<?php

namespace App\Services\Moderation\Rules;

use App\Contracts\Moderation\ModerationRuleInterface;

class UnicodeObfuscationRule implements ModerationRuleInterface
{
    public function analyze(string $content, string $normalized, array $context = []): array
    {
        $findings = [];
        $weights = app('config')->get('content_moderation.weights', []);

        // Detect homoglyph / lookalike characters
        // Common spam tactic: replace Latin chars with Cyrillic, Greek, or special Unicode
        $suspiciousPatterns = [
            // Mixed script detection: Latin + Cyrillic in same word
            '/\b(?=\S*[\x{0400}-\x{04FF}])(?=\S*[a-zA-Z])\S+\b/u',
            // Zero-width characters
            '/[\x{200B}\x{200C}\x{200D}\x{FEFF}\x{00AD}]/u',
            // Invisible formatting characters
            '/[\x{2060}\x{2061}\x{2062}\x{2063}\x{2064}]/u',
            // Fullwidth Latin letters (used to bypass filters)
            '/[\x{FF01}-\x{FF5E}]/u',
            // Mathematical alphanumeric symbols used as text
            '/[\x{1D400}-\x{1D7FF}]/u',
        ];

        $matchCount = 0;
        foreach ($suspiciousPatterns as $pattern) {
            if (preg_match($pattern, $content)) {
                $matchCount++;
            }
        }

        if ($matchCount > 0) {
            $findings[] = [
                'rule'     => 'unicode_obfuscation',
                'score'    => ($weights['unicode_obfuscation'] ?? 30) * $matchCount,
                'reason'   => 'Contains suspicious Unicode characters/obfuscation (' . $matchCount . ' pattern(s) matched)',
                'links'    => [],
                'domains'  => [],
                'keywords' => [],
            ];
        }

        return $findings;
    }
}