<?php

/*
 * 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\PHP\Shim;

/**
 * Normalizer is a PHP fallback implementation of the Normalizer class provided by the intl extension.
 *
 * It has been validated with Unicode 6.3 Normalization Conformance Test.
 * See http://www.unicode.org/reports/tr15/ for detailed info about Unicode normalizations.
 */
class Normalizer
{
    const NONE = 1;
    const FORM_D  = 2;
    const FORM_KD = 3;
    const FORM_C  = 4;
    const FORM_KC = 5;
    const NFD  = 2;
    const NFKD = 3;
    const NFC  = 4;
    const NFKC = 5;

    protected static $C;
    protected static $D;
    protected static $KD;
    protected static $cC;
    protected static $ulen_mask = array("\xC0" => 2, "\xD0" => 2, "\xE0" => 3, "\xF0" => 4);
    protected static $ASCII = "\x20\x65\x69\x61\x73\x6E\x74\x72\x6F\x6C\x75\x64\x5D\x5B\x63\x6D\x70\x27\x0A\x67\x7C\x68\x76\x2E\x66\x62\x2C\x3A\x3D\x2D\x71\x31\x30\x43\x32\x2A\x79\x78\x29\x28\x4C\x39\x41\x53\x2F\x50\x22\x45\x6A\x4D\x49\x6B\x33\x3E\x35\x54\x3C\x44\x34\x7D\x42\x7B\x38\x46\x77\x52\x36\x37\x55\x47\x4E\x3B\x4A\x7A\x56\x23\x48\x4F\x57\x5F\x26\x21\x4B\x3F\x58\x51\x25\x59\x5C\x09\x5A\x2B\x7E\x5E\x24\x40\x60\x7F\x00\x01\x02\x03\x04\x05\x06\x07\x08\x0B\x0C\x0D\x0E\x0F\x10\x11\x12\x13\x14\x15\x16\x17\x18\x19\x1A\x1B\x1C\x1D\x1E\x1F";

    public static function isNormalized($s, $form = self::NFC)
    {
        if (strspn($s .= '', self::$ASCII) === strlen($s)) {
            return true;
        }
        if (self::NFC === $form && preg_match('//u', $s) && !preg_match('/[^\x00-\x{2FF}]/u', $s)) {
            return true;
        }

        return false; // Pretend false as quick checks implementented in PHP won't be so quick
    }

    public static function normalize($s, $form = self::NFC)
    {
        if (!preg_match('//u', $s .= '')) {
            return false;
        }

        switch ($form) {
            case self::NONE: return $s;
            case self::NFC:  $C = true;  $K = false; break;
            case self::NFD:  $C = false; $K = false; break;
            case self::NFKC: $C = true;  $K = true;  break;
            case self::NFKD: $C = false; $K = true;  break;
            default: return false;
        }

        if ('' === $s) {
            return '';
        }

        if ($K && empty(self::$KD)) {
            self::$KD = static::getData('compatibilityDecomposition');
        }

        if (empty(self::$D)) {
            self::$D = static::getData('canonicalDecomposition');
            self::$cC = static::getData('combiningClass');
        }

        if ($C) {
            if (empty(self::$C)) {
                self::$C = static::getData('canonicalComposition');
            }

            return self::recompose(self::decompose($s, $K));
        } else {
            return self::decompose($s, $K);
        }
    }

    protected static function recompose($s)
    {
        $ASCII = self::$ASCII;
        $compMap = self::$C;
        $combClass = self::$cC;
        $ulen_mask = self::$ulen_mask;

        $result = $tail = '';

        $i = $s[0] < "\x80" ? 1 : $ulen_mask[$s[0] & "\xF0"];
        $len = strlen($s);

        $last_uchr = substr($s, 0, $i);
        $last_ucls = isset($combClass[$last_uchr]) ? 256 : 0;

        while ($i < $len) {
            if ($s[$i] < "\x80") {
                // ASCII chars

                if ($tail) {
                    $last_uchr .= $tail;
                    $tail = '';
                }

                if ($j = strspn($s, $ASCII, $i + 1)) {
                    $last_uchr .= substr($s, $i, $j);
                    $i += $j;
                }

                $result .= $last_uchr;
                $last_uchr = $s[$i];
                ++$i;
            } else {
                $ulen = $ulen_mask[$s[$i] & "\xF0"];
                $uchr = substr($s, $i, $ulen);

                if ($last_uchr < "\xE1\x84\x80" || "\xE1\x84\x92" < $last_uchr
                    ||   $uchr < "\xE1\x85\xA1" || "\xE1\x85\xB5" < $uchr
                    || $last_ucls) {
                    // Table lookup and combining chars composition

                    $ucls = isset($combClass[$uchr]) ? $combClass[$uchr] : 0;

                    if (isset($compMap[$last_uchr.$uchr]) && (!$last_ucls || $last_ucls < $ucls)) {
                        $last_uchr = $compMap[$last_uchr.$uchr];
                    } elseif ($last_ucls = $ucls) {
                        $tail .= $uchr;
                    } else {
                        if ($tail) {
                            $last_uchr .= $tail;
                            $tail = '';
                        }

                        $result .= $last_uchr;
                        $last_uchr = $uchr;
                    }
                } else {
                    // Hangul chars

                    $L = ord($last_uchr[2]) - 0x80;
                    $V = ord($uchr[2]) - 0xA1;
                    $T = 0;

                    $uchr = substr($s, $i + $ulen, 3);

                    if ("\xE1\x86\xA7" <= $uchr && $uchr <= "\xE1\x87\x82") {
                        $T = ord($uchr[2]) - 0xA7;
                        0 > $T && $T += 0x40;
                        $ulen += 3;
                    }

                    $L = 0xAC00 + ($L * 21 + $V) * 28 + $T;
                    $last_uchr = chr(0xE0 | $L >> 12).chr(0x80 | $L >> 6 & 0x3F).chr(0x80 | $L & 0x3F);
                }

                $i += $ulen;
            }
        }

        return $result.$last_uchr.$tail;
    }

    protected static function decompose($s, $c)
    {
        $result = '';

        $ASCII = self::$ASCII;
        $decompMap = self::$D;
        $combClass = self::$cC;
        $ulen_mask = self::$ulen_mask;
        if ($c) {
            $compatMap = self::$KD;
        }

        $c = array();
        $i = 0;
        $len = strlen($s);

        while ($i < $len) {
            if ($s[$i] < "\x80") {
                // ASCII chars

                if ($c) {
                    ksort($c);
                    $result .= implode('', $c);
                    $c = array();
                }

                $j = 1 + strspn($s, $ASCII, $i + 1);
                $result .= substr($s, $i, $j);
                $i += $j;
            } else {
                $ulen = $ulen_mask[$s[$i] & "\xF0"];
                $uchr = substr($s, $i, $ulen);
                $i += $ulen;

                if (isset($combClass[$uchr])) {
                    // Combining chars, for sorting

                    isset($c[$combClass[$uchr]]) or $c[$combClass[$uchr]] = '';
                    $c[$combClass[$uchr]] .= isset($compatMap[$uchr]) ? $compatMap[$uchr] : (isset($decompMap[$uchr]) ? $decompMap[$uchr] : $uchr);
                } else {
                    if ($c) {
                        ksort($c);
                        $result .= implode('', $c);
                        $c = array();
                    }

                    if ($uchr < "\xEA\xB0\x80" || "\xED\x9E\xA3" < $uchr) {
                        // Table lookup

                        $j = isset($compatMap[$uchr]) ? $compatMap[$uchr] : (isset($decompMap[$uchr]) ? $decompMap[$uchr] : $uchr);

                        if ($uchr != $j) {
                            $uchr = $j;

                            $j = strlen($uchr);
                            $ulen = $uchr[0] < "\x80" ? 1 : $ulen_mask[$uchr[0] & "\xF0"];

                            if ($ulen != $j) {
                                // Put trailing chars in $s

                                $j -= $ulen;
                                $i -= $j;

                                if (0 > $i) {
                                    $s = str_repeat(' ', -$i).$s;
                                    $len -= $i;
                                    $i = 0;
                                }

                                while ($j--) {
                                    $s[$i + $j] = $uchr[$ulen + $j];
                                }

                                $uchr = substr($uchr, 0, $ulen);
                            }
                        }
                    } else {
                        // Hangul chars

                        $uchr = unpack('C*', $uchr);
                        $j = (($uchr[1] - 224) << 12) + (($uchr[2] - 128) << 6) + $uchr[3] - 0xAC80;

                        $uchr = "\xE1\x84".chr(0x80 + (int) ($j / 588))
                              ."\xE1\x85".chr(0xA1 + (int) (($j % 588) / 28));

                        if ($j %= 28) {
                            $uchr .= $j < 25
                                ? ("\xE1\x86".chr(0xA7 + $j))
                                : ("\xE1\x87".chr(0x67 + $j));
                        }
                    }

                    $result .= $uchr;
                }
            }
        }

        if ($c) {
            ksort($c);
            $result .= implode('', $c);
        }

        return $result;
    }

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