Code Coverage
 
Lines
Functions and Methods
Classes and Traits
Total
95.45% covered (success)
95.45%
42 / 44
71.43% covered (warning)
71.43%
5 / 7
 CRAP
0.00% covered (danger)
0.00%
0 / 1
RemoteIcuCollation
95.45% covered (success)
95.45%
42 / 44
71.43% covered (warning)
71.43%
5 / 7
 13
0.00% covered (danger)
0.00%
0 / 1
 __construct
100.00% covered (success)
100.00%
3 / 3
100.00% covered (success)
100.00%
1 / 1
 1
 getSortKey
100.00% covered (success)
100.00%
1 / 1
100.00% covered (success)
100.00%
1 / 1
 1
 encode
100.00% covered (success)
100.00%
4 / 4
100.00% covered (success)
100.00%
1 / 1
 2
 decode
100.00% covered (success)
100.00%
8 / 8
100.00% covered (success)
100.00%
1 / 1
 2
 getSortKeys
100.00% covered (success)
100.00%
18 / 18
100.00% covered (success)
100.00%
1 / 1
 2
 getFirstLetter
0.00% covered (danger)
0.00%
0 / 1
0.00% covered (danger)
0.00%
0 / 1
 2
 doGetSortKeys
88.89% covered (warning)
88.89%
8 / 9
0.00% covered (danger)
0.00%
0 / 1
 4.02
1<?php
2
3use MediaWiki\Shell\ShellboxClientFactory;
4use Shellbox\RPC\RpcClient;
5
6/**
7 * An ICU collation that uses a remote server to compute sort keys. This can be
8 * used in conjunction with $wgTempCategoryCollations to migrate to a different
9 * version of ICU.
10 */
11class RemoteIcuCollation extends Collation {
12    private RpcClient $rpcClient;
13    private string $locale;
14
15    public function __construct( ShellboxClientFactory $shellboxClientFactory, string $locale ) {
16        $this->rpcClient = $shellboxClientFactory->getRpcClient(
17            [ 'service' => 'icu-collation' ] );
18        $this->locale = $locale;
19    }
20
21    public function getSortKey( $string ) {
22        return $this->getSortKeys( [ $string ] )[0];
23    }
24
25    /**
26     * Encode an array of binary strings as a string
27     *
28     * @param string[] $strings
29     * @return string
30     */
31    private static function encode( $strings ) {
32        $ret = '';
33        foreach ( $strings as $s ) {
34            $ret .= sprintf( "%08x", strlen( $s ) ) . $s;
35        }
36        return $ret;
37    }
38
39    /**
40     * Decode the value returned by encode()
41     *
42     * @param string $blob
43     * @return string[]
44     */
45    private static function decode( $blob ) {
46        $p = 0;
47        $ret = [];
48        while ( $p < strlen( $blob ) ) {
49            $len = intval( substr( $blob, $p, 8 ), 16 );
50            $p += 8;
51            $ret[] = substr( $blob, $p, $len );
52            $p += $len;
53        }
54        return $ret;
55    }
56
57    public function getSortKeys( $strings ) {
58        if ( !count( $strings ) ) {
59            return [];
60        }
61        $blob = $this->rpcClient->call(
62            'icu-collation',
63            self::class . '::' . 'doGetSortKeys',
64            [
65                $this->locale,
66                self::encode( array_values( $strings ) )
67            ],
68            [
69                'classes' => [ parent::class, self::class ],
70                'binary' => true
71            ]
72        );
73        return array_combine(
74            array_keys( $strings ),
75            self::decode( $blob )
76        );
77    }
78
79    public function getFirstLetter( $string ) {
80        // @phan-suppress-previous-line PhanPluginNeverReturnMethod
81        throw new RuntimeException( __METHOD__ . ': not implemented' );
82    }
83
84    /**
85     * The remote entry point. Get sort keys for an encoded list of inputs.
86     *
87     * @param string $locale The ICU locale
88     * @param string $blob The input array encoded with encode()
89     * @return string The encoded result
90     */
91    public static function doGetSortKeys( $locale, $blob ) {
92        $mainCollator = Collator::create( $locale );
93        if ( !$mainCollator ) {
94            throw new RuntimeException( "Invalid ICU locale specified for collation: $locale" );
95        }
96
97        // If the special suffix for numeric collation is present, turn on numeric collation.
98        if ( str_ends_with( $locale, '-u-kn' ) ) {
99            $mainCollator->setAttribute( Collator::NUMERIC_COLLATION, Collator::ON );
100        }
101        $ret = [];
102        foreach ( self::decode( $blob ) as $string ) {
103            $ret[] = $mainCollator->getSortKey( $string );
104        }
105        return self::encode( $ret );
106    }
107}