feat: New features for clustering

sdcat/cluster/commands.py

@@ -25,7 +25,11 @@
 @click.option('--det-dir', help='Input folder(s) with raw detection results', multiple=True)
 @click.option('--save-dir', help='Output directory to save clustered detection results')
 @click.option('--device', help='Device to use.', type=int)
-def run_cluster(det_dir, save_dir, device, config_ini):
+@click.option('--alpha', help='Alpha is a parameter that controls the linkage.', type=float, default=0.92)
+@click.option('--cluster_selection_epsilon', help='Epsilon is a parameter that controls the linkage', type=float, default=0.0)
+@click.option('--min_cluster_size', help='The minimum number of samples in a group for that group to be considered a cluster', type=int, default=2)
+
+def run_cluster(det_dir, save_dir, device, config_ini, alpha, cluster_selection_epsilon, min_cluster_size):
     config = cfg.Config(config_ini)
     max_area = int(config('cluster', 'max_area'))
     min_area = int(config('cluster', 'min_area'))
@@ -36,9 +40,6 @@ def run_cluster(det_dir, save_dir, device, config_ini):
     longitude = float(config('cluster', 'longitude'))
     min_score = float(config('cluster', 'min_score'))
     min_similarity = float(config('cluster', 'min_similarity'))
-    alpha = float(config('cluster', 'alpha'))
-    cluster_selection_epsilon = float(config('cluster', 'cluster_selection_epsilon'))
-    min_cluster_size = int(config('cluster', 'min_cluster_size'))
     model = config('cluster', 'model')
 
     if device:

sdcat/cluster/utils.py

@@ -110,7 +110,7 @@ def crop_square_image(row, square_dim: int):
 
         if Path(row.crop_path).exists():  # If the crop already exists, skip it
             return
-
+        print("Image widht ", row.image_width)
         x1 = int(row.image_width * row.x)
         y1 = int(row.image_height * row.y)
         x2 = int(row.image_width * row.xx)

sdcat/detect/commands.py

@@ -298,7 +298,9 @@ def run_detect(show: bool, image_dir: str, save_dir: str, model: str,
         df_combined['w'] = (df_combined['xx'] - df_combined['x'])
         df_combined['h'] = (df_combined['xy'] - df_combined['y'])
 
+        # Save DataFrame to CSV file including image_width and image_height columns
         df_combined.to_csv(pred_out_csv.as_posix(), index=False, header=True)
+
         info(f'Found {len(pred_list)} total localizations in {f} with {len(df_combined)} after NMS')
         info(f'Slice width: {slice_size_width} height: {slice_size_height}')