<?php // vi: set fenc=utf-8 ts=4 sw=4 et:
/*
 * Copyright (C) 2013 Nicolas Grekas - p@tchwork.com
 *
 * This library is free software; you can redistribute it and/or modify it
 * under the terms of the (at your option):
 * Apache License v2.0 (http://apache.org/licenses/LICENSE-2.0.txt), or
 * GNU General Public License v2.0 (http://gnu.org/licenses/gpl-2.0.txt).
 */

namespace Patchwork;

use Normalizer as n;

/**
 * UTF-8 Grapheme Cluster aware string manipulations implementing the quasi complete
 * set of native PHP string functions that need UTF-8 awareness and more.
 * Missing are printf-family functions.
 */
class Utf8
{
    protected static

    $commonCaseFold = array(
        array('µ','ſ',"\xCD\x85",'ς',"\xCF\x90","\xCF\x91","\xCF\x95","\xCF\x96","\xCF\xB0","\xCF\xB1","\xCF\xB5","\xE1\xBA\x9B","\xE1\xBE\xBE"),
        array('μ','s','ι',       'σ','β',       'θ',       'φ',       'π',       'κ',       'ρ',       'ε',       "\xE1\xB9\xA1",'ι'           )
    ),
    $cp1252 = array('€','‚','ƒ','„','…','†','‡','ˆ','‰','Š','‹','Œ','Ž','‘','’','“','”','•','–','—','˜','™','š','›','œ','ž','Ÿ'),
    $utf8   = array('€','‚','ƒ','„','…','†','‡','ˆ','‰','Š','‹','Œ','Ž','‘','’','“','”','•','–','—','˜','™','š','›','œ','ž','Ÿ');


    static function isUtf8($s)
    {
        return (bool) preg_match('//u', $s); // Since PHP 5.2.5, this also excludes invalid five and six bytes sequences
    }

    // Generic UTF-8 to ASCII transliteration

    static function toAscii($s, $subst_chr = '?')
    {
        if (preg_match("/[\x80-\xFF]/", $s))
        {
            static $translitExtra = array();
            $translitExtra or $translitExtra = static::getData('translit_extra');

            $s = n::normalize($s, n::NFKC);

/**/        $glibc = 'glibc' === ICONV_IMPL;

            preg_match_all('/./u', $s, $s);

            foreach ($s[0] as &$c)
            {
                if (! isset($c[1])) continue;

/**/            if ($glibc)
/**/            {
                    $t = iconv('UTF-8', 'ASCII//TRANSLIT', $c);
/**/            }
/**/            else
/**/            {
                    $t = iconv('UTF-8', 'ASCII//IGNORE//TRANSLIT', $c);

                    if (! isset($t[0])) $t = '?';
                    else if (isset($t[1])) $t = ltrim($t, '\'`"^~');
/**/            }

                if ('?' === $t)
                {
                    if (isset($translitExtra[$c]))
                    {
                        $t = $translitExtra[$c];
                    }
                    else
                    {
                        $t = n::normalize($c, n::NFD);

                        if ($t[0] < "\x80") $t = $t[0];
                        else $t = $subst_chr;
                    }
                }

                $c = $t;
            }

            $s = implode('', $s[0]);
        }

        return $s;
    }

    static function filter($var, $normalization_form = 4 /* n::NFC */, $leading_combining = '◌')
    {
        switch (gettype($var))
        {
        case 'array':
            foreach ($var as $k => $v) $var[$k] = static::filter($v, $normalization_form, $leading_combining);
            break;

        case 'object':
            foreach ($var as $k => $v) $var->$k = static::filter($v, $normalization_form, $leading_combining);
            break;

        case 'string':
            if (false !== strpos($var, "\r"))
            {
                // Workaround https://bugs.php.net/65732
                $var = str_replace("\r\n", "\n", $var);
                $var = strtr($var, "\r", "\n");
            }

            if (preg_match('/[\x80-\xFF]/', $var))
            {
                if (n::isNormalized($var, $normalization_form)) $n = '-';
                else
                {
                    $n = n::normalize($var, $normalization_form);
                    if (isset($n[0])) $var = $n;
                    else $var = static::utf8_encode($var);
                }

                if ($var[0] >= "\x80" && isset($n[0], $leading_combining[0]) && preg_match('/^\p{Mn}/u', $var))
                {
                    // Prevent leading combining chars
                    // for NFC-safe concatenations.
                    $var = $leading_combining . $var;
                }
            }
            break;
        }

        return $var;
    }

    // Unicode transformation for caseless matching
    // see http://unicode.org/reports/tr21/tr21-5.html

    static function strtocasefold($s, $full = true)
    {
        $s = str_replace(self::$commonCaseFold[0], self::$commonCaseFold[1], $s);

        if ($full)
        {
            static $fullCaseFold = false;
            $fullCaseFold || $fullCaseFold = static::getData('caseFolding_full');

            $s = str_replace($fullCaseFold[0], $fullCaseFold[1], $s);
        }

        return static::strtolower($s);
    }

    // Generic case sensitive collation support for self::strnatcmp()

    static function strtonatfold($s)
    {
        $s = n::normalize($s, n::NFD);
        return preg_replace('/\p{Mn}+/u', '', $s);
    }

    // PHP string functions that need UTF-8 awareness

    static function filter_input($type, $var, $filter = FILTER_DEFAULT, $option = null)
    {
        if (4 > func_num_args()) $var = filter_input($type, $var, $filter);
        else $var = filter_input($type, $var, $filter, $option);

        return static::filter($var);
    }

    static function filter_input_array($type, $def = null, $add_empty = true)
    {
        if (2 > func_num_args()) $a = filter_input_array($type);
        else $a = filter_input_array($type, $def, $add_empty);

        return static::filter($a);
    }

    static function json_decode($json, $assoc = false, $depth = 512, $options = 0)
    {
/**/    if (PHP_VERSION_ID < 50400)
/**/    {
            $json = json_decode($json, $assoc, $depth);
/**/    }
/**/    else
/**/    {
            $json = json_decode($json, $assoc, $depth, $options);
/**/    }

        return static::filter($json);
    }

    static function substr($s, $start, $len = 2147483647)
    {
/**/    static $bug62759;

/**/    isset($bug62759) or $bug62759 = extension_loaded('intl') && 'à' === grapheme_substr('éà', 1, -2);

/**/    if ($bug62759)
/**/    {
            return PHP\Shim\Intl::grapheme_substr_workaround62759($s, $start, $len);
/**/    }
/**/    else
/**/    {
            return grapheme_substr($s, $start, $len);
/**/    }
    }

    static function strlen($s) {return grapheme_strlen($s);}
    static function strpos ($s, $needle, $offset = 0) {return grapheme_strpos ($s, $needle, $offset);}
    static function strrpos($s, $needle, $offset = 0) {return grapheme_strrpos($s, $needle, $offset);}

    static function stripos($s, $needle, $offset = 0)
    {
/**/    if (50418 > PHP_VERSION_ID || 50500 == PHP_VERSION_ID)
/**/    {
            // Don't use grapheme_stripos because of https://bugs.php.net/61860
            if (! preg_match('//u', $s .= '')) return false;
            if ($offset < 0) $offset = 0;
            if (! $needle = mb_stripos($s, $needle .= '', $offset, 'UTF-8')) return $needle;
            return grapheme_strlen(iconv_substr($s, 0, $needle, 'UTF-8'));
/**/    }
/**/    else
/**/    {
            return grapheme_stripos($s, $needle, $offset);
/**/    }
    }

    static function strripos($s, $needle, $offset = 0)
    {
/**/    if (50418 > PHP_VERSION_ID || 50500 == PHP_VERSION_ID)
/**/    {
            // Don't use grapheme_strripos because of https://bugs.php.net/61860
            if (! preg_match('//u', $s .= '')) return false;
            if ($offset < 0) $offset = 0;
            if (! $needle = mb_strripos($s, $needle .= '', $offset, 'UTF-8')) return $needle;
            return grapheme_strlen(iconv_substr($s, 0, $needle, 'UTF-8'));
/**/    }
/**/    else
/**/    {
            return grapheme_strripos($s, $needle, $offset);
/**/    }
    }

    static function stristr($s, $needle, $before_needle = false)
    {
        if ('' === $needle .= '') return false;
        return mb_stristr($s, $needle, $before_needle, 'UTF-8');
    }

    static function strstr  ($s, $needle, $before_needle = false) {return grapheme_strstr($s, $needle, $before_needle);}
    static function strrchr ($s, $needle, $before_needle = false) {return mb_strrchr ($s, $needle, $before_needle, 'UTF-8');}
    static function strrichr($s, $needle, $before_needle = false) {return mb_strrichr($s, $needle, $before_needle, 'UTF-8');}

    static function strtolower($s) {return mb_strtolower($s, 'UTF-8');}
    static function strtoupper($s) {return mb_strtoupper($s, 'UTF-8');}

    static function wordwrap($s, $width = 75, $break = "\n", $cut = false)
    {
        if (false === wordwrap('-', $width, $break, $cut)) return false;

        is_string($break) or $break = (string) $break;

        $w = '';
        $s = explode($break, $s);
        $iLen = count($s);
        $chars = array();

        if (1 === $iLen && '' === $s[0])
            return '';

        for ($i = 0; $i < $iLen; ++$i)
        {
            if ($i)
            {
                $chars[] = $break;
                $w .= '#';
            }

            $c = $s[$i];
            unset($s[$i]);

            foreach (self::str_split($c) as $c)
            {
                $chars[] = $c;
                $w .= ' ' === $c ? ' ' : '?';
            }
        }

        $s = '';
        $j = 0;
        $b = $i = -1;
        $w = wordwrap($w, $width, '#', $cut);

        while (false !== $b = strpos($w, '#', $b+1))
        {
            for (++$i; $i < $b; ++$i)
            {
                $s .= $chars[$j];
                unset($chars[$j++]);
            }

            if ($break === $chars[$j] || ' ' === $chars[$j]) unset($chars[$j++]);
            $s .= $break;
        }

        return $s . implode('', $chars);
    }

    static function chr($c)
    {
        if (0x80 > $c %= 0x200000) return chr($c);
        if (0x800 > $c) return chr(0xC0 | $c>>6) . chr(0x80 | $c & 0x3F);
        if (0x10000 > $c) return chr(0xE0 | $c>>12) . chr(0x80 | $c>>6 & 0x3F) . chr(0x80 | $c & 0x3F);
        return chr(0xF0 | $c>>18) . chr(0x80 | $c>>12 & 0x3F) . chr(0x80 | $c>>6 & 0x3F) . chr(0x80 | $c & 0x3F);
    }

    static function count_chars($s, $mode = 0)
    {
        if (1 != $mode) user_error(__METHOD__ . '(): the only allowed $mode is 1', E_USER_WARNING);
        $s = self::str_split($s);
        return array_count_values($s);
    }

    static function ltrim($s, $charlist = INF)
    {
        $charlist = INF === $charlist ? '\s' : self::rxClass($charlist);
        return preg_replace("/^{$charlist}+/u", '', $s);
    }

    static function ord($s)
    {
        $a = ($s = unpack('C*', substr($s, 0, 4))) ? $s[1] : 0;
        if (0xF0 <= $a) return (($a - 0xF0)<<18) + (($s[2] - 0x80)<<12) + (($s[3] - 0x80)<<6) + $s[4] - 0x80;
        if (0xE0 <= $a) return (($a - 0xE0)<<12) + (($s[2] - 0x80)<<6) + $s[3] - 0x80;
        if (0xC0 <= $a) return (($a - 0xC0)<<6) + $s[2] - 0x80;
        return $a;
    }

    static function rtrim($s, $charlist = INF)
    {
        $charlist = INF === $charlist ? '\s' : self::rxClass($charlist);
        return preg_replace("/{$charlist}+$/u", '', $s);
    }

    static function trim($s, $charlist = INF) {return self::rtrim(self::ltrim($s, $charlist), $charlist);}

    static function str_ireplace($search, $replace, $subject, &$count = null)
    {
        $search = (array) $search;

        foreach ($search as $i => $s)
        {
            if ('' === $s .= '') $s = '/^(?<=.)$/';
            else $s = '/' . preg_quote($s, '/') . '/ui';

            $search[$i] = $s;
        }

        $subject = preg_replace($search, $replace, $subject, -1, $replace);
        $count = $replace;

        return $subject;
    }

    static function str_pad($s, $len, $pad = ' ', $type = STR_PAD_RIGHT)
    {
        $slen = grapheme_strlen($s);
        if ($len <= $slen) return $s;

        $padlen = grapheme_strlen($pad);
        $freelen = $len - $slen;
        $len = $freelen % $padlen;

        if (STR_PAD_RIGHT == $type) return $s . str_repeat($pad, $freelen / $padlen) . ($len ? grapheme_substr($pad, 0, $len) : '');
        if (STR_PAD_LEFT  == $type) return      str_repeat($pad, $freelen / $padlen) . ($len ? grapheme_substr($pad, 0, $len) : '') . $s;
        if (STR_PAD_BOTH  == $type)
        {
            $freelen /= 2;

            $type = ceil($freelen);
            $len = $type % $padlen;
            $s .= str_repeat($pad, $type / $padlen) . ($len ? grapheme_substr($pad, 0, $len) : '');

            $type = floor($freelen);
            $len = $type % $padlen;
            return str_repeat($pad, $type / $padlen) . ($len ? grapheme_substr($pad, 0, $len) : '') . $s;
        }

        user_error(__METHOD__ . '(): Padding type has to be STR_PAD_LEFT, STR_PAD_RIGHT, or STR_PAD_BOTH', E_USER_WARNING);
    }

    static function str_shuffle($s)
    {
        $s = self::str_split($s);
        shuffle($s);
        return implode('', $s);
    }

    static function str_split($s, $len = 1)
    {
        if (1 > $len = (int) $len)
        {
            $len = func_get_arg(1);
            return str_split($s, $len);
        }

/**/    if (extension_loaded('intl'))
/**/    {
            $a = array();
            $p = 0;
            $l = strlen($s);

            while ($p < $l) $a[] = grapheme_extract($s, 1, GRAPHEME_EXTR_COUNT, $p, $p);
/**/    }
/**/    else
/**/    {
            preg_match_all('/' . GRAPHEME_CLUSTER_RX . '/u', $s, $a);
            $a = $a[0];
/**/    }

        if (1 == $len) return $a;

        $s = array();
        $p = -1;

        foreach ($a as $l => $a)
        {
            if ($l % $len) $s[$p] .= $a;
            else $s[++$p] = $a;
        }

        return $s;
    }

    static function str_word_count($s, $format = 0, $charlist = '')
    {
        $charlist = self::rxClass($charlist, '\pL');
        $s = preg_split("/({$charlist}+(?:[\p{Pd}’']{$charlist}+)*)/u", $s, -1, PREG_SPLIT_DELIM_CAPTURE);

        $charlist = array();
        $len = count($s);

        if (1 == $format) for ($i = 1; $i < $len; $i+=2) $charlist[] = $s[$i];
        else if (2 == $format)
        {
            $offset = grapheme_strlen($s[0]);
            for ($i = 1; $i < $len; $i+=2)
            {
                $charlist[$offset] = $s[$i];
                $offset += grapheme_strlen($s[$i]) + grapheme_strlen($s[$i+1]);
            }
        }
        else $charlist = ($len - 1) / 2;

        return $charlist;
    }

    static function strcmp       ($a, $b) {return $a . '' === $b . '' ? 0 : strcmp(n::normalize($a, n::NFD), n::normalize($b, n::NFD));}
    static function strnatcmp    ($a, $b) {return $a . '' === $b . '' ? 0 : strnatcmp(self::strtonatfold($a), self::strtonatfold($b));}
    static function strcasecmp   ($a, $b) {return self::strcmp   (static::strtocasefold($a), static::strtocasefold($b));}
    static function strnatcasecmp($a, $b) {return self::strnatcmp(static::strtocasefold($a), static::strtocasefold($b));}
    static function strncasecmp  ($a, $b, $len) {return self::strncmp(static::strtocasefold($a), static::strtocasefold($b), $len);}
    static function strncmp      ($a, $b, $len) {return self::strcmp(self::substr($a, 0, $len), self::substr($b, 0, $len));}

    static function strcspn($s, $charlist, $start = 0, $len = 2147483647)
    {
        if ('' === $charlist .= '') return null;
        if ($start || 2147483647 != $len) $s = self::substr($s, $start, $len);

        return preg_match('/^(.*?)' . self::rxClass($charlist) . '/us', $s, $len) ? grapheme_strlen($len[1]) : grapheme_strlen($s);
    }

    static function strpbrk($s, $charlist)
    {
        if (preg_match('/' . self::rxClass($charlist) . '/us', $s, $m)) return substr($s, strpos($s, $m[0]));
        else return false;
    }

    static function strrev($s)
    {
        $s = self::str_split($s);
        return implode('', array_reverse($s));
    }

    static function strspn($s, $mask, $start = 0, $len = 2147483647)
    {
        if ($start || 2147483647 != $len) $s = self::substr($s, $start, $len);
        return preg_match('/^' . self::rxClass($mask) . '+/u', $s, $s) ? grapheme_strlen($s[0]) : 0;
    }

    static function strtr($s, $from, $to = INF)
    {
        if (INF !== $to)
        {
            $from = self::str_split($from);
            $to   = self::str_split($to);

            $a = count($from);
            $b = count($to);

                 if ($a > $b) $from = array_slice($from, 0, $b);
            else if ($a < $b) $to   = array_slice($to  , 0, $a);

            $from = array_combine($from, $to);
        }

        return strtr($s, $from);
    }

    static function substr_compare($a, $b, $offset, $len = 2147483647, $i = 0)
    {
        $a = self::substr($a, $offset, $len);
        return $i ? static::strcasecmp($a, $b) : self::strcmp($a, $b);
    }

    static function substr_count($s, $needle, $offset = 0, $len = 2147483647)
    {
        return substr_count(self::substr($s, $offset, $len), $needle);
    }

    static function substr_replace($s, $replace, $start, $len = 2147483647)
    {
        $s       = self::str_split($s);
        $replace = self::str_split($replace);
        array_splice($s, $start, $len, $replace);
        return implode('', $s);
    }

    static function ucfirst($s)
    {
        $c = iconv_substr($s, 0, 1, 'UTF-8');
        return static::ucwords($c) . substr($s, strlen($c));
    }

    static function lcfirst($s)
    {
        $c = iconv_substr($s, 0, 1, 'UTF-8');
        return static::strtolower($c) . substr($s, strlen($c));
    }

    static function ucwords($s)
    {
        return mb_convert_case($s, MB_CASE_TITLE, 'UTF-8');
    }

    static function number_format($number, $decimals = 0, $dec_point = '.', $thousands_sep = ',')
    {
/**/    if (PHP_VERSION_ID < 50400)
/**/    {
            if (isset($thousands_sep[1]) || isset($dec_point[1]))
            {
                return str_replace(
                    array('.', ','),
                    array($dec_point, $thousands_sep),
                    number_format($number, $decimals, '.', ',')
                );
            }
/**/    }

        return number_format($number, $decimals, $dec_point, $thousands_sep);
    }

    static function utf8_encode($s)
    {
        $s = utf8_encode($s);
        if (false === strpos($s, "\xC2")) return $s;
        else return str_replace(self::$cp1252, self::$utf8, $s);
    }

    static function utf8_decode($s)
    {
        $s = str_replace(self::$utf8, self::$cp1252, $s);
        return utf8_decode($s);
    }


    protected static function rxClass($s, $class = '')
    {
        $class = array($class);

        foreach (self::str_split($s) as $s)
        {
            if ('-' === $s) $class[0] = '-' . $class[0];
            else if (!isset($s[2])) $class[0] .= preg_quote($s, '/');
            else if (1 === iconv_strlen($s, 'UTF-8')) $class[0] .= $s;
            else $class[] = $s;
        }

        $class[0] = '[' . $class[0] . ']';

        if (1 === count($class)) return $class[0];
        else return '(?:' . implode('|', $class) . ')';
    }

    protected static function getData($file)
    {
        $file = __DIR__ . '/Utf8/data/' . $file . '.ser';
        if (file_exists($file)) return unserialize(file_get_contents($file));
        else return false;
    }
}