MIF_E31230910_MP-HRIS-WEB/app/Services/FaceRecognitionService.php

195 lines
6.4 KiB
PHP

<?php
namespace App\Services;
use App\Enums\StatusVerifikasiWajah;
use Illuminate\Support\Facades\DB;
use Illuminate\Support\Facades\Log;
use Illuminate\Support\Facades\Http;
use Illuminate\Http\UploadedFile;
use Illuminate\Support\Facades\Storage;
use App\Jobs\ProcessFaceEnrollment;
class FaceRecognitionService
{
private function getFlaskUrl(): string
{
return config('services.flask.url', 'http://127.0.0.1:5000');
}
private function getFlaskHeaders(): array
{
return [
'X-API-Key' => config('services.flask.api_key', ''),
];
}
public function enrollFace($userId, UploadedFile $video)
{
$user = DB::table('users')->where('id', $userId)->first();
if (!$user) {
throw new \Exception('User tidak ditemukan.', 404);
}
$existing = DB::table('data_wajah')->where('id_user', $userId)->first();
if ($existing && $existing->is_verified == StatusVerifikasiWajah::APPROVED) {
if ($existing->face_embeddings !== null) {
throw new \Exception('Wajah Anda sudah terverifikasi. Hubungi HRD untuk reset.', 400);
}
Log::warning("User {$userId} is_verified=APPROVED tapi embedding kosong, izinkan re-enrollment.");
}
$videoStoragePath = "face_videos/{$userId}";
if (Storage::disk('local')->exists($videoStoragePath)) {
Storage::disk('local')->deleteDirectory($videoStoragePath);
}
Storage::disk('local')->makeDirectory($videoStoragePath);
$video->storeAs($videoStoragePath, 'enrollment.mp4', 'local');
DB::table('data_wajah')->updateOrInsert(
['id_user' => $userId],
[
'path_model_yml' => null,
'path_model_pkl' => null,
'path_scaler_pkl' => null,
'path_video' => "{$videoStoragePath}/enrollment.mp4",
'jumlah_frame' => null,
'face_embeddings' => null,
'jumlah_embedding' => null,
'embedding_generated_at' => null,
'is_verified' => StatusVerifikasiWajah::PENDING,
'last_updated' => now(),
]
);
DB::table('users')
->where('id', $userId)
->update(['is_face_registered' => 1]);
ProcessFaceEnrollment::dispatch($userId, $videoStoragePath);
return [
'status' => true,
'message' => 'Video wajah berhasil dikirim. Proses ekstraksi sedang berjalan. Setelah selesai, HRD akan memverifikasi data wajah Anda.',
];
}
/**
* Verifikasi wajah menggunakan pendekatan embedding + cosine similarity.
* File dikirim ke Flask untuk generate embedding, lalu dibandingkan dengan database.
*/
public function verifyFace($userId, UploadedFile $file)
{
$probeEmbedding = $this->getEmbedding($file);
$allEmbeddings = DB::table('data_wajah')
->where('is_verified', StatusVerifikasiWajah::APPROVED)
->whereNotNull('face_embeddings')
->get(['id_user', 'face_embeddings']);
if ($allEmbeddings->isEmpty()) {
throw new \Exception('Belum ada data wajah terverifikasi di sistem.');
}
$bestMatch = null;
$bestScore = -1;
foreach ($allEmbeddings as $data) {
$storedEmbedding = json_decode($data->face_embeddings, true);
if (!is_array($storedEmbedding)) continue;
$similarity = $this->cosineSimilarity($probeEmbedding, $storedEmbedding);
if ($similarity > $bestScore) {
$bestScore = $similarity;
$bestMatch = $data->id_user;
}
}
$threshold = 0.6;
$isMatch = $bestScore >= $threshold && $bestMatch == $userId;
$verificationStatus = 'REJECTED';
if ($isMatch) {
$verificationStatus = 'MATCH';
} elseif ($bestScore >= $threshold && $bestMatch != $userId) {
$verificationStatus = 'MISMATCH';
}
Log::info("[FaceVerify] User {$userId}: best_match={$bestMatch}, score={$bestScore}, status={$verificationStatus}");
return [
'verified' => $isMatch,
'confidence' => round($bestScore, 4),
'svm_df' => null,
'verification_status' => $verificationStatus,
'blur_score' => $probeEmbedding['blur_score'] ?? null,
'predicted_user' => $bestMatch,
'expected_user' => $userId,
'message' => $isMatch ? 'Wajah cocok' : 'Wajah tidak cocok',
];
}
private function getEmbedding(UploadedFile $file): array
{
$isVideo = str_starts_with($file->getMimeType() ?? '', 'video/');
$response = Http::timeout(120)
->withHeaders($this->getFlaskHeaders())
->attach(
'file',
fopen($file->getRealPath(), 'r'),
$isVideo ? 'verify.mp4' : 'verify.jpg'
)
->post($this->getFlaskUrl() . '/get-embedding', [
'is_video' => $isVideo ? 'true' : 'false',
]);
if (!$response->successful()) {
$body = $response->json();
$msg = $body['message'] ?? 'Flask ML API tidak merespons.';
Log::error("[FaceVerify] Flask /get-embedding error: HTTP {$response->status()} - {$msg}");
throw new \Exception("Gagal mengekstrak embedding wajah. {$msg}");
}
$output = $response->json();
if (!$output || $output['status'] !== 'success' || !isset($output['embedding'])) {
throw new \Exception('Respons embedding tidak valid.');
}
return $output;
}
private function cosineSimilarity(array $embeddingResponse, array $storedEmbedding): float
{
$a = $embeddingResponse['embedding'] ?? $embeddingResponse;
$b = $storedEmbedding;
if (count($a) !== count($b) || count($a) === 0) {
return 0.0;
}
$dotProduct = 0.0;
$normA = 0.0;
$normB = 0.0;
for ($i = 0; $i < count($a); $i++) {
$dotProduct += $a[$i] * $b[$i];
$normA += $a[$i] * $a[$i];
$normB += $b[$i] * $b[$i];
}
$normA = sqrt($normA);
$normB = sqrt($normB);
if ($normA == 0 || $normB == 0) {
return 0.0;
}
return $dotProduct / ($normA * $normB);
}
}