some cleaning and comments

main.py

@@ -9,7 +9,7 @@ from scipy.spatial import distance
 import argparse
 
 
-def generate_folders( dataset_path, name, list_categories, scenario):
+def generate_folders( dataset_path, name, list_categories):
     full_name = dataset_path + '/' + name
     is_exist = os.path.exists(full_name)
     if not is_exist:
@@ -68,37 +68,30 @@ if __name__ == '__main__':
     parser.add_argument('--dataset_id', type=str, default='', required=True)
     parser.add_argument('--occlusion_target_min', type=float, default='', required=True)
     parser.add_argument('--occlusion_target_max', type=float, default='', required=True)
-    #parser.add_argument('--rearrange', dest='rearrange', default=False, action='store_true')
-    #parser.add_argument('--compute', dest='compute', default=False, action='store_true')
     parser.add_argument('--rearrange', type=str, default='no', required=True)
     parser.add_argument('--compute', type=str, default='no', required=True)
     # Parse the argument
     args = parser.parse_args()
 
-    scenario = "Worlds"
-
     ### parameters ###
     Categories = [] # to read
     Nb_instance = 1
     occ_target_min = args.occlusion_target_min
     occ_target_max = args.occlusion_target_max
 
-    dataset_src = f"/gpfsscratch/rech/uli/ubn15wo/DATA/data{args.dataset_id}"
-    #dataset_src = "/media/gduret/DATA/dataset/s2rg/Fruits_all_medium/data"
+    dataset_src = f"/gpfsscratch/rech/uli/ubn15wo/DATA/data{args.dataset_id}" #TODO, path of the raw data to process.
 
-    choice = "low" # depth of rgb resolution datas
+    choice = "low" # depth of rgb resolution datas #TODO, low is the adviced value. 
     data_options = {"high": "ground_truth_rgb",
                     "low": "ground_truth_depth"}
     dataset_type = data_options[choice]
-    dataset_path = f"/gpfsscratch/rech/uli/ubn15wo/FruitBin{args.dataset_id}" #GUIMOD_New_{choice}_{args.dataset_id}"
-    #dataset_path = f"/home/gduret/Documents/FruitBin{args.dataset_id}/"
-    dataset_name = f"FruitBin_{choice}_{Nb_instance}_{occ_target_min}_{occ_target_max}"
-    #dataset_name = f"/gpfsscratch/rech/uli/ubn15wo/dataset_new{args.dataset_id}/s2rg/Fruits_all_medium/GUIMOD_{choice}"
-    list_categories = ["banana1", "kiwi1", "pear2", "apricot", "orange2", "peach1", "lemon2", "apple2"]
-    Nb_camera = 15
-    #Nb_world = 10000
+    dataset_path = f"/gpfsscratch/rech/uli/ubn15wo/FruitBin{args.dataset_id}" #TODO,  path and name of the destination for the precessed dataset.
+    dataset_name = f"FruitBin_{choice}_{Nb_instance}_{occ_target_min}_{occ_target_max}" #TODO, name of the subdataset preprocessed for scenarios. 
+
+    list_categories = ["banana1", "kiwi1", "pear2", "apricot", "orange2", "peach1", "lemon2", "apple2"] #TODO, to change if different objects
+    Nb_camera = 15  # TODO, to change if different number of cameras. 
 
-    generate_folders(dataset_path , dataset_name, list_categories, scenario)
+    generate_folders(dataset_path , dataset_name, list_categories)
 
     if choice == 'high':
         camera = np.matrix([[1386.4138492513919, 0.0, 960.5],
@@ -117,15 +110,14 @@ if __name__ == '__main__':
                         [0.0, (2 / 3), 0.0],
                         [0.0, 0.0, 1.0]])
 
-    new_size = (640, 480)
+    new_size = (640, 480)  # size used for training baseline of 6D pose estimation
 
     new_camera = trans @ camera
 
-    #np.savetxt(f'{dataset_name}/Generated/camera_{choice}.txt', camera)
     print("rearrange", args.rearrange)
     print("compute", args.compute)
 
-    if args.rearrange == 'yes': 
+    if args.rearrange == 'yes': # step nedeed before process, to do only one time
         reform_data(dataset_src, dataset_path, dataset_type, Nb_camera, args.World_begin, args.Nb_worlds)
 
     objs = {"banana1": [ 0.02949700132012367249, 0.1511049866676330566, 0.06059300713241100311 ],
@@ -156,6 +148,6 @@ if __name__ == '__main__':
         bbox = get_3D_bbox(ext)
         np.savetxt(f'{dataset_path}/{dataset_name}/Generated/{categories}/{categories}_bbox_3d.txt', bbox)  # save
 
-    if args.compute == 'yes' : 
+    if args.compute == 'yes' : # process of a sub dataset for specific scenarios, it can be repeated to generate multiple ready to train sub-datasets 
         process_compute(dataset_path, dataset_path+'/'+dataset_name, camera, new_camera, new_size, Nb_camera, args.World_begin, args.Nb_worlds, list_categories, occ_target_min, occ_target_max, False)