Arguments

class deeptrack.features.Arguments(_input: Any = [], **kwargs: Dict[str, Any])

Bases: Feature

A convenience container for pipeline arguments.

The Arguments feature allows dynamic control of pipeline behavior by providing a container for arguments that can be modified or overridden at runtime. This is particularly useful when working with parameterized pipelines, such as toggling behaviors based on whether an image is a label or a raw input.

Methods

get(image, **kwargs)

Passes the input image through unchanged, while allowing for property overrides.

Examples

A typical use-case is:

>>> arguments = Arguments(is_label=False)
>>> image_loader = (
...     LoadImage(path="./image.png") >>
...     GaussianNoise(sigma = (1 - arguments.is_label) * 5)
...     )
>>> image_loader.bind_arguments(arguments)

>>> image_loader()  # Image with added noise.
>>> image_loader(is_label=True)  # Raw image with no noise.

For a non-mathematical dependence, create a local link to the property as follows:

>>> arguments = Arguments(is_label=False)
>>> image_loader = (
...     LoadImage(path="./image.png") >>
...     GaussianNoise(
...         is_label=arguments.is_label,
...         sigma=lambda is_label: 0 if is_label else 5
...     )
... )
>>> image_loader.bind_arguments(arguments)

>>> image_loader()              # Image with added noise
>>> image_loader(is_label=True) # Raw image with no noise

Keep in mind that, if any dependent property is non-deterministic, they may permanently change:

>>> arguments = Arguments(noise_max_sigma=5)
>>> image_loader = (
...     LoadImage(path="./image.png") >>
...     GaussianNoise(
...         noise_max_sigma=5,
...         sigma=lambda noise_max_sigma: rand() * noise_max_sigma
...     )
... )

>>> image_loader.bind_arguments(arguments)

>>> image_loader().get_property("sigma") # Example: 3.27...
>>> image_loader(noise_max_sigma=0) # 0
>>> image_loader().get_property("sigma") # Example: 1.93...

As with any feature, all arguments can be passed by deconstructing the properties dict:

>>> arguments = Arguments(is_label=False, noise_sigma=5)
>>> image_loader = (
...     LoadImage(path="./image.png") >>
...     GaussianNoise(
...         sigma=lambda is_label, noise_sigma: (
...             0 if is_label else noise_sigma
...         )
...         **arguments.properties
...     )
... )
>>> image_loader.bind_arguments(arguments)

>>> image_loader()  # Image with added noise.
>>> image_loader(is_label=True)  # Raw image with no noise.

Methods Summary

get(image, **kwargs)

Process the input image and allow property overrides.

Methods Documentation

get(image: Any, **kwargs: Dict[str, Any]) → Any

Process the input image and allow property overrides.

This method does not modify the input image but provides a mechanism for overriding arguments dynamically during pipeline execution.

Parameters

imageAny

The input image to be passed through unchanged.

**kwargsAny

Key-value pairs for overriding pipeline properties.

Returns

Any

The unchanged input image.