xref: /dokuwiki/inc/Utf8/Sort.php (revision f9aa34a37441e1fedd970151f80edc3bd4223e7d)
12d85e841SAndreas Gohr<?php
22d85e841SAndreas Gohr
32d85e841SAndreas Gohrnamespace dokuwiki\Utf8;
42d85e841SAndreas Gohr
52d85e841SAndreas Gohr/**
62d85e841SAndreas Gohr * DokuWiki sort functions
72d85e841SAndreas Gohr *
82d85e841SAndreas Gohr * When "intl" extension is available, all sorts are done using a collator.
92d85e841SAndreas Gohr * Otherwise, primitive PHP functions are called.
102d85e841SAndreas Gohr *
112d85e841SAndreas Gohr * The collator is created using the locale given in $conf['lang'].
122d85e841SAndreas Gohr * It always uses case insensitive "natural" ordering in its collation.
132d85e841SAndreas Gohr * The fallback solution uses the primitive PHP functions that return almost the same results
142d85e841SAndreas Gohr * when the input is text with only [A-Za-z0-9] characters.
152d85e841SAndreas Gohr *
162d85e841SAndreas Gohr * @license    GPL 2 (http://www.gnu.org/licenses/gpl.html)
172d85e841SAndreas Gohr * @author     Moisés Braga Ribeiro <moisesbr@gmail.com>
182d85e841SAndreas Gohr */
192d85e841SAndreas Gohrclass Sort
202d85e841SAndreas Gohr{
212d85e841SAndreas Gohr    /** @var \Collator[] language specific collators, usually only one */
22*f9aa34a3SAndreas Gohr    protected static $collator = [];
23*f9aa34a3SAndreas Gohr
24*f9aa34a3SAndreas Gohr    /** @var bool should the intl extension be used if available? For testing only */
25*f9aa34a3SAndreas Gohr    protected static $useIntl = true;
262d85e841SAndreas Gohr
272d85e841SAndreas Gohr    /**
282d85e841SAndreas Gohr     * Initialization of a collator using $conf['lang'] as the locale.
292d85e841SAndreas Gohr     * The initialization is done only once, except when $reload is set to true.
302d85e841SAndreas Gohr     * The collation takes "natural ordering" into account, that is, "page 2" is before "page 10".
312d85e841SAndreas Gohr     *
322d85e841SAndreas Gohr     * @param bool $reload Usually false; true forces collator re-creation
332d85e841SAndreas Gohr     * @return \Collator Returns a configured collator or null if the collator cannot be created.
342d85e841SAndreas Gohr     *
352d85e841SAndreas Gohr     * @author Moisés Braga Ribeiro <moisesbr@gmail.com>
362d85e841SAndreas Gohr     */
372d85e841SAndreas Gohr    protected static function getCollator($reload = false)
382d85e841SAndreas Gohr    {
392d85e841SAndreas Gohr        global $conf;
402d85e841SAndreas Gohr        $lc = $conf['lang'];
412d85e841SAndreas Gohr
422d85e841SAndreas Gohr        // check if intl extension is available
43*f9aa34a3SAndreas Gohr        if (!self::$useIntl || !class_exists('\Collator')) {
442d85e841SAndreas Gohr            return null;
452d85e841SAndreas Gohr        }
462d85e841SAndreas Gohr
472d85e841SAndreas Gohr        // load collator if not available yet
482d85e841SAndreas Gohr        if ($reload || !isset(self::$collator[$lc])) {
492d85e841SAndreas Gohr            $collator = \Collator::create($lc);
502d85e841SAndreas Gohr            if (!isset($collator)) return null;
512d85e841SAndreas Gohr            $collator->setAttribute(\Collator::NUMERIC_COLLATION, \Collator::ON);
522d85e841SAndreas Gohr            self::$collator[$lc] = $collator;
532d85e841SAndreas Gohr        }
542d85e841SAndreas Gohr
552d85e841SAndreas Gohr        return self::$collator[$lc];
562d85e841SAndreas Gohr    }
572d85e841SAndreas Gohr
582d85e841SAndreas Gohr    /**
59*f9aa34a3SAndreas Gohr     * Enable or disable the use of the intl extension collator
60*f9aa34a3SAndreas Gohr     *
61*f9aa34a3SAndreas Gohr     * This is mostly used for testing and should not be used in normal code
62*f9aa34a3SAndreas Gohr     *
63*f9aa34a3SAndreas Gohr     * @param bool $use
64*f9aa34a3SAndreas Gohr     */
65*f9aa34a3SAndreas Gohr    public static function useIntl($use = true)
66*f9aa34a3SAndreas Gohr    {
67*f9aa34a3SAndreas Gohr        self::$useIntl = $use;
68*f9aa34a3SAndreas Gohr        self::$collator = [];
69*f9aa34a3SAndreas Gohr    }
70*f9aa34a3SAndreas Gohr
71*f9aa34a3SAndreas Gohr    /**
722d85e841SAndreas Gohr     * Drop-in replacement for strcmp(), strcasecmp(), strnatcmp() and strnatcasecmp().
732d85e841SAndreas Gohr     * It uses a collator-based comparison, or strnatcasecmp() as a fallback.
742d85e841SAndreas Gohr     *
752d85e841SAndreas Gohr     * @param string $str1 The first string.
762d85e841SAndreas Gohr     * @param string $str2 The second string.
772d85e841SAndreas Gohr     * @return int Returns < 0 if $str1 is less than $str2; > 0 if $str1 is greater than $str2, and 0 if they are equal.
782d85e841SAndreas Gohr     *
792d85e841SAndreas Gohr     * @author Moisés Braga Ribeiro <moisesbr@gmail.com>
802d85e841SAndreas Gohr     */
812d85e841SAndreas Gohr    public static function strcmp($str1, $str2)
822d85e841SAndreas Gohr    {
832d85e841SAndreas Gohr        $collator = self::getCollator();
842d85e841SAndreas Gohr        if (isset($collator)) {
852d85e841SAndreas Gohr            return $collator->compare($str1, $str2);
862d85e841SAndreas Gohr        } else {
872d85e841SAndreas Gohr            return strnatcasecmp($str1, $str2);
882d85e841SAndreas Gohr        }
892d85e841SAndreas Gohr    }
902d85e841SAndreas Gohr
912d85e841SAndreas Gohr    /**
922d85e841SAndreas Gohr     * Drop-in replacement for sort().
932d85e841SAndreas Gohr     * It uses a collator-based sort, or sort() with flags SORT_NATURAL and SORT_FLAG_CASE as a fallback.
942d85e841SAndreas Gohr     *
952d85e841SAndreas Gohr     * @param array $array The input array.
962d85e841SAndreas Gohr     * @return bool Returns true on success or false on failure.
972d85e841SAndreas Gohr     *
982d85e841SAndreas Gohr     * @author Moisés Braga Ribeiro <moisesbr@gmail.com>
992d85e841SAndreas Gohr     */
1002d85e841SAndreas Gohr    public static function sort(&$array)
1012d85e841SAndreas Gohr    {
1022d85e841SAndreas Gohr        $collator = self::getCollator();
1032d85e841SAndreas Gohr        if (isset($collator)) {
1042d85e841SAndreas Gohr            return $collator->sort($array);
1052d85e841SAndreas Gohr        } else {
1062d85e841SAndreas Gohr            return sort($array, SORT_NATURAL | SORT_FLAG_CASE);
1072d85e841SAndreas Gohr        }
1082d85e841SAndreas Gohr    }
1092d85e841SAndreas Gohr
1102d85e841SAndreas Gohr    /**
1112d85e841SAndreas Gohr     * Drop-in replacement for ksort().
1122d85e841SAndreas Gohr     * It uses a collator-based sort, or ksort() with flags SORT_NATURAL and SORT_FLAG_CASE as a fallback.
1132d85e841SAndreas Gohr     *
1142d85e841SAndreas Gohr     * @param array $array The input array.
1152d85e841SAndreas Gohr     * @return bool Returns true on success or false on failure.
1162d85e841SAndreas Gohr     *
1172d85e841SAndreas Gohr     * @author Moisés Braga Ribeiro <moisesbr@gmail.com>
1182d85e841SAndreas Gohr     */
1192d85e841SAndreas Gohr    public static function ksort(&$array)
1202d85e841SAndreas Gohr    {
1212d85e841SAndreas Gohr        $collator = self::getCollator();
1222d85e841SAndreas Gohr        if (isset($collator)) {
1232d85e841SAndreas Gohr            return uksort($array, array($collator, 'compare'));
1242d85e841SAndreas Gohr        } else {
1252d85e841SAndreas Gohr            return ksort($array, SORT_NATURAL | SORT_FLAG_CASE);
1262d85e841SAndreas Gohr        }
1272d85e841SAndreas Gohr    }
1282d85e841SAndreas Gohr
1292d85e841SAndreas Gohr    /**
1302d85e841SAndreas Gohr     * Drop-in replacement for asort(), natsort() and natcasesort().
1312d85e841SAndreas Gohr     * It uses a collator-based sort, or asort() with flags SORT_NATURAL and SORT_FLAG_CASE as a fallback.
1322d85e841SAndreas Gohr     *
1332d85e841SAndreas Gohr     * @param array $array The input array.
1342d85e841SAndreas Gohr     * @return bool Returns true on success or false on failure.
1352d85e841SAndreas Gohr     *
1362d85e841SAndreas Gohr     * @author Moisés Braga Ribeiro <moisesbr@gmail.com>
1372d85e841SAndreas Gohr     */
1382d85e841SAndreas Gohr    public static function asort(&$array)
1392d85e841SAndreas Gohr    {
1402d85e841SAndreas Gohr        $collator = self::getCollator();
1412d85e841SAndreas Gohr        if (isset($collator)) {
1422d85e841SAndreas Gohr            return $collator->asort($array);
1432d85e841SAndreas Gohr        } else {
1442d85e841SAndreas Gohr            return asort($array, SORT_NATURAL | SORT_FLAG_CASE);
1452d85e841SAndreas Gohr        }
1462d85e841SAndreas Gohr    }
1472d85e841SAndreas Gohr
1482d85e841SAndreas Gohr    /**
1492d85e841SAndreas Gohr     * Drop-in replacement for asort(), natsort() and natcasesort() when the parameter is an array of filenames.
1502d85e841SAndreas Gohr     * Filenames may not be equal to page names, depending on the setting in $conf['fnencode'],
1512d85e841SAndreas Gohr     * so the correct behavior is to sort page names and reflect this sorting in the filename array.
1522d85e841SAndreas Gohr     *
1532d85e841SAndreas Gohr     * @param array $array The input array.
1542d85e841SAndreas Gohr     * @return bool Returns true on success or false on failure.
1552d85e841SAndreas Gohr     *
1562d85e841SAndreas Gohr     * @author Moisés Braga Ribeiro <moisesbr@gmail.com>
1572d85e841SAndreas Gohr     */
1582d85e841SAndreas Gohr    public static function asortFN(&$array)
1592d85e841SAndreas Gohr    {
1602d85e841SAndreas Gohr        $collator = self::getCollator();
1612d85e841SAndreas Gohr        return uasort($array, function ($fn1, $fn2) use ($collator) {
1622d85e841SAndreas Gohr            if (isset($collator)) {
1632d85e841SAndreas Gohr                return $collator->compare(utf8_decodeFN($fn1), utf8_decodeFN($fn2));
1642d85e841SAndreas Gohr            } else {
1652d85e841SAndreas Gohr                return strnatcasecmp(utf8_decodeFN($fn1), utf8_decodeFN($fn2));
1662d85e841SAndreas Gohr            }
1672d85e841SAndreas Gohr        });
1682d85e841SAndreas Gohr    }
1692d85e841SAndreas Gohr
1702d85e841SAndreas Gohr}
171