Code Coverage
 
Lines
Functions and Methods
Classes and Traits
Total
0.00% covered (danger)
0.00%
0 / 42
0.00% covered (danger)
0.00%
0 / 4
 CRAP
0.00% covered (danger)
0.00%
0 / 1
NearMatchPicker
0.00% covered (danger)
0.00%
0 / 42
0.00% covered (danger)
0.00%
0 / 4
 380
0.00% covered (danger)
0.00%
0 / 1
 __construct
0.00% covered (danger)
0.00%
0 / 3
0.00% covered (danger)
0.00%
0 / 1
 2
 pickBest
0.00% covered (danger)
0.00%
0 / 29
0.00% covered (danger)
0.00%
0 / 1
 132
 checkAllMatches
0.00% covered (danger)
0.00%
0 / 8
0.00% covered (danger)
0.00%
0 / 1
 42
 checkOneMatch
0.00% covered (danger)
0.00%
0 / 2
0.00% covered (danger)
0.00%
0 / 1
 2
1<?php
2
3namespace CirrusSearch;
4
5use Language;
6use MediaWiki\Logger\LoggerFactory;
7use Title;
8
9/**
10 * Picks the best "near match" title.
11 *
12 * This program is free software; you can redistribute it and/or modify
13 * it under the terms of the GNU General Public License as published by
14 * the Free Software Foundation; either version 2 of the License, or
15 * (at your option) any later version.
16 *
17 * This program is distributed in the hope that it will be useful,
18 * but WITHOUT ANY WARRANTY; without even the implied warranty of
19 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
20 * GNU General Public License for more details.
21 *
22 * You should have received a copy of the GNU General Public License along
23 * with this program; if not, write to the Free Software Foundation, Inc.,
24 * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
25 * http://www.gnu.org/copyleft/gpl.html
26 */
27class NearMatchPicker {
28    /**
29     * @var Language language to use during normalization process
30     */
31    private $language;
32    /**
33     * @var string the search term
34     */
35    private $term;
36    /**
37     * @var array[] Potential near matches
38     */
39    private $titles;
40
41    /**
42     * @param Language $language to use during normalization process
43     * @param string $term the search term
44     * @param array[] $titles Array of arrays, each with optional keys:
45     *   titleMatch => a title if the title matched
46     *   redirectMatches => an array of redirect matches, one per matched redirect
47     */
48    public function __construct( $language, $term, $titles ) {
49        $this->language = $language;
50        $this->term = $term;
51        $this->titles = $titles;
52    }
53
54    /**
55     * Pick the best near match if possible.
56     *
57     * @return Title|null title if there is a near match and null otherwise
58     */
59    public function pickBest() {
60        if ( !$this->titles ) {
61            return null;
62        }
63        if ( !$this->term ) {
64            return null;
65        }
66        if ( count( $this->titles ) === 1 ) {
67            if ( isset( $this->titles[ 0 ][ 'titleMatch' ] ) ) {
68                return $this->titles[ 0 ][ 'titleMatch' ];
69            }
70            if ( isset( $this->titles[ 0 ][ 'redirectMatches' ][ 0 ] ) ) {
71                return $this->titles[ 0 ][ 'redirectMatches' ][ 0 ];
72            }
73            LoggerFactory::getInstance( 'CirrusSearch' )->info(
74                'NearMatchPicker built with busted matches.  Assuming no near match' );
75            return null;
76        }
77
78        $transformers = [
79            static function ( $term ) {
80                return $term;
81            },
82            [ $this->language, 'lc' ],
83            [ $this->language, 'ucwords' ],
84        ];
85
86        foreach ( $transformers as $transformer ) {
87            $transformedTerm = call_user_func( $transformer, $this->term );
88            $found = null;
89            foreach ( $this->titles as $title ) {
90                $match = $this->checkAllMatches( $transformer, $transformedTerm, $title );
91                if ( $match ) {
92                    // @phan-suppress-next-line PhanSuspiciousValueComparisonInLoop
93                    if ( $found === null ) {
94                        $found = $match;
95                    } else {
96                        // Found more than one result so we try another transformer
97                        $found = null;
98                        break;
99                    }
100                }
101
102            }
103            if ( $found ) {
104                return $found;
105            }
106        }
107
108        // Didn't find anything
109        return null;
110    }
111
112    /**
113     * Check a single title's worth of matches.  The big thing here is that titles cannot compete with themselves.
114     * @param callable $transformer
115     * @param string $transformedTerm
116     * @param array $allMatchedTitles
117     * @return null|Title null if no title matches and the actual title (either of the page or of a redirect to the
118     *       page) if one did match
119     */
120    private function checkAllMatches( $transformer, $transformedTerm, $allMatchedTitles ) {
121        if ( isset( $allMatchedTitles[ 'titleMatch' ] ) &&
122                $this->checkOneMatch( $transformer, $transformedTerm, $allMatchedTitles[ 'titleMatch' ] ) ) {
123            return $allMatchedTitles[ 'titleMatch' ];
124        }
125        if ( isset( $allMatchedTitles[ 'redirectMatches' ] ) ) {
126            foreach ( $allMatchedTitles[ 'redirectMatches' ] as $redirectMatch ) {
127                if ( $this->checkOneMatch( $transformer, $transformedTerm, $redirectMatch ) ) {
128                    return $redirectMatch;
129                }
130            }
131        }
132        return null;
133    }
134
135    /**
136     * @param callable $transformer
137     * @param string $transformedTerm
138     * @param Title $matchedTitle
139     * @return bool
140     */
141    private function checkOneMatch( $transformer, $transformedTerm, $matchedTitle ) {
142        $transformedTitle = call_user_func( $transformer, $matchedTitle->getText() );
143        return $transformedTerm === $transformedTitle;
144    }
145}