refactoring

experiment/hdbscan/main_kitti_processing_hdbscan_voxel.py

@@ -21,22 +21,30 @@
 
 def segment_pcds_by_hdbscan(start_index, end_index):
 
-    file_name = "experiment_bin_0704_4_sem_voxel_offset0_T0l03/start{}_end{}.pickle".format(
-        start_index, end_index
+    file_name = (
+        "experiment_bin_0704_4_sem_voxel_offset0_T0l03/start{}_end{}.pickle".format(
+            start_index, end_index
+        )
     )
 
-    with open(file_name, 'rb') as file:
+    with open(file_name, "rb") as file:
         data = pickle.load(file)
 
     pcd_hdbscan_voxel_down_points = np.asarray(data[2]["voxel_pcd_original_points"])
     pcd_hdbscan_voxel_down = o3d.geometry.PointCloud()
-    pcd_hdbscan_voxel_down.points = o3d.utility.Vector3dVector(pcd_hdbscan_voxel_down_points)
+    pcd_hdbscan_voxel_down.points = o3d.utility.Vector3dVector(
+        pcd_hdbscan_voxel_down_points
+    )
 
     clusterer = HDBSCAN()
     clusters = clusterer.fit_predict(np.asarray(pcd_hdbscan_voxel_down.points))
 
     return (
-        {"hdbscan_clustered_voxel_pcd_original_points": np.asarray(pcd_hdbscan_voxel_down.points)},
+        {
+            "hdbscan_clustered_voxel_pcd_original_points": np.asarray(
+                pcd_hdbscan_voxel_down.points
+            )
+        },
         {"hdbscan_voxel_trace_original": data[3]["voxel_trace_original"]},
         {"hdbscan_clusters": clusters},
         {"inst_label_array_for_clustering": data[6]["inst_label_array_for_clustering"]},
@@ -54,10 +62,8 @@ def process_kitti_hdbscan(from_num, to_num):
 
         result_tuple = segment_pcds_by_hdbscan(start_index, end_index)
 
-        file_name = (
-            "experiment_bin_0704_4_sem_voxel_offset0_T0l03_hdbscan/start{}_end{}.pickle".format(
-                start_index, end_index
-            )
+        file_name = "experiment_bin_0704_4_sem_voxel_offset0_T0l03_hdbscan/start{}_end{}.pickle".format(
+            start_index, end_index
         )
         new_file = open(file_name, "w")
         new_file.close()

experiment/hdbscan/main_kitti_processing_metrics_hdbscan.py

@@ -109,14 +109,14 @@ def main():
             ):  # в облаке нет инстансов => пропускаем
                 skipped += 1
                 print(
-                    "start_index={}, end_index={} skip".format(
-                        start_index, end_index
-                    )
+                    "start_index={}, end_index={} skip".format(start_index, end_index)
                 )
                 current_from_num = end_index
                 continue
 
-            clusters_list_without_noise = convert_clusters_to_list_of_point_arrays(clusters_array)[1:]
+            clusters_list_without_noise = convert_clusters_to_list_of_point_arrays(
+                clusters_array
+            )[1:]
 
             pred_inst_array = build_pred_inst_array(
                 copy.deepcopy(inst_label_array_for_clustering),

experiment/main_calc_metrics_by_csv.py

@@ -15,6 +15,17 @@
 import csv
 
 
+def calculate_sum_ones_zeros(values):
+    sum, ones, zeros = 0, 0, 0
+    for value in values:
+        sum += float(value)
+        if float(value) == 1.0:
+            ones += 1
+        elif float(value) == 0.0:
+            zeros += 1
+    return sum, ones, zeros
+
+
 def calculate_metrics(file_name):
     values_pres = []
     values_recall = []
@@ -26,35 +37,9 @@ def calculate_metrics(file_name):
             values_recall.append(row["recall"])
             values_fScore.append(row["fScore"])
 
-    sum_pres = 0
-    pres1 = 0
-    pres0 = 0
-    for value in values_pres:
-        sum_pres += float(value)
-        if float(value) == 1.0:
-            pres1 += 1
-        if float(value) == 0.0:
-            pres0 += 1
-
-    sum_recall = 0
-    recall1 = 0
-    recall0 = 0
-    for value in values_recall:
-        sum_recall += float(value)
-        if float(value) == 1.0:
-            recall1 += 1
-        if float(value) == 0.0:
-            recall0 += 1
-
-    sum_fScore = 0
-    fscore1 = 0
-    fscore0 = 0
-    for value in values_fScore:
-        sum_fScore += float(value)
-        if float(value) == 1.0:
-            fscore1 += 1
-        if float(value) == 0.0:
-            fscore0 += 1
+    sum_pres, pres1, pres0 = calculate_sum_ones_zeros(values_pres)
+    sum_recall, recall1, recall0 = calculate_sum_ones_zeros(values_recall)
+    sum_fScore, fscore1, fscore0 = calculate_sum_ones_zeros(values_fScore)
 
     print(
         "precision={}, 1={}, 0={}".format(