TicketAssistent/app/Services/TicketNormalizationService.php

<?php

namespace App\Services;

use App\Services\Llm\LlmClientInterface;

class TicketNormalizationService
{
    public function __construct(
        private readonly LlmClientInterface $llmClient,
        private readonly AppSettingsService $settings,
    ) {}

    public function normalize(string $original): array
    {
        $fallback = $this->fallbackNormalize($original);
        $fallbackLanguage = $this->detectLanguage($original);

        if (! (bool) config('services.llm.ranking_enabled', true)) {
            return [
                'normalized_message' => $fallback['text'],
                'redaction_report' => [
                    'mode' => 'fallback_regex',
                    'pii_types' => $fallback['pii_types'],
                    'language' => $fallbackLanguage,
                    'reason' => 'llm_normalization_disabled',
                ],
            ];
        }

        $basePrompt = $this->settings->getPrompt('normalization', 'Rewrite and redact PII. Return JSON.');
        $prompt = $basePrompt."\n\n".
            'Also detect the original language. Return JSON with keys: normalized_message, redaction_report. '.
            "redaction_report must include pii_types and language as an ISO 639-1 code such as nl, en, de, fr.\n\n".
            "Original question:\n\"\"\"\n{$original}\n\"\"\"";

        try {
            $raw = $this->llmClient->generate($prompt, ['expect_json' => true, 'task' => 'normalization']);
            $decoded = $this->decodeJsonResponse($raw);

            if (is_array($decoded) && ! empty($decoded['normalized_message'])) {
                return [
                    'normalized_message' => (string) $decoded['normalized_message'],
                    'redaction_report' => [
                        'mode' => 'llm',
                        'pii_types' => $decoded['redaction_report']['pii_types'] ?? [],
                        'language' => $this->normalizeLanguageCode($decoded['redaction_report']['language'] ?? $fallbackLanguage),
                        'notes' => $decoded['redaction_report']['notes'] ?? null,
                        'raw' => $decoded,
                    ],
                ];
            }
        } catch (\Throwable $e) {
            return [
                'normalized_message' => $fallback['text'],
                'redaction_report' => [
                    'mode' => 'fallback_regex',
                    'pii_types' => $fallback['pii_types'],
                    'language' => $fallbackLanguage,
                    'reason' => 'llm_exception',
                    'error' => $e->getMessage(),
                ],
            ];
        }

        return [
            'normalized_message' => $fallback['text'],
            'redaction_report' => [
                'mode' => 'fallback_regex',
                'pii_types' => $fallback['pii_types'],
                'language' => $fallbackLanguage,
                'reason' => 'llm_invalid_json_or_missing_fields',
            ],
        ];
    }

    private function fallbackNormalize(string $text): array
    {
        $pii = [];

        // Replace highly-structured values first so looser patterns (phone) do not corrupt them.
        $orderedPatterns = [
            'email' => '/[A-Z0-9._%+-]+@[A-Z0-9.-]+\.[A-Z]{2,}/i',
            'iban' => '/\b[A-Z]{2}[0-9]{2}[A-Z0-9]{10,30}\b/i',
            'url' => '/https?:\/\/\S+/i',
            'ip' => '/\b(?:\d{1,3}\.){3}\d{1,3}\b/',
            // Require separators to avoid matching long account-like sequences.
            'phone' => '/(?<![A-Z0-9])(?:\+?\d{1,3}[\s\-]?)?(?:\(?\d{2,4}\)?[\s\-])\d[\d\s\-]{4,}\d(?![A-Z0-9])/i',
        ];

        foreach ($orderedPatterns as $type => $pattern) {
            if (preg_match($pattern, $text) === 1) {
                $pii[] = $type;
                $text = preg_replace($pattern, '['.strtoupper($type).']', $text) ?? $text;
            }
        }

        $text = preg_replace('/\s+/', ' ', trim($text)) ?? trim($text);

        return ['text' => $text, 'pii_types' => array_values(array_unique($pii))];
    }

    private function detectLanguage(string $text): string
    {
        $lower = mb_strtolower($text);
        $dutchSignals = [' ik ', ' mijn ', ' een ', ' het ', ' de ', ' hoe ', ' niet ', ' wordt ', ' domeinnaam ', ' website '];
        $englishSignals = [' i ', ' my ', ' the ', ' how ', ' not ', ' website ', ' domain ', ' redirected '];

        $padded = ' '.$lower.' ';
        $nl = 0;
        $en = 0;

        foreach ($dutchSignals as $signal) {
            $nl += substr_count($padded, $signal);
        }

        foreach ($englishSignals as $signal) {
            $en += substr_count($padded, $signal);
        }

        return $en > $nl ? 'en' : 'nl';
    }

    private function normalizeLanguageCode(mixed $language): string
    {
        $value = mb_strtolower(trim((string) $language));

        return match (true) {
            str_starts_with($value, 'nl'), str_contains($value, 'dutch'), str_contains($value, 'nederlands') => 'nl',
            str_starts_with($value, 'en'), str_contains($value, 'english') => 'en',
            str_starts_with($value, 'de'), str_contains($value, 'german'), str_contains($value, 'duits') => 'de',
            str_starts_with($value, 'fr'), str_contains($value, 'french'), str_contains($value, 'frans') => 'fr',
            default => 'nl',
        };
    }

    private function decodeJsonResponse(string $raw): ?array
    {
        $raw = trim($raw);
        $decoded = json_decode($raw, true);
        if (is_array($decoded)) {
            return $decoded;
        }

        if (preg_match('/```(?:json)?\s*(\{.*\})\s*```/is', $raw, $matches) === 1) {
            $decoded = json_decode(trim($matches[1]), true);
            if (is_array($decoded)) {
                return $decoded;
            }
        }

        $start = strpos($raw, '{');
        $end = strrpos($raw, '}');
        if ($start !== false && $end !== false && $end > $start) {
            $candidate = substr($raw, $start, $end - $start + 1);
            $decoded = json_decode($candidate, true);
            if (is_array($decoded)) {
                return $decoded;
            }
        }

        return null;
    }
}