<?php
namespace ActiveTwist\SEO\Utils;

class Markdown
{
    public static function convert($html)
    {
        if (empty($html)) {
            return '';
        }

        // Suppress warnings for malformed HTML
        $dom = new \DOMDocument();
        $html = mb_convert_encoding($html, 'HTML-ENTITIES', 'UTF-8');
        @$dom->loadHTML($html, LIBXML_HTML_NOIMPLIED | LIBXML_HTML_NODEFDTD);

        return self::process_node($dom);
    }

    private static function process_node($node)
    {
        if ($node instanceof \DOMText) {
            return $node->wholeText; // Preserve whitespace logic in render
        }

        if (!$node instanceof \DOMElement && !$node instanceof \DOMDocument) {
            return '';
        }

        $output = '';
        foreach ($node->childNodes as $child) {
            $output .= self::process_node($child);
        }

        if ($node instanceof \DOMDocument) {
            return trim($output);
        }

        $tag = strtolower($node->tagName);

        switch ($tag) {
            case 'h1':
                return "\n# " . trim($output) . "\n\n";
            case 'h2':
                return "\n## " . trim($output) . "\n\n";
            case 'h3':
                return "\n### " . trim($output) . "\n\n";
            case 'h4':
                return "\n#### " . trim($output) . "\n\n";
            case 'h5':
                return "\n##### " . trim($output) . "\n\n";
            case 'h6':
                return "\n###### " . trim($output) . "\n\n";

            case 'p':
                if (trim($output) === '')
                    return '';
                return trim($output) . "\n\n";

            case 'br':
                return "  \n";

            case 'hr':
                return "\n---\n\n";

            case 'strong':
            case 'b':
                // Extract whitespace
                $leading = '';
                $trailing = '';
                if (preg_match('/^(\s+)/', $output, $matches)) {
                    $leading = $matches[1];
                }
                if (preg_match('/(\s+)$/', $output, $matches)) {
                    $trailing = $matches[1];
                }
                $trimmed = trim($output);
                if ($trimmed === '')
                    return $output; // Just whitespace or empty
                return $leading . "**" . $trimmed . "**" . $trailing;

            case 'em':
            case 'i':
                // Extract whitespace
                $leading = '';
                $trailing = '';
                if (preg_match('/^(\s+)/', $output, $matches)) {
                    $leading = $matches[1];
                }
                if (preg_match('/(\s+)$/', $output, $matches)) {
                    $trailing = $matches[1];
                }
                $trimmed = trim($output);
                if ($trimmed === '')
                    return $output; // Just whitespace or empty
                return $leading . "*" . $trimmed . "*" . $trailing;

            case 'code':
                // Check if inside pre (handled by pre)
                if ($node->parentNode && $node->parentNode instanceof \DOMElement && strtolower($node->parentNode->tagName) === 'pre') {
                    return $output;
                }
                return "`" . $output . "`";

            case 'pre':
                // Attempt to guess language or just use generic
                return "\n```\n" . trim($output) . "\n```\n\n";

            case 'blockquote':
                $lines = explode("\n", trim($output));
                $quoted = array_map(function ($line) {
                    return "> " . $line;
                }, $lines);
                return "\n" . implode("\n", $quoted) . "\n\n";

            case 'ul':
            case 'ol':
                return "\n" . trim($output) . "\n\n";

            case 'li':
                // Determine list type from parent
                $marker = '-';
                if ($node->parentNode && $node->parentNode instanceof \DOMElement && strtolower($node->parentNode->tagName) === 'ol') {
                    // Calculate index
                    $index = 1;
                    $prev = $node->previousSibling;
                    while ($prev) {
                        if ($prev instanceof \DOMElement && strtolower($prev->tagName) === 'li') {
                            $index++;
                        }
                        $prev = $prev->previousSibling;
                    }
                    $marker = "$index.";
                }
                return $marker . " " . trim($output) . "\n";

            case 'a':
                $href = $node->getAttribute('href');
                $title = $node->getAttribute('title');
                $text = $output;
                if (!$href)
                    return $text;
                $markdown = "[$text]($href";
                if ($title) {
                    $markdown .= " \"$title\"";
                }
                $markdown .= ")";
                return $markdown;

            case 'img':
                $src = $node->getAttribute('src');
                $alt = $node->getAttribute('alt');
                $title = $node->getAttribute('title');
                if (!$src)
                    return '';
                $markdown = "![$alt]($src";
                if ($title) {
                    $markdown .= " \"$title\"";
                }
                $markdown .= ")";
                return $markdown;

            case 'div':
            case 'span':
            case 'article':
            case 'section':
                return $output; // Just pass through text content

            default:
                return $output;
        }
    }
}