simba-uw-tf-dev 5.2.7__tar.gz → 5.2.8__tar.gz

{simba_uw_tf_dev-5.2.7 → simba_uw_tf_dev-5.2.8}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: simba_uw_tf_dev
-Version: 5.2.7
+Version: 5.2.8
 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-5.2.7 → simba_uw_tf_dev-5.2.8}/requirements_gpu.txt

@@ -4,4 +4,5 @@ shap==0.46.1.dev78
 cuml-cu12==24.12.0
 torch==2.5.0
 ultralytics==8.4.9
+pynvvideocodec==2.1.0
 nvidia-cuda-runtime-cu12

{simba_uw_tf_dev-5.2.7 → simba_uw_tf_dev-5.2.8}/setup.py

@@ -29,7 +29,7 @@ with open(GPU_REQUIREMENTS_PATH, "r") as f:
 # Setup configuration
 setuptools.setup(
     name="simba_uw_tf_dev",
-    version="5.2.7",
+    version="5.2.8",
     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-5.2.7 → simba_uw_tf_dev-5.2.8}/simba/data_processors/cuda/utils.py

@@ -315,7 +315,8 @@ def get_nvc_decoder(video_path: Union[str, os.PathLike],
                     gpu_id: int = 0,
                     use_device_memory: bool = False):
 
-    from simba.utils.checks import (check_file_exist_and_readable, check_instance, check_int)
+    from simba.utils.checks import (check_file_exist_and_readable,
+                                    check_instance, check_int)
     from simba.utils.errors import SimBAGPUError
 
     if nvc is None:

{simba_uw_tf_dev-5.2.7 → simba_uw_tf_dev-5.2.8}/simba/data_processors/movement_calculator.py

@@ -18,9 +18,9 @@ from simba.utils.checks import (
     check_that_column_exist, check_valid_boolean, check_valid_dict,
     check_valid_lst, check_valid_tuple)
 from simba.utils.errors import FrameRangeError, InvalidInputError, NoDataError
+from simba.utils.warnings import MissingFileWarning
 from simba.utils.printing import SimbaTimer, stdout_information, stdout_success
-from simba.utils.read_write import (find_files_of_filetypes_in_directory,
-                                    get_fn_ext, read_df, seconds_to_timestamp)
+from simba.utils.read_write import (find_files_of_filetypes_in_directory, get_fn_ext, read_df, seconds_to_timestamp)
 
 START, END = 'START', 'END'
 
@@ -151,9 +151,6 @@ class MovementCalculator(ConfigReader, FeatureExtractionMixin):
                     raise FrameRangeError(msg=f'Cannot compute movement between frame {start_frm} (time s: {start_time}) and frame {end_frm} (time s: {end_time}) in video {video_name}. The data only has {len(self.data_df)} frames.', source=self.__class__.__name__)
                 self.data_df = self.data_df.loc[start_frm:end_frm].reset_index()
                 if self.verbose: stdout_information(msg=f"Slicing video {video_name} between frames {start_frm} and {end_frm}...)")
-            else:
-                start_frm, end_frm = int(0 * self.fps), int(899 * self.fps)
-                self.data_df = self.data_df.loc[start_frm:end_frm].reset_index()
             if self.bp_list:
                 check_that_column_exist(df=self.data_df, column_name=self.bp_list, file_name=file_path)
                 self.data_df = self.data_df[self.bp_list]
@@ -212,7 +209,7 @@ if __name__ == "__main__" and not hasattr(sys, 'ps1'):
     runner.save()
 
 
-# TIME_TIME_PATH = r"F:\troubleshooting\sam\sam\project_folder\DATA MASTERSHEET (1).csv"
+# TIME_TIME_PATH = r"F:\troubleshooting\sam\sam\project_folder\DATA MASTERSHEET_0428_PRE_ISO.csv"
 # time_stamps = pd.read_csv(TIME_TIME_PATH)[['VIDEO_FILE_NAME', 'START', 'END']].set_index('VIDEO_FILE_NAME').to_dict(orient='index')
 # test = MovementCalculator(config_path=r"F:\troubleshooting\sam\sam\project_folder\project_config.ini",
 #                           body_parts=['center'],  #['Simon CENTER OF GRAVITY', 'JJ CENTER OF GRAVITY', 'Animal_1 CENTER OF GRAVITY']

{simba_uw_tf_dev-5.2.7 → simba_uw_tf_dev-5.2.8}/simba/labelling/labelling_interface.py

@@ -317,9 +317,8 @@ class LabellingInterface(ConfigReader):
         start_frm, end_frm = int(start_frm), int(end_frm)
         if start_frm >= end_frm:
             raise FrameRangeError(msg=f'The RANGE START FRAME ({start_frm}) has to be smaller than the RANGE END FRAME ({end_frm})', source=self.__class__.__name__)
-        for frm_no in range(int(start_frm), int(end_frm) + 1):
-            for target in self.clf_names:
-                self.data_df_targets[target].loc[frm_no] = self.clf_cbs[target].get()
+        for target in self.clf_names:
+            self.data_df_targets[target].loc[start_frm:end_frm] = int(self.clf_cbs[target].get())
         self.set_img(img_lbl=self.img_lbl, video_path=self.video_path, img_idx=end_frm)
         self.current_frm_eb.entry_set(end_frm)
         self.img_idx = deepcopy(end_frm)

{simba_uw_tf_dev-5.2.7 → simba_uw_tf_dev-5.2.8}/simba/labelling/standard_labeller.py

@@ -28,7 +28,8 @@ from simba.utils.checks import (check_file_exist_and_readable, check_int,
                                 check_valid_dataframe, check_valid_dict,
                                 check_video_and_data_frm_count_align)
 from simba.utils.enums import Formats, TagNames
-from simba.utils.errors import FrameRangeError, NoDataError, NoFilesFoundError
+from simba.utils.errors import (FrameRangeError, NoDataError,
+                                NoFilesFoundError, SimbaError)
 from simba.utils.lookups import (get_current_time, get_display_resolution,
                                  get_img_resize_info,
                                  get_labelling_img_kbd_bindings,
@@ -166,7 +167,7 @@ class LabellingInterface(ConfigReader):
         self.forward_btn = SimbaButton(parent=self.navigation_frm, txt="", img='play_large_green', tooltip_txt="+1 FRAME", cmd=self.__advance_frame, cmd_kwargs={'frm_id': lambda: self.img_idx + 1})
         self.last_frm_btn = SimbaButton(parent=self.navigation_frm, txt="", img='skip_forward_large', tooltip_txt="GO TO LAST FRAME", cmd=self.__advance_frame, cmd_kwargs={'frm_id': self.max_frm_id})
 
-        self.set_img(img_lbl=self.img_lbl, video_path=self.video_path, img_idx=self.img_idx, display_size=self.img_display_size)
+        self.set_img(img_lbl=self.img_lbl, video_path=self.cap, img_idx=self.img_idx, display_size=self.img_display_size)
 
         self.navigation_frm.grid(row=2, column=0, sticky=NSEW, pady=10, padx=10, columnspan=1)
         self.current_frm_eb.grid(row=0, column=0, sticky=EW, columnspan=4)
@@ -350,7 +351,7 @@ class LabellingInterface(ConfigReader):
                 else:
                     self.clf_cbs[clf].set(bool(self.data_df_targets[clf].loc[frm_id]))
             self.img_idx = deepcopy(frm_id)
-            self.set_img(img_lbl=self.img_lbl, video_path=self.video_path, img_idx=frm_id, display_size=self.img_display_size)
+            self.set_img(img_lbl=self.img_lbl, video_path=self.cap, img_idx=frm_id, display_size=self.img_display_size)
 
     def __save_behavior_in_range(self,
                                  start_frm: int,
@@ -361,13 +362,12 @@ class LabellingInterface(ConfigReader):
         start_frm, end_frm = int(start_frm), int(end_frm)
         if start_frm >= end_frm:
             raise FrameRangeError(msg=f'The RANGE START FRAME ({start_frm}) has to be smaller than the RANGE END FRAME ({end_frm})', source=self.__class__.__name__)
-        for frm_no in range(int(start_frm), int(end_frm) + 1):
-            for target in self.clf_names:
-                self.data_df_targets[target].loc[frm_no] = int(self.clf_cbs[target].get())
-        self.set_img(img_lbl=self.img_lbl, video_path=self.video_path, img_idx=end_frm, display_size=self.img_display_size)
+        for target in self.clf_names:
+            self.data_df_targets[target].loc[start_frm:end_frm] = int(self.clf_cbs[target].get())
         self.current_frm_eb.entry_set(end_frm)
         self.img_idx = deepcopy(end_frm)
         self.__create_print_statements(frame=(start_frm, end_frm))
+        self.main_window.after(1, lambda: self.set_img(img_lbl=self.img_lbl, video_path=self.cap, img_idx=end_frm, display_size=self.img_display_size))
 
     def save_behavior_in_frm(self,
                              frame_idx: int,
@@ -380,8 +380,7 @@ class LabellingInterface(ConfigReader):
         try:
             write_df(self.save_df, self.file_type, self.targets_inserted_file_path)
         except Exception as e:
-            print(e, f"SIMBA ERROR: File for video {self.video_name} could not be saved.")
-            raise FileExistsError
+            raise SimbaError(msg=f"Annotation file for video {self.video_name} could not be saved at {self.targets_inserted_file_path}. {e}", source=self.__class__.__name__)
         stdout_success(msg=f"SAVED: Annotation file for video {self.video_name} saved within the {self.targets_folder} directory at file path: {self.targets_inserted_file_path}.", source=self.__class__.__name__)
         if not self.config.has_section("Last saved frames"):
             self.config.add_section("Last saved frames")
@@ -455,3 +454,7 @@ class LabellingInterface(ConfigReader):
 #                               threshold_dict={'Attack': 0.4},
 #                               setting='from_scratch',
 #                               continuing=False)
+# _ = LabellingInterface(config_path=r"F:\troubleshooting\sophiaa\project_folder\project_config.ini",
+#                        file_path=r"F:\troubleshooting\sophiaa\project_folder\videos\Choice222.mp4",
+#                        thresholds=None,
+#                        continuing=False)

{simba_uw_tf_dev-5.2.7 → simba_uw_tf_dev-5.2.8}/simba/model/yolo_fit.py

@@ -200,6 +200,19 @@ if __name__ == "__main__" and not hasattr(sys, 'ps1'):
 # fitter.run()
 #
 
+#
+# fitter = FitYolo(weights_path=r"F:\yolo_weights\yolo26n.pt",
+#                  model_yaml=r"F:\irondog\data\yolo\map.yaml",
+#                  save_path=r"F:\irondog\data\yolo\mdl",
+#                  epochs=1500,
+#                  batch=16,
+#                  workers=3,
+#                  format=None,
+#                  device=0,
+#                  imgsz=256)
+# fitter.run()
+#
+
 
 
 # fitter = FitYolo(weights_path=r"C:\Users\sroni\Downloads\yolo26n.pt",

{simba_uw_tf_dev-5.2.7 → simba_uw_tf_dev-5.2.8}/simba/third_party_label_appenders/transform/sam3_to_yolo_bbox.py

@@ -13,7 +13,7 @@ To merge this project with others that share the same class names and task type,
 import os
 import random
 import time
-from typing import Optional, Tuple, Union
+from typing import List, Optional, Tuple, Union
 
 import cv2
 
@@ -32,14 +32,15 @@ import numpy as np
 
 from simba.third_party_label_appenders.converters import create_yolo_yaml
 from simba.utils.checks import (check_file_exist_and_readable, check_float,
-                                check_if_dir_exists, check_int,
+                                check_if_dir_exists, check_instance, check_int,
                                 check_nvidea_gpu_available, check_str,
-                                check_valid_boolean, check_valid_tuple)
+                                check_valid_boolean, check_valid_lst,
+                                check_valid_tuple)
 from simba.utils.errors import NoFilesFoundError, SimBAPAckageVersionError
 from simba.utils.printing import (SimbaTimer, stdout_information,
                                   stdout_success, stdout_warning)
 from simba.utils.read_write import (create_directory,
-                                    find_all_videos_in_directory,
+                                    find_all_videos_in_directory, get_fn_ext,
                                     get_pkg_version, get_video_meta_data,
                                     read_frm_of_video, recursive_file_search)
 from simba.utils.yolo import create_yolo_sample_visualizations
@@ -57,7 +58,7 @@ class SAM3ToYoloBBox:
        * :class:`~simba.third_party_label_appenders.transform.merge_yolo_projects.MergeYoloProjects` — merge multiple YOLO ``map.yaml`` projects into one training set.
        * :class:`~simba.third_party_label_appenders.transform.sam3_to_yolo_seg.SAM3ToYoloSeg` — SAM3 to YOLO **segmentation** labels from the same predictor stack.
 
-    :param Union[str, os.PathLike] video_dir: Directory containing input videos.
+    :param Union[str, os.PathLike, List[Union[str, os.PathLike]]] video_data: Input videos. Accepts: (1) a directory containing video files (combine with ``recursive=True`` to also search subdirectories), (2) a path to a single video file, or (3) a list of video file paths.
     :param Union[str, os.PathLike] sam_path: Path to SAM3 model weights (e.g. sam3.pt).
     :param Union[str, os.PathLike] save_dir: Root output directory for the YOLO project.
     :param str txt_prompt: Text prompt for SAM3 (e.g. "mouse", "mouse tail").
@@ -70,7 +71,7 @@ class SAM3ToYoloBBox:
     :param Optional[Union[Tuple[int, int, int], bool]] clahe: If True, applies CLAHE with default params. If tuple of (clip_limit, tile_x, tile_y), applies CLAHE with those params. Default False.
     :param float buffer_pct: Fraction to expand each box by (e.g. 0.1 adds 10% of width/height on each side). Default 0.0.
     :param int consecutive_miss_limit: If this many consecutive frames yield no detection, skip to the next video. Default 100.
-    :param bool recursive: If True, search video_dir and all subdirectories for videos. Default False.
+    :param bool recursive: If True and ``video_data`` is a directory, search it and all subdirectories for videos. Ignored if ``video_data`` is a file path or a list. Default False.
     :param Optional[int] seed: Random seed for reproducible frame sampling.
     :param Optional[int] max_detections: Maximum number of detections to keep per frame (sorted by confidence descending). If None, all detections above ``conf`` are kept. Default None.
     :param bool visualize: If True, saves annotated images with bounding-box overlays to a ``visualizations`` subfolder inside ``save_dir``. Useful for verifying SAM3 annotation quality. Default False.
@@ -83,12 +84,12 @@ class SAM3ToYoloBBox:
     :raises SimBAPAckageVersionError: If ``ultralytics`` is not installed, or ``SAM3SemanticPredictor`` cannot be imported.
 
     :example:
-    >>> runner = SAM3ToYoloBBox(video_dir=r'/path/to/videos', sam_path=r'/path/to/sam3.pt', save_dir=r'/path/to/yolo_project', txt_prompt='mouse', n_frames=50)
+    >>> runner = SAM3ToYoloBBox(video_data=r'/path/to/videos', sam_path=r'/path/to/sam3.pt', save_dir=r'/path/to/yolo_project', txt_prompt='mouse', n_frames=50)
     >>> runner.run()
     """
 
     def __init__(self,
-                 video_dir: Union[str, os.PathLike],
+                 video_data: Union[str, os.PathLike, List[Union[str, os.PathLike]]],
                  sam_path: Union[str, os.PathLike],
                  save_dir: Union[str, os.PathLike],
                  txt_prompt: str = 'mouse',
@@ -115,7 +116,15 @@ class SAM3ToYoloBBox:
         if SAM3SemanticPredictor is None:
             raise SimBAPAckageVersionError(msg='Could not import SAM3SemanticPredictor from ultralytics.models.sam. Install a compatible ultralytics build with SAM3 support.', source=self.__class__.__name__)
 
-        check_if_dir_exists(in_dir=video_dir, source=f'{self.__class__.__name__} video_dir')
+        check_instance(source=f'{self.__class__.__name__} video_data', instance=video_data, accepted_types=(str, os.PathLike, list))
+        if isinstance(video_data, list):
+            check_valid_lst(data=video_data, source=f'{self.__class__.__name__} video_data', valid_dtypes=(str, os.PathLike), min_len=1)
+            for v in video_data:
+                check_file_exist_and_readable(file_path=v)
+        elif os.path.isfile(video_data):
+            check_file_exist_and_readable(file_path=video_data)
+        else:
+            check_if_dir_exists(in_dir=video_data, source=f'{self.__class__.__name__} video_data')
         check_file_exist_and_readable(file_path=sam_path)
         check_if_dir_exists(in_dir=save_dir, source=f'{self.__class__.__name__} save_dir')
         check_str(name=f'{self.__class__.__name__} txt_prompt', value=txt_prompt)
@@ -135,14 +144,18 @@ class SAM3ToYoloBBox:
         check_float(name=f'{self.__class__.__name__} io_timeout', value=io_timeout, min_value=0.0)
         if max_detections is not None: check_int(name=f'{self.__class__.__name__} max_detections', value=max_detections, min_value=1)
         if seed is not None: check_int(name=f'{self.__class__.__name__} seed', value=seed)
-        self.video_dir, self.sam_path, self.save_dir, self.txt_prompt = video_dir, sam_path, save_dir, txt_prompt
+        self.video_data, self.sam_path, self.save_dir, self.txt_prompt = video_data, sam_path, save_dir, txt_prompt
         self.n_frames, self.names, self.train_val_split, self.conf, self.imgsz = n_frames, names, train_val_split, conf, sam_imgsz
         self.greyscale, self.clahe, self.buffer_pct, self.consecutive_miss_limit, self.max_detections, self.seed, self.verbose, self.visualize, self.min_frame_gap, self.io_timeout = greyscale, clahe, buffer_pct, consecutive_miss_limit, max_detections, seed, verbose, visualize, min_frame_gap, io_timeout
         self.name_map = {name: idx for idx, name in enumerate(names)}
-        if recursive:
-            self.video_paths = recursive_file_search(directory=video_dir, extensions=[".avi", ".mp4", ".mov", ".flv", ".m4v", ".webm", ".h264"], as_dict=True, raise_error=True)
+        if isinstance(video_data, list):
+            self.video_paths = {get_fn_ext(filepath=v)[1]: str(v) for v in video_data}
+        elif isinstance(video_data, (str, os.PathLike)) and os.path.isfile(video_data):
+            self.video_paths = {get_fn_ext(filepath=video_data)[1]: str(video_data)}
+        elif recursive:
+            self.video_paths = recursive_file_search(directory=video_data, extensions=[".avi", ".mp4", ".mov", ".flv", ".m4v", ".webm", ".h264"], as_dict=True, raise_error=True)
         else:
-            self.video_paths = find_all_videos_in_directory(directory=video_dir, as_dict=True, raise_error=True)
+            self.video_paths = find_all_videos_in_directory(directory=video_data, as_dict=True, raise_error=True)
         if shuffle_videos:
             items = list(self.video_paths.items())
             random.shuffle(items)
@@ -313,10 +326,10 @@ class SAM3ToYoloBBox:
         return '\n'.join(lines) + '\n' if lines else ''
 
 
-# runner = SAM3ToYoloBBox(video_dir=r'F:\netholabs\V6\batch_5\mp4s',
+# runner = SAM3ToYoloBBox(video_data=r'F:\irondog\data\uncompressed',
 #                         sam_path=r'D:\sam3\sam3.pt',
-#                         save_dir=r'F:\netholabs\V6\cage_3\yolo_project_0413',
-#                         txt_prompt='mouse',
+#                         save_dir=r'F:\irondog\data\yolo',
+#                         txt_prompt='dog',
 #                         n_frames=25,
 #                         verbose=True,
 #                         conf=0.25,
@@ -327,7 +340,7 @@ class SAM3ToYoloBBox:
 #                         shuffle_videos=True,
 #                         visualize=True)
 # runner.run()
-#
+
 
 
 # runner = SAM3ToYoloBBox(video_dir=r'F:\netholabs\V6\cage_3\samples',

{simba_uw_tf_dev-5.2.7 → simba_uw_tf_dev-5.2.8}/simba_uw_tf_dev.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: simba-uw-tf-dev
-Version: 5.2.7
+Version: 5.2.8
 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-5.2.7 → simba_uw_tf_dev-5.2.8}/simba_uw_tf_dev.egg-info/requires.txt

@@ -150,4 +150,5 @@ shap==0.46.1.dev78
 cuml-cu12==24.12.0
 torch==2.5.0
 ultralytics==8.4.9
+pynvvideocodec==2.1.0
 nvidia-cuda-runtime-cu12