Code Coverage |
||||||||||
Lines |
Functions and Methods |
Classes and Traits |
||||||||
| Total | |
0.00% |
0 / 47 |
|
0.00% |
0 / 5 |
CRAP | |
0.00% |
0 / 1 |
| SearchExactMatchRescorer | |
0.00% |
0 / 46 |
|
0.00% |
0 / 5 |
380 | |
0.00% |
0 / 1 |
| rescore | |
0.00% |
0 / 28 |
|
0.00% |
0 / 1 |
110 | |||
| getReplacedRedirect | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
| redirectTargetsToRedirect | |
0.00% |
0 / 10 |
|
0.00% |
0 / 1 |
30 | |||
| pullFront | |
0.00% |
0 / 3 |
|
0.00% |
0 / 1 |
2 | |||
| getRedirectTarget | |
0.00% |
0 / 4 |
|
0.00% |
0 / 1 |
6 | |||
| 1 | <?php |
| 2 | /** |
| 3 | * Rescores results from a prefix search/opensearch to make sure the |
| 4 | * exact match is the first result. |
| 5 | * |
| 6 | * @license GPL-2.0-or-later |
| 7 | * @file |
| 8 | */ |
| 9 | |
| 10 | namespace MediaWiki\Search; |
| 11 | |
| 12 | use MediaWiki\MediaWikiServices; |
| 13 | use MediaWiki\Page\PageIdentity; |
| 14 | use MediaWiki\Title\Title; |
| 15 | |
| 16 | /** |
| 17 | * An utility class to rescore search results by looking for an exact match |
| 18 | * in the db and add the page found to the first position. |
| 19 | * |
| 20 | * NOTE: extracted from TitlePrefixSearch |
| 21 | * @ingroup Search |
| 22 | */ |
| 23 | class SearchExactMatchRescorer { |
| 24 | /** |
| 25 | * @var ?string set when a redirect returned from the engine is replaced by the exact match |
| 26 | */ |
| 27 | private ?string $replacedRedirect; |
| 28 | |
| 29 | /** |
| 30 | * Default search backend does proper prefix searching, but custom backends |
| 31 | * may sort based on other algorithms that may cause the exact title match |
| 32 | * to not be in the results or be lower down the list. |
| 33 | * @param string $search the query |
| 34 | * @param int[] $namespaces |
| 35 | * @param string[] $srchres results |
| 36 | * @param int $limit the max number of results to return |
| 37 | * @return string[] munged results |
| 38 | */ |
| 39 | public function rescore( $search, $namespaces, $srchres, $limit ) { |
| 40 | $this->replacedRedirect = null; |
| 41 | // Pick namespace (based on PrefixSearch::defaultSearchBackend) |
| 42 | $ns = in_array( NS_MAIN, $namespaces ) ? NS_MAIN : reset( $namespaces ); |
| 43 | $t = Title::newFromText( $search, $ns ); |
| 44 | if ( !$t || !$t->exists() ) { |
| 45 | // No exact match so just return the search results |
| 46 | return $srchres; |
| 47 | } |
| 48 | $string = $t->getPrefixedText(); |
| 49 | $key = array_search( $string, $srchres ); |
| 50 | if ( $key !== false ) { |
| 51 | // Exact match was in the results so just move it to the front |
| 52 | return $this->pullFront( $key, $srchres ); |
| 53 | } |
| 54 | // Exact match not in the search results so check for some redirect handling cases |
| 55 | if ( $t->isRedirect() ) { |
| 56 | $target = $this->getRedirectTarget( $t ); |
| 57 | $key = array_search( $target, $srchres ); |
| 58 | if ( $key !== false ) { |
| 59 | // Exact match is a redirect to one of the returned matches so pull the |
| 60 | // returned match to the front. This might look odd but the alternative |
| 61 | // is to put the redirect in front and drop the match. The name of the |
| 62 | // found match is often more descriptive/better formed than the name of |
| 63 | // the redirect AND by definition they share a prefix. Hopefully this |
| 64 | // choice is less confusing and more helpful. But it might not be. But |
| 65 | // it is the choice we're going with for now. |
| 66 | return $this->pullFront( $key, $srchres ); |
| 67 | } |
| 68 | $redirectTargetsToRedirect = $this->redirectTargetsToRedirect( $srchres ); |
| 69 | if ( isset( $redirectTargetsToRedirect[$target] ) ) { |
| 70 | // The exact match and something in the results list are both redirects |
| 71 | // to the same thing! In this case we prefer the match the user typed. |
| 72 | $this->replacedRedirect = array_splice( $srchres, $redirectTargetsToRedirect[$target], 1 )[0]; |
| 73 | array_unshift( $srchres, $string ); |
| 74 | return $srchres; |
| 75 | } |
| 76 | } else { |
| 77 | $redirectTargetsToRedirect = $this->redirectTargetsToRedirect( $srchres ); |
| 78 | if ( isset( $redirectTargetsToRedirect[$string] ) ) { |
| 79 | // The exact match is the target of a redirect already in the results list so remove |
| 80 | // the redirect from the results list and push the exact match to the front |
| 81 | array_splice( $srchres, $redirectTargetsToRedirect[$string], 1 ); |
| 82 | array_unshift( $srchres, $string ); |
| 83 | return $srchres; |
| 84 | } |
| 85 | } |
| 86 | |
| 87 | // Exact match is totally unique from the other results so just add it to the front |
| 88 | array_unshift( $srchres, $string ); |
| 89 | // And roll one off the end if the results are too long |
| 90 | if ( count( $srchres ) > $limit ) { |
| 91 | array_pop( $srchres ); |
| 92 | } |
| 93 | return $srchres; |
| 94 | } |
| 95 | |
| 96 | /** |
| 97 | * Redirect initially returned by the search engine that got replaced by a better match: |
| 98 | * - exact match to a redirect to the same page |
| 99 | * - exact match to the target page |
| 100 | * @return string|null the replaced redirect or null if nothing was replaced |
| 101 | */ |
| 102 | public function getReplacedRedirect(): ?string { |
| 103 | return $this->replacedRedirect; |
| 104 | } |
| 105 | |
| 106 | /** |
| 107 | * @param string[] $titles |
| 108 | * @return array redirect target prefixedText to index of title in titles |
| 109 | * that is a redirect to it. |
| 110 | */ |
| 111 | private function redirectTargetsToRedirect( array $titles ) { |
| 112 | $result = []; |
| 113 | foreach ( $titles as $key => $titleText ) { |
| 114 | $title = Title::newFromText( $titleText ); |
| 115 | if ( !$title || !$title->isRedirect() ) { |
| 116 | continue; |
| 117 | } |
| 118 | $target = $this->getRedirectTarget( $title ); |
| 119 | if ( !$target ) { |
| 120 | continue; |
| 121 | } |
| 122 | $result[$target] = $key; |
| 123 | } |
| 124 | return $result; |
| 125 | } |
| 126 | |
| 127 | /** |
| 128 | * Returns an array where the element of $array at index $key becomes |
| 129 | * the first element. |
| 130 | * @param int $key key to pull to the front |
| 131 | * @param array $array |
| 132 | * @return array $array with the item at $key pulled to the front |
| 133 | */ |
| 134 | private function pullFront( $key, array $array ) { |
| 135 | $cut = array_splice( $array, $key, 1 ); |
| 136 | array_unshift( $array, $cut[0] ); |
| 137 | return $array; |
| 138 | } |
| 139 | |
| 140 | /** |
| 141 | * Get a redirect's destination from a title |
| 142 | * @param PageIdentity $page A page to redirect. It may not redirect or even exist |
| 143 | * @return null|string If title exists and redirects, get the destination's prefixed name |
| 144 | */ |
| 145 | private function getRedirectTarget( PageIdentity $page ) { |
| 146 | $redirectStore = MediaWikiServices::getInstance()->getRedirectStore(); |
| 147 | $redir = $redirectStore->getRedirectTarget( $page ); |
| 148 | |
| 149 | // Needed to get the text needed for display. |
| 150 | $redir = Title::castFromLinkTarget( $redir ); |
| 151 | return $redir ? $redir->getPrefixedText() : null; |
| 152 | } |
| 153 | } |
| 154 | |
| 155 | /** @deprecated class alias since 1.46 */ |
| 156 | class_alias( SearchExactMatchRescorer::class, 'SearchExactMatchRescorer' ); |