simba-uw-tf-dev 3.2.2__tar.gz → 3.2.4__tar.gz

{simba_uw_tf_dev-3.2.2 → simba_uw_tf_dev-3.2.4}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: simba_uw_tf_dev
-Version: 3.2.2
+Version: 3.2.4
 Summary: Toolkit for computer classification and analysis of behaviors in experimental animals
 Home-page: https://github.com/sgoldenlab/simba
 Author: Simon Nilsson, Jia Jie Choong, Sophia Hwang

{simba_uw_tf_dev-3.2.2 → simba_uw_tf_dev-3.2.4}/Simba_UW_tf_dev.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: simba-uw-tf-dev
-Version: 3.2.2
+Version: 3.2.4
 Summary: Toolkit for computer classification and analysis of behaviors in experimental animals
 Home-page: https://github.com/sgoldenlab/simba
 Author: Simon Nilsson, Jia Jie Choong, Sophia Hwang

{simba_uw_tf_dev-3.2.2 → simba_uw_tf_dev-3.2.4}/setup.py

@@ -29,7 +29,7 @@ with open(GPU_REQUIREMENTS_PATH, "r") as f:
 # Setup configuration
 setuptools.setup(
     name="simba_uw_tf_dev",
-    version="3.2.2",
+    version="3.2.4",
     author="Simon Nilsson, Jia Jie Choong, Sophia Hwang",
     author_email="sronilsson@gmail.com",
     description="Toolkit for computer classification and analysis of behaviors in experimental animals",

{simba_uw_tf_dev-3.2.2 → simba_uw_tf_dev-3.2.4}/simba/bounding_box_tools/yolo/yolo_pose_inference.py

@@ -18,10 +18,6 @@ COORD_COLS = ['X1', 'Y1', 'X2', 'Y2', 'X3', 'Y3', 'X4', 'Y4']
 NEAREST = 'nearest'
 
 
-
-
-
-
 class YOLOPoseInference():
 
     """
@@ -168,21 +164,21 @@ if __name__ == "__main__" and not hasattr(sys, 'ps1'):
     inference.run()
 
 
-video_path = r"/mnt/c/troubleshooting/mitra/project_folder/videos/501_MA142_Gi_CNO_0521.mp4"
-video_path = "/mnt/d/netholabs/yolo_videos/2025-05-28_19-46-56.mp4"
-video_path = "/mnt/d/netholabs/yolo_videos/2025-05-28_19-50-23.mp4"
-
-save_dir = r'/mnt/d/netholabs/yolo_videos_out'
-i = YOLOPoseInference(weights_path=r"/mnt/d/netholabs/yolo_mdls_1/train/weights/best.pt",
-                        video_path=video_path,
-                        save_dir=save_dir,
-                        verbose=True,
-                        device=0,
-                        format='onnx',
-                        stream=True,
-                        keypoint_names=('nose', 'ear_left', 'ear_right', 'lateral_left', 'center', 'lateral_right', 'tail_base'),
-                        batch_size=32,
-                        imgsz=320,
-                        interpolate=True,
-                         threshold=0.8)
-#i.run()
+# video_path = r"/mnt/c/troubleshooting/mitra/project_folder/videos/501_MA142_Gi_CNO_0521.mp4"
+# video_path = "/mnt/d/netholabs/yolo_videos/2025-05-28_19-46-56.mp4"
+# video_path = "/mnt/d/netholabs/yolo_videos/2025-05-28_19-50-23.mp4"
+#
+# save_dir = r'/mnt/d/netholabs/yolo_videos_out'
+# i = YOLOPoseInference(weights_path=r"/mnt/d/netholabs/yolo_mdls_1/train/weights/best.pt",
+#                         video_path=video_path,
+#                         save_dir=save_dir,
+#                         verbose=True,
+#                         device=0,
+#                         format='onnx',
+#                         stream=True,
+#                         keypoint_names=('nose', 'ear_left', 'ear_right', 'lateral_left', 'center', 'lateral_right', 'tail_base'),
+#                         batch_size=32,
+#                         imgsz=320,
+#                         interpolate=True,
+#                         threshold=0.8)
+# #i.run()

{simba_uw_tf_dev-3.2.2 → simba_uw_tf_dev-3.2.4}/simba/bounding_box_tools/yolo/yolo_pose_visualizer.py

@@ -70,7 +70,7 @@ class YOLOPoseVisualizer():
     and optional filtering. The result is saved as a new annotated video, and supports multicore parallel rendering
     for efficient processing of long videos.
 
-    .. seelalso::
+    .. seealso::
        To create YOLO pose data, see `:func:~simba.bounding_box_tools.yolo.yolo_pose_inference.YOLOPoseInference`
 
     :param Union[str, os.PathLike] data_path: Path to the CSV file containing keypoint data (output from YOLO pose inference).

{simba_uw_tf_dev-3.2.2 → simba_uw_tf_dev-3.2.4}/simba/bounding_box_tools/yolo/yolo_visualize.py

@@ -34,7 +34,7 @@ class YOLOVisualizer():
         """
         Visualize the results of YOLO model.
 
-        .. seelalso::
+        .. seealso::
            :func:`simba.bounding_box_tools.yolo.model.inference_yolo`
 
         .. video:: _static/img/YOLOVisualizer.webm

{simba_uw_tf_dev-3.2.2 → simba_uw_tf_dev-3.2.4}/simba/data_processors/cuda/image.py

@@ -49,7 +49,7 @@ from simba.utils.printing import SimbaTimer, stdout_success
 from simba.utils.read_write import (
     check_if_hhmmss_timestamp_is_valid_part_of_video, get_fn_ext,
     get_memory_usage_array, get_video_meta_data, read_df,
-    read_img_batch_from_video_gpu, create_directory)
+    read_img_batch_from_video_gpu, create_directory, concatenate_videos_in_folder, read_img_batch_from_video)
 
 warnings.simplefilter('ignore', category=NumbaPerformanceWarning)
 
@@ -830,25 +830,24 @@ def slice_imgs(video_path: Union[str, os.PathLike],
     """
     THREADS_PER_BLOCK = (32, 32, 1)
 
-    video_meta_data = get_video_meta_data(video_path=video_path)
+    video_meta_data = get_video_meta_data(video_path=video_path, fps_as_int=False)
     video_meta_data['frame_count'] = shapes.shape[0]
     n, w, h = video_meta_data['frame_count'], video_meta_data['width'], video_meta_data['height']
     is_color = ImageMixin.is_video_color(video=video_path)
-    timer, save_temp_dir, results = SimbaTimer(start=True), None, None
+    timer, save_temp_dir, results, video_out_path = SimbaTimer(start=True), None, None, None
     if save_dir is None:
         results = np.zeros((n, h, w, 3), dtype=np.uint8) if is_color else np.zeros((n, h, w), dtype=np.uint8)
     else:
         save_temp_dir = os.path.join(save_dir, f'temp_{video_meta_data["video_name"]}')
         create_directory(path=save_temp_dir, overwrite=True)
+        video_out_path = os.path.join(save_dir, f'{video_meta_data["video_name"]}.mp4')
     for batch_cnt, start_img_idx in enumerate(range(0, n, batch_size)):
         end_img_idx = start_img_idx + batch_size
         end_img_idx = video_meta_data['frame_count'] if end_img_idx > video_meta_data['frame_count'] else end_img_idx
         batch_n = end_img_idx - start_img_idx
         if verbose: print(f'Processing images {start_img_idx} - {end_img_idx} (of {n})...')
         if save_dir is not None:
-            batch_save_path = os.path.join(save_temp_dir, f'{video_meta_data["video_name"]}.mp4')
-            results = np.zeros((batch_n, h, w, 3), dtype=np.uint8) if is_color else np.zeros((batch_n, h, w), dtype=np.uint8)
-
+            batch_save_path = os.path.join(save_temp_dir, f'{batch_cnt}.mp4')
         batch_imgs = read_img_batch_from_video_gpu(video_path=video_path, start_frm=start_img_idx, end_frm=end_img_idx-1)
         batch_imgs = np.stack(list(batch_imgs.values()), axis=0)
         batch_shapes = shapes[start_img_idx:end_img_idx].astype(np.int32)
@@ -865,9 +864,10 @@ def slice_imgs(video_path: Union[str, os.PathLike],
         if save_dir is None:
             results[start_img_idx: end_img_idx] = batch_img_dev.copy_to_host()
         else:
-            results[0: batch_n] = batch_img_dev.copy_to_host()
-            results = {k: v for k, v in enumerate(results)}
-            ImageMixin().img_stack_to_video(imgs=results, fps=video_meta_data['fps'], save_path=batch_save_path, verbose=verbose)
+            results = {k: v for k, v in enumerate(batch_img_dev.copy_to_host())}
+            ImageMixin().img_stack_to_video(imgs=results, fps=video_meta_data['fps'], save_path=batch_save_path, verbose=False)
+    if save_dir:
+        concatenate_videos_in_folder(in_folder=save_temp_dir, save_path=video_out_path, remove_splits=True, gpu=True)
     timer.stop_timer()
     if verbose:
         stdout_success(msg='Shapes sliced in video.', elapsed_time=timer.elapsed_time_str)
@@ -1333,6 +1333,7 @@ video_path = "/mnt/c/troubleshooting/RAT_NOR/project_folder/videos/03152021_NOB_
 data_path = "/mnt/c/troubleshooting/RAT_NOR/project_folder/csv/outlier_corrected_movement_location/03152021_NOB_IOT_8.csv"
 save_dir = '/mnt/d/netholabs/yolo_videos/input/mp4_20250606083508'
 
+get_video_meta_data(video_path)
 
 nose_arr = read_df(file_path=data_path, file_type='csv', usecols=['Nose_x', 'Nose_y', 'Tail_base_x', 'Tail_base_y', 'Lat_left_x', 'Lat_left_y', 'Lat_right_x', 'Lat_right_y']).values.reshape(-1, 4, 2).astype(np.int32) ## READ THE BODY-PART THAT DEFINES THE HULL AND CONVERT TO ARRAY
 
@@ -1341,7 +1342,7 @@ polygons = GeometryMixin().multiframe_minimum_rotated_rectangle(shapes=polygons)
 polygon_lst = [] # GET THE POINTS OF THE RECTANGLES
 for i in polygons: polygon_lst.append(np.array(i.exterior.coords))
 polygons = np.stack(polygon_lst, axis=0)
-sliced_imgs = slice_imgs(video_path=video_path, shapes=polygons, batch_size=500, save_dir=save_dir) #SLICE THE RECTANGLES IN THE VIDEO.
+sliced_imgs = slice_imgs(video_path=video_path, shapes=polygons, batch_size=1000, save_dir=save_dir) #SLICE THE RECTANGLES IN THE VIDEO.
 
 #sliced_imgs =  {k: v for k, v in enumerate(sliced_imgs)}
 

{simba_uw_tf_dev-3.2.2 → simba_uw_tf_dev-3.2.4}/simba/feature_extractors/feature_subsets.py

@@ -16,6 +16,7 @@ from itertools import combinations
 from simba.feature_extractors.perimeter_jit import jitted_hull
 from simba.mixins.config_reader import ConfigReader
 from simba.mixins.feature_extraction_mixin import FeatureExtractionMixin
+from simba.mixins.feature_extraction_supplement_mixin import FeatureExtractionSupplemental
 from simba.mixins.train_model_mixin import TrainModelMixin
 from simba.roi_tools.roi_utils import get_roi_dict_from_dfs
 from simba.utils.checks import (
@@ -33,7 +34,6 @@ from simba.utils.read_write import (copy_files_in_directory,
                                     remove_multiple_folders, write_df)
 
 SHAPE_TYPE = "Shape_type"
-
 TWO_POINT_BP_DISTANCES = 'TWO-POINT BODY-PART DISTANCES (MM)'
 WITHIN_ANIMAL_THREE_POINT_ANGLES = 'WITHIN-ANIMAL THREE-POINT BODY-PART ANGLES (DEGREES)'
 WITHIN_ANIMAL_THREE_POINT_HULL = "WITHIN-ANIMAL THREE-POINT CONVEX HULL PERIMETERS (MM)"
@@ -43,6 +43,10 @@ ANIMAL_CONVEX_HULL_AREA = "ENTIRE ANIMAL CONVEX HULL AREA (MM2)"
 FRAME_BP_MOVEMENT = "FRAME-BY-FRAME BODY-PART MOVEMENTS (MM)"
 FRAME_BP_TO_ROI_CENTER = "FRAME-BY-FRAME BODY-PART DISTANCES TO ROI CENTERS (MM)"
 FRAME_BP_INSIDE_ROI = "FRAME-BY-FRAME BODY-PARTS INSIDE ROIS (BOOLEAN)"
+ARENA_EDGE = "BODY-PART DISTANCES TO VIDEO FRAME EDGE (MM)"
+
+
+
 
 FEATURE_FAMILIES = [TWO_POINT_BP_DISTANCES,
                     WITHIN_ANIMAL_THREE_POINT_ANGLES,
@@ -52,7 +56,8 @@ FEATURE_FAMILIES = [TWO_POINT_BP_DISTANCES,
                     ANIMAL_CONVEX_HULL_AREA,
                     FRAME_BP_MOVEMENT,
                     FRAME_BP_TO_ROI_CENTER,
-                    FRAME_BP_INSIDE_ROI]
+                    FRAME_BP_INSIDE_ROI,
+                    ARENA_EDGE]
 
 
 class FeatureSubsetsCalculator(ConfigReader, TrainModelMixin):
@@ -202,6 +207,16 @@ class FeatureSubsetsCalculator(ConfigReader, TrainModelMixin):
                     distance = FeatureExtractionMixin.framewise_euclidean_distance_roi(location_1=bp_arr, location_2=center_point, px_per_mm=self.px_per_mm)
                     self.results[f"{animal} {bp} to {roi_name} center distance (mm)"] = distance
 
+    def __get_distances_to_frm_edge(self):
+        for animal, animal_bps in self.animal_bps.items():
+            for bp in animal_bps:
+                check_valid_dataframe(df=self.data_df, source=self.file_path, required_fields=[f"{bp}_x", f"{bp}_y"])
+                bp_arr = self.data_df[[f"{bp}_x", f"{bp}_y"]].values.astype(np.float32)
+                distance = FeatureExtractionSupplemental().border_distances(data=bp_arr, pixels_per_mm=self.px_per_mm, img_resolution=np.array([self.video_width, self.video_height], dtype=np.int32), time_window=1, fps=1)
+                self.results[f"{animal} {bp} to left video edge distance (mm)"] = distance[:, 0]
+                self.results[f"{animal} {bp} to right video edge distance (mm)"] = distance[:, 1]
+                self.results[f"{animal} {bp} to top video edge distance (mm)"] = distance[:, 2]
+                self.results[f"{animal} {bp} to bottom video edge distance (mm)"] = distance[:, 3]
 
     def __get_inside_roi(self):
         for animal, animal_bps in self.animal_bps.items():
@@ -279,7 +294,8 @@ class FeatureSubsetsCalculator(ConfigReader, TrainModelMixin):
             save_path = os.path.join(self.temp_dir, f'{self.video_name}.{self.file_type}')
             self.results = pd.DataFrame()
             print(f'Analyzing video {self.video_name}... ({file_cnt+1}/{len(self.data_paths)})')
-            _, self.px_per_mm, self.fps = self.read_video_info(video_name=self.video_name)
+            self.video_info, self.px_per_mm, self.fps = self.read_video_info(video_name=self.video_name)
+            self.video_width, self.video_height = self.video_info['Resolution_width'].values[0], self.video_info['Resolution_height'].values[0]
             self.data_df = read_df(file_path=file_path, file_type=self.file_type)
             for family_cnt, feature_family in enumerate(self.feature_families):
                 print(f"Analyzing {self.video_name} and {feature_family} (Video {file_cnt + 1}/{len(self.outlier_corrected_paths)}, Family {family_cnt + 1}/{len(self.feature_families)})...")
@@ -301,6 +317,8 @@ class FeatureSubsetsCalculator(ConfigReader, TrainModelMixin):
                     self.__get_roi_center_distances()
                 elif feature_family == FRAME_BP_INSIDE_ROI:
                     self.__get_inside_roi()
+                elif feature_family == ARENA_EDGE:
+                    self.__get_distances_to_frm_edge()
 
             self.results = self.results.add_suffix('_FEATURE_SUBSET')
             self.results = self.results[sorted(self.results.columns)]
@@ -320,6 +338,17 @@ class FeatureSubsetsCalculator(ConfigReader, TrainModelMixin):
         self.timer.stop_timer()
         stdout_success(msg="Feature sub-sets calculations complete!", elapsed_time=self.timer.elapsed_time_str)
 
+
+
+
+# test = FeatureSubsetsCalculator(config_path=r"C:\troubleshooting\mitra\project_folder\project_config.ini",
+#                                 feature_families=[ARENA_EDGE],
+#                                 append_to_features_extracted=False,
+#                                 file_checks=True,
+#                                 append_to_targets_inserted=False,
+#                                 save_dir=r"C:\troubleshooting\mitra\project_folder\csv\feature_subset")
+# test.run()
+
 #
 # test = FeatureSubsetsCalculator(config_path=r"D:\Stretch\Stretch\project_folder\project_config.ini",
 #                                 feature_families=[TWO_POINT_BP_DISTANCES],

{simba_uw_tf_dev-3.2.2 → simba_uw_tf_dev-3.2.4}/simba/plotting/cue_light_visualizer.py

@@ -114,6 +114,11 @@ class CueLightVisualizer(ConfigReader):
     .. notes:
        `Cue light tutorials <https://github.com/sgoldenlab/simba/blob/master/docs/cue_light_tutorial.md>`__.
 
+    .. video:: _static/img/CueLightVisualizer.webm
+       :width: 800
+       :autoplay:
+       :loop:
+
     :examples:
     >>> cue_light_visualizer = CueLightVisualizer(config_path='SimBAConfig', cue_light_names=['Cue_light'], video_path='VideoPath', video_setting=True, frame_setting=False)
     >>> cue_light_visualizer.run()

{simba_uw_tf_dev-3.2.2 → simba_uw_tf_dev-3.2.4}/simba/ui/pop_ups/subset_feature_extractor_pop_up.py

@@ -22,6 +22,7 @@ ANIMAL_CONVEX_HULL_AREA = "ENTIRE ANIMAL CONVEX HULL AREA (MM2)"
 FRAME_BP_MOVEMENT = "FRAME-BY-FRAME BODY-PART MOVEMENTS (MM)"
 FRAME_BP_TO_ROI_CENTER = "FRAME-BY-FRAME BODY-PART DISTANCES TO ROI CENTERS (MM)"
 FRAME_BP_INSIDE_ROI = "FRAME-BY-FRAME BODY-PARTS INSIDE ROIS (BOOLEAN)"
+ARENA_EDGE = "BODY-PART DISTANCES TO VIDEO FRAME EDGE (MM)"
 
 FEATURE_FAMILIES = [TWO_POINT_BP_DISTANCES,
                     WITHIN_ANIMAL_THREE_POINT_ANGLES,
@@ -31,7 +32,8 @@ FEATURE_FAMILIES = [TWO_POINT_BP_DISTANCES,
                     ANIMAL_CONVEX_HULL_AREA,
                     FRAME_BP_MOVEMENT,
                     FRAME_BP_TO_ROI_CENTER,
-                    FRAME_BP_INSIDE_ROI]
+                    FRAME_BP_INSIDE_ROI,
+                    ARENA_EDGE]
 
 class FeatureSubsetExtractorPopUp(PopUpMixin, ConfigReader):
     """

{simba_uw_tf_dev-3.2.2 → simba_uw_tf_dev-3.2.4}/simba/utils/read_write.py

@@ -2249,7 +2249,7 @@ def read_img_batch_from_video_gpu(video_path: Union[str, os.PathLike],
     """
 
     check_file_exist_and_readable(file_path=video_path)
-    video_meta_data = get_video_meta_data(video_path=video_path)
+    video_meta_data = get_video_meta_data(video_path=video_path, fps_as_int=False)
     if start_frm is not None:
         check_int(name=read_img_batch_from_video_gpu.__name__, value=start_frm, min_value=0, max_value=video_meta_data["frame_count"])
     else: