fix: filtre historique, fuite mémoire OpenCV, export IA, tests

- Historique : le filtre par type de cible est désormais appliqué au
  chargement ; correction du piège PopupMenuItem(value: null) qui
  empêchait l'option « Tous » de réinitialiser le filtre ; icône
  colorée quand un filtre est actif
- OpenCV : libération des Mat natifs (img, gray, blurred, circles)
  dans un finally — detectTarget tourne toutes les secondes pendant
  l'aperçu caméra et faisait grimper la mémoire native en continu
- Export IA : distance, arme et id de session réels transmis depuis
  SessionProvider au lieu des placeholders (25 m / "Unknown")
- Tests : remplacement du test widget cassé (BullyApp sans providers)
  par 14 tests qui passent — calcul de score concentrique (centre,
  hors cible, ratio d'image, anneaux personnalisés), agrégation des
  scores, analyse de groupement, et rendu du widget StatsCard

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

lib/services/ai_export_service.dart

@@ -53,6 +53,8 @@ class AiExportService {
     required double targetCenterY,
     required double targetRadius,
     required List<Shot> shots,
+    int distanceMeters = 25,
+    String weaponName = 'Unknown',
     String? apiUrl,
   }) async {
     try {
@@ -109,9 +111,9 @@ class AiExportService {
         "device_info": deviceData,
         "target_metadata": {
           "type": targetType.name,
-          "distance_meters": 25, // Default/placeholder
-          "weapon": "Unknown", // Default/placeholder
-          // The backend could extract exact width/height from the image. 
+          "distance_meters": distanceMeters,
+          "weapon": weaponName,
+          // The backend could extract exact width/height from the image.
         },
         "plotting": {
           "target_corners": corners,

lib/services/opencv_target_service.dart

@@ -26,23 +26,34 @@ class TargetDetectionResult {
 
 class OpenCVTargetService {
   /// Detect the main target (center and radius) from an image file
+  ///
+  /// IMPORTANT : les Mat OpenCV sont de la mémoire NATIVE, invisible pour le
+  /// garbage collector Dart. Cette méthode est appelée en boucle (~1 s)
+  /// pendant l'aperçu caméra : sans dispose() explicite dans le finally, la
+  /// mémoire native grimpe en continu tant que l'utilisateur vise.
   Future<TargetDetectionResult> detectTarget(String imagePath) async {
+    cv.Mat? img;
+    cv.Mat? gray;
+    cv.Mat? blurred;
+    cv.Mat? circles;
+    cv.Mat? looseCircles;
+
     try {
       // Read image
-      final img = cv.imread(imagePath, flags: cv.IMREAD_COLOR);
+      img = cv.imread(imagePath, flags: cv.IMREAD_COLOR);
       if (img.isEmpty) {
         return TargetDetectionResult.failure();
       }
 
       // Convert to grayscale
-      final gray = cv.cvtColor(img, cv.COLOR_BGR2GRAY);
+      gray = cv.cvtColor(img, cv.COLOR_BGR2GRAY);
 
       // Apply Gaussian blur to reduce noise
-      final blurred = cv.gaussianBlur(gray, (9, 9), 2, sigmaY: 2);
+      blurred = cv.gaussianBlur(gray, (9, 9), 2, sigmaY: 2);
 
-      // Detect circles using Hough Transform
-      // Parameters need to be tuned for the specific target type
-      final circles = cv.HoughCircles(
+      // Detect circles using Hough Transform.
+      // HoughCircles returns a Mat of shape (1, N) of Vec3f (x, y, r).
+      circles = cv.HoughCircles(
         blurred,
         cv.HOUGH_GRADIENT,
         1, // dp
@@ -55,26 +66,9 @@ class OpenCVTargetService {
         maxRadius: img.cols ~/ 2,
       );
 
-      // HoughCircles returns a Mat of shape (1, N, 3) where N is number of circles.
-      // In opencv_dart, we cannot iterate easily.
-      // However, we can access data via pointer if needed, or check if Vec3f is supported.
-      // Given the user report, `at<Vec3f>` likely failed compilation or runtime.
-      // Let's use a safer approach: assume standard memory layout (x, y, r, x, y, r...).
-      // Or use `at<double>` carefully.
-
-      // Better yet: try to use `circles.data` if available, but it returns a Pointer.
-      // Let's stick to `at` but use `double` and manual offset if Vec3f fails.
-      // actually, let's try to trust `at<double>` for flattened access OR `at<Vec3f>`.
-      // NOTE: `at<Vec3f>` was reported as "method at not defined for VecPoint2f" earlier, NOT for Mat.
-      // The user error was for `VecPoint2f`. `Mat` definitely has `at`.
-      // BUT `VecPoint2f` is a List-like structure in Dart wrapper.
-      // usage of `at` on `VecPoint2f` was the error.
-      // Here `circles` IS A MAT. So `at` IS defined.
-      // However, to be safe and robust, and to implement clustering...
-
       if (circles.isEmpty) {
         // Try with different parameters if first attempt fails (more lenient)
-        final looseCircles = cv.HoughCircles(
+        looseCircles = cv.HoughCircles(
           blurred,
           cv.HOUGH_GRADIENT,
           1,
@@ -93,8 +87,15 @@ class OpenCVTargetService {
 
       return _findBestConcentricCircles(circles, img.cols, img.rows);
     } catch (e) {
-      // print('Error detecting target with OpenCV: $e');
       return TargetDetectionResult.failure();
+    } finally {
+      // _findBestConcentricCircles a déjà extrait les données dans des listes
+      // Dart avant qu'on arrive ici : libérer les Mat est donc toujours sûr.
+      img?.dispose();
+      gray?.dispose();
+      blurred?.dispose();
+      circles?.dispose();
+      looseCircles?.dispose();
     }
   }