sportslabkit.utils#

Overview#

Classes#
`HiddenPrints`	-
`MovieIterator`	-
`ImageIterator`	-

Function#
`tqdm`(args, level, *kwargs)	Wrapper for tqdm.tqdm that uses the logger’s level.
`read_image`(img)	Reads an image from a file, URL, a numpy array, or a torch tensor.
`auto_string_parser`(value)	Auxiliary function to parse string values.
`count_iter_items`(iterable)	Consume an iterable not reading it into memory; return the number of items.
`load_config`(yaml_path)	Load config from yaml file.
`write_config`(yaml_path, cfg)	Write config to yaml file.
`pil2cv`(image)	Convert PIL image to OpenCV image.
`cv2pil`(image, convert_bgr2rgb)	Convert OpenCV image to PIL image.
`get_fps`(path)	-
`make_video`(frames, outpath, vcodec, pix_fmt, preset, crf, ss, t, c, height, width, input_framerate, logging, custom_ffmpeg)	Make video from a list of opencv format frames.
`merge_dict_of_lists`(d1, d2)	Merge two dicts of lists.
`get_git_root`()	Get the root of the git repository.
`download_file_from_google_drive`(id, destination)	-
`get_confirm_token`(response)	-
`save_response_content`(response, destination)	-
`increment_path`(path, exist_ok, mkdir)	Increments a path (appends a suffix) if it already exists.
`load_keypoints`(keypoint_json)	Loads source and target keypoints from a JSON file.
`sanitize_url_name`(url)	Sanitize the URL to create a safe filename for caching.
`fetch_or_cache_model`(url, dst, hash_prefix, progress)	Fetches a model from a URL or uses a cached version if it exists.

Attributes#
`PathLike`	-

Classes#

class sportslabkit.utils.HiddenPrints[source]#

class sportslabkit.utils.MovieIterator(path: str)[source]#

class sportslabkit.utils.ImageIterator(path: str)[source]#

Functions#

sportslabkit.utils.tqdm(*args, level: str = 'INFO', **kwargs) → collections.abc.Iterable#

Wrapper for tqdm.tqdm that uses the logger’s level.

Parameters:

*args – Arguments to pass to tqdm.tqdm
**kwargs – Keyword arguments to pass to tqdm.tqdm
level (str, optional) – Logging level to set. Defaults to “INFO”.

Returns:

Iterable from tqdm progress bar

Return type:

Iterable

sportslabkit.utils.read_image(img)[source]#

Reads an image from a file, URL, a numpy array, or a torch tensor. :param img: The image to read. :type img: str, Path, Image.Image, np.ndarray, or torch.Tensor

Returns:: The image as a numpy array.
Return type:: np.ndarray

sportslabkit.utils.auto_string_parser(value: str) → Any[source]#

Auxiliary function to parse string values.

Parameters:: value (str) – String value to parse.
Returns:: Parsed string value.
Return type:: value (any)

sportslabkit.utils.count_iter_items(iterable: collections.abc.Iterable) → int[source]#

Consume an iterable not reading it into memory; return the number of items.

Parameters:: iterable (Iterable) – Iterable object
Returns:: Number of items
Return type:: int

sportslabkit.utils.load_config(yaml_path: str) → omegaconf.OmegaConf[source]#

Load config from yaml file.

Parameters:: yaml_path (str) – Path to yaml file
Returns:: Config object loaded from yaml file
Return type:: OmegaConf

sportslabkit.utils.write_config(yaml_path: str, cfg: omegaconf.OmegaConf) → None[source]#

Write config to yaml file.

Parameters:

yaml_path (str) – Path to yaml file
cfg (OmegaConf) – Config object

sportslabkit.utils.pil2cv(image: PIL.Image.Image) → numpy.typing.NDArray[numpy.uint8][source]#

Convert PIL image to OpenCV image.

Parameters:: image (Image.Image) – PIL image
Returns:: Numpy Array (OpenCV image)
Return type:: NDArray[np.uint8]

sportslabkit.utils.cv2pil(image: numpy.typing.NDArray[numpy.uint8], convert_bgr2rgb=True) → PIL.Image.Image[source]#

Convert OpenCV image to PIL image.

Parameters:: image (NDArray[np.uint8]) – Numpy Array (OpenCV image)
Returns:: PIL image
Return type:: Image.Image

sportslabkit.utils.get_fps(path)[source]#

sportslabkit.utils.make_video(frames: collections.abc.Iterable[numpy.typing.NDArray[numpy.uint8]], outpath: sportslabkit.types.types.PathLike, vcodec: str = 'libx264', pix_fmt: str = 'yuv420p', preset: str = 'medium', crf: int | None = None, ss: int | None = None, t: int | None = None, c: str | None = None, height: int | None = -1, width: int | None = -1, input_framerate: int | None = None, logging: bool = False, custom_ffmpeg: str | None = None) → None[source]#

Make video from a list of opencv format frames.

Parameters:

frames (Iterable[NDArray[np.uint8]]) – List of opencv format frames
outpath (str) – Path to output video file
vcodec (str) – Video codec.
preset (str) –
Video encoding preset. A preset is a collection of options that will provide a certain encoding speed to compression ratio. A slower preset will provide better compression (compression is quality per filesize). Use the slowest preset that you have patience for. The available presets in descending order of speed are:
- ultrafast
- superfast
- veryfast
- faster
- fast
- medium (default preset)
- slow
- slower
- veryslow
Defaults to medium.
crf (int) – Constant Rate Factor. Use the crf (Constant Rate Factor) parameter to control the output quality. The lower crf, the higher the quality (range: 0-51). Visually lossless compression corresponds to -crf 18. Use the preset parameter to control the speed of the compression process. Defaults to 23.
ss (int) – Start-time of the clip in seconds. Defaults to 0.
t (Optional[int]) – Duration of the clip in seconds. Defaults to None.
c (bool) – copies the first video, audio, and subtitle bitstream from the input to the output file without re-encoding them. Defaults to False.
height (int) – Video height. Defaults to None.
width (int) – Video width. Defaults to None.
input_framerate (int) – Input framerate. Defaults to 25.
logging (bool) – Logging. Defaults to False.

Todo

add FPS option
functionality to use PIL image
reconsider compression (current compression is not good)

sportslabkit.utils.merge_dict_of_lists(d1: dict, d2: dict) → dict[source]#

Merge two dicts of lists.

Parameters:

d1 (dict) – The first dict to merge.
d2 (dict) – The second dict to merge.

Returns:

The merged dict.

Return type:

dict

sportslabkit.utils.get_git_root()[source]#: Get the root of the git repository.

sportslabkit.utils.download_file_from_google_drive(id, destination)[source]#

sportslabkit.utils.get_confirm_token(response)[source]#

sportslabkit.utils.save_response_content(response, destination)[source]#

sportslabkit.utils.increment_path(path: str | Path, exist_ok: bool = False, mkdir: bool = False) → pathlib.Path[source]#

Increments a path (appends a suffix) if it already exists.

Parameters:

path (Union[str, Path]) – The path to increment.
exist_ok (bool, optional) – If set to True, no increment will be done. Defaults to False.
mkdir (bool, optional) – If set to True, the directory will be created. Defaults to False.

Returns:

The incremented path.

Return type:

Path

sportslabkit.utils.load_keypoints(keypoint_json)[source]#

Loads source and target keypoints from a JSON file.

Parameters:: keypoint_json (str) – Path to JSON file containing keypoints.
Returns:: Source keypoints. target_keypoints (np.ndarray): Target keypoints.
Return type:: source_keypoints (np.ndarray)

sportslabkit.utils.sanitize_url_name(url: str) → str[source]#

Sanitize the URL to create a safe filename for caching.

Parameters:: url (str) – The URL to sanitize.
Returns:: A sanitized version of the URL suitable for use as a filename.
Return type:: str

sportslabkit.utils.fetch_or_cache_model(url: str, dst: PathLike | None = None, hash_prefix: str | None = None, progress: bool = True) → str[source]#

Fetches a model from a URL or uses a cached version if it exists.

Parameters:

url (str) – URL of the object to download.
dst (PathLike | None, optional) – Full path where object will be saved. Defaults to None.
hash_prefix (str | None, optional) – Hash prefix to validate downloaded file. Defaults to None.
progress (bool, optional) – Whether to show download progress. Defaults to True.

Returns:

The path to the downloaded or cached file.

Return type:

str

Attributes#

sportslabkit.utils.PathLike :TypeAlias#