } |