<?php

declare(strict_types=1);

/**
 * Bidi.php
 *
 * @since     2011-05-23
 * @category  Library
 * @package   Unicode
 * @author    Nicola Asuni <info@tecnick.com>
 * @copyright 2011-2026 Nicola Asuni - Tecnick.com LTD
 * @license   https://www.gnu.org/copyleft/lesser.html GNU-LGPL v3 (see LICENSE)
 * @link      https://github.com/tecnickcom/tc-lib-unicode
 *
 * This file is part of tc-lib-unicode software library.
 */

namespace Com\Tecnick\Unicode;

use Com\Tecnick\Unicode\Bidi\Shaping;
use Com\Tecnick\Unicode\Bidi\StepI;
use Com\Tecnick\Unicode\Bidi\StepL;
use Com\Tecnick\Unicode\Bidi\StepN;
use Com\Tecnick\Unicode\Bidi\StepP;
use Com\Tecnick\Unicode\Bidi\StepW;
use Com\Tecnick\Unicode\Bidi\StepX;
use Com\Tecnick\Unicode\Bidi\StepXten;
use Com\Tecnick\Unicode\Data\Constant as UniConstant;
use Com\Tecnick\Unicode\Data\Type as UniType;
use Com\Tecnick\Unicode\Exception as UnicodeException;

/**
 * Com\Tecnick\Unicode\Bidi
 *
 * @since     2015-07-13
 * @category  Library
 * @package   Unicode
 * @author    Nicola Asuni <info@tecnick.com>
 * @copyright 2011-2026 Nicola Asuni - Tecnick.com LTD
 * @license   https://www.gnu.org/copyleft/lesser.html GNU-LGPL v3 (see LICENSE)
 * @link      https://github.com/tecnickcom/tc-lib-unicode
 */
class Bidi
{
    /**
     * The input contains strong right-to-left characters (bidi class R, AL or AN)
     */
    public const CONTAINS_RTL = 1;

    /**
     * The input contains Arabic characters (bidi class AL or AN)
     */
    public const CONTAINS_ARABIC = 2;

    /**
     * The input contains explicit formatting characters
     * (LRE, RLE, PDF, LRO, RLO, LRI, RLI, FSI, PDI)
     */
    public const CONTAINS_FORMATTING = 4;

    /**
     * String to process
     */
    protected string $str = '';

    /**
     * Array of UTF-8 chars
     *
     * @var array<string>
     */
    protected array $chrarr = [];

    /**
     * Array of UTF-8 codepoints
     *
     * @var array<int>
     */
    protected array $ordarr = [];

    /**
     * Processed string (null until it is built)
     */
    protected ?string $bidistr = null;

    /**
     * Array of processed UTF-8 chars (null until it is built)
     *
     * @var ?array<string>
     */
    protected ?array $bidichrarr = null;

    /**
     * Array of processed UTF-8 codepoints
     *
     * @var array<int>
     */
    protected array $bidiordarr = [];

    /**
     * If 'R' forces RTL, if 'L' forces LTR
     */
    protected string $forcedir = '';

    /**
     * If true enable shaping
     */
    protected bool $shaping = true;

    /**
     * Content flags of the input, as a combination of the CONTAINS_* constants
     */
    protected int $content = 0;

    /**
     * Convert object
     */
    protected Convert $conv;

    /**
     * Reverse the RTL substrings using the Bidirectional Algorithm
     * http://unicode.org/reports/tr9/
     *
     * @param ?string $str      String to convert (if null it will be generated from $chrarr or $ordarr)
     * @param ?array<string>  $chrarr   Array of UTF-8 chars (if empty it will be generated from $str or $ordarr)
     * @param ?array<int>  $ordarr   Array of UTF-8 codepoints (if empty it will be generated from $str or $chrarr)
     * @param string|TextDirection $forcedir If 'R' forces RTL, if 'L' forces LTR ('' auto), or a TextDirection case
     * @param bool   $shaping  If true enable the shaping algorithm
     *
     * @throws UnicodeException
     */
    public function __construct(
        ?string $str = null,
        ?array $chrarr = null,
        ?array $ordarr = null,
        string|TextDirection $forcedir = '',
        bool $shaping = true,
    ) {
        if ($str === null && ($chrarr === null || $chrarr === []) && ($ordarr === null || $ordarr === [])) {
            throw new UnicodeException('empty input');
        }

        $this->conv = new Convert();
        $this->setInput($str, $chrarr, $ordarr, $forcedir);
        $this->scanInput();

        // The explicit formatting characters are removed by X9, so an input that contains
        // them is processed even when it holds no right-to-left character.
        if (!$this->isRtlMode() && !$this->hasFormatting()) {
            $this->bidistr = $this->str;
            $this->bidichrarr = $this->chrarr;
            $this->bidiordarr = $this->ordarr;
            return;
        }

        $this->shaping = $shaping && $this->isArabic();

        $this->process();
    }

    /**
     * Set Input data
     *
     * @param ?string $str      String to convert (if null it will be generated from $chrarr or $ordarr)
     * @param ?array<string>  $chrarr   Array of UTF-8 chars (if empty it will be generated from $str or $ordarr)
     * @param ?array<int>  $ordarr   Array of UTF-8 codepoints (if empty it will be generated from $str or $chrarr)
     * @param string|TextDirection $forcedir If 'R' forces RTL, if 'L' forces LTR ('' auto), or a TextDirection case
     *
     * @throws UnicodeException
     * @SuppressWarnings("PHPMD.CyclomaticComplexity")
     */
    protected function setInput(
        ?string $str = null,
        ?array $chrarr = null,
        ?array $ordarr = null,
        string|TextDirection $forcedir = '',
    ): void {
        if ($str === null) {
            $str = '';
            if (($chrarr === null || $chrarr === []) && ($ordarr !== null && $ordarr !== [])) {
                $chrarr = $this->conv->ordArrToChrArr($ordarr);
            }
            if ($chrarr !== null && $chrarr !== []) {
                $str = \implode('', $chrarr);
            }
        }

        if ($chrarr === null || $chrarr === []) {
            $chrarr = $this->conv->strToChrArr($str);
        }

        if ($ordarr === null || $ordarr === []) {
            $ordarr = $this->conv->chrArrToOrdArr($chrarr);
        }

        if (\count($chrarr) !== \count($ordarr)) {
            throw new UnicodeException('the input forms contain a different number of characters');
        }

        $this->str = $str;
        $this->chrarr = $chrarr;
        $this->ordarr = $ordarr;
        $this->forcedir = TextDirection::fromLoose($forcedir)->value;
    }

    /**
     * Classify the input once into the content flags used to decide whether the
     * bidirectional algorithm and the shaping have to run.
     */
    protected function scanInput(): void
    {
        foreach ($this->ordarr as $ord) {
            if (
                $ord >= UniConstant::LRE && $ord <= UniConstant::RLO
                || $ord >= UniConstant::LRI && $ord <= UniConstant::PDI
            ) {
                $this->content |= self::CONTAINS_FORMATTING;
                continue;
            }

            $type = UniType::getType($ord);
            if ($type === 'AL' || $type === 'AN') {
                $this->content |= self::CONTAINS_ARABIC | self::CONTAINS_RTL;
                continue;
            }

            if ($type === 'R') {
                $this->content |= self::CONTAINS_RTL;
            }
        }
    }

    /**
     * Returns the processed array of UTF-8 codepoints
     *
     * @return array<int>
     */
    public function getOrdArray(): array
    {
        return $this->bidiordarr;
    }

    /**
     * Returns the processed array of UTF-8 chars
     *
     * @return array<string>
     *
     * @throws UnicodeException
     */
    public function getChrArray(): array
    {
        if ($this->bidichrarr === null) {
            $this->bidichrarr = $this->conv->ordArrToChrArr($this->bidiordarr);
        }

        return $this->bidichrarr;
    }

    /**
     * Returns the number of characters in the processed string
     *
     * @throws UnicodeException
     */
    public function getNumChars(): int
    {
        return \count($this->getChrArray());
    }

    /**
     * Returns the processed string
     *
     * @throws UnicodeException
     */
    public function getString(): string
    {
        if ($this->bidistr === null) {
            $this->bidistr = \implode('', $this->getChrArray());
        }

        return $this->bidistr;
    }

    /**
     * Returns an array with processed chars as keys
     *
     * @return array<int, true>
     */
    public function getCharKeys(): array
    {
        return \array_fill_keys($this->bidiordarr, true);
    }

    /**
     * P1. Split the text into separate paragraphs.
     *     A paragraph separator is kept with the previous paragraph.
     *
     * @return array<int, array<int, int>>
     */
    protected function getParagraphs(): array
    {
        $paragraph = [
            0 => [],
        ];
        $pdx = 0; // paragraphs index
        foreach ($this->ordarr as $ord) {
            $paragraph[$pdx][] = $ord;
            if (UniType::getType($ord) === 'B') {
                ++$pdx;
                $paragraph[$pdx] = [];
            }
        }

        return $paragraph;
    }

    /**
     * Process the string
     *
     * @SuppressWarnings("PHPMD.CyclomaticComplexity")
     */
    protected function process(): void
    {
        // split the text into separate paragraphs.
        $paragraph = $this->getParagraphs();

        // Within each paragraph, apply all the other rules of this algorithm.
        foreach ($paragraph as $par) {
            // A trailing paragraph separator produces an empty final paragraph; it contributes
            // nothing and would otherwise run through the full Step* pipeline for no result.
            if ($par === []) {
                continue;
            }

            $pel = $this->getPel($par);
            $stepx = new StepX($par, $pel);
            $stepx10 = new StepXten($stepx->getChrData(), $pel);
            $ilrs = $stepx10->getIsolatedLevelRunSequences();
            $chardata = [];
            $maxlevel = 0;
            foreach ($ilrs as $ilr) {
                $stepw = new StepW($ilr);
                $stepn = new StepN($stepw->getSequence());
                $stepi = new StepI($stepn->getSequence());
                $ilr = $stepi->getSequence();
                if ($this->shaping) {
                    $shaping = new Shaping($ilr, $par);
                    $ilr = $shaping->getSequence();
                }

                \array_push($chardata, ...$ilr['item']);

                if ($ilr['maxlevel'] > $maxlevel) {
                    $maxlevel = $ilr['maxlevel'];
                }
            }

            $stepl = new StepL($chardata, $pel, $maxlevel);
            $chardata = $stepl->getChrData();
            foreach ($chardata as $chardatum) {
                $this->bidiordarr[] = $chardatum['char'];
            }

            // add back the paragraph separators ($par is non-empty, so end() returns a codepoint)
            $lastchar = \end($par);
            if ($lastchar < 0) {
                continue;
            }

            if (UniType::getType($lastchar) !== 'B') {
                continue;
            }

            $this->bidiordarr[] = $lastchar;
        }
    }

    /**
     * Get the paragraph embedding level
     *
     * @param array<int> $par Paragraph
     */
    protected function getPel(array $par): int
    {
        if ($this->forcedir === 'R') {
            return 1;
        }

        if ($this->forcedir === 'L') {
            return 0;
        }

        $stepp = new StepP($par);
        return $stepp->getPel();
    }

    /**
     * Check if the input contains Arabic characters
     */
    protected function isArabic(): bool
    {
        return ($this->content & self::CONTAINS_ARABIC) !== 0;
    }

    /**
     * Check if the input contains explicit formatting characters
     */
    protected function hasFormatting(): bool
    {
        return ($this->content & self::CONTAINS_FORMATTING) !== 0;
    }

    /**
     * Check if the input contains right-to-left characters to process
     */
    protected function isRtlMode(): bool
    {
        return $this->forcedir === 'R' || ($this->content & self::CONTAINS_RTL) !== 0;
    }
}