ConditionalSetFeature

class deeptrack.features.ConditionalSetFeature(on_false: Feature | None = None, on_true: Feature | None = None, condition: PropertyLike[str | bool] = True, **kwargs: dict[str, Any])

Bases: StructuralFeature

Conditionally resolves one of two features based on a condition.

This feature allows dynamically selecting and resolving one of two child features depending on whether a specified condition evaluates to True or False.

The condition parameter specifies either: - A boolean value (default is True). - The name of a property to listen to. For example, if condition=”is_label”, the selected feature can be toggled as follows:

>>> feature.resolve(is_label=True)   # Resolves `on_true`
>>> feature.resolve(is_label=False)  # Resolves `on_false`
>>> feature.update(is_label=True)    # Updates both features

Both on_true and on_false are updated during each call, even if only one is resolved.

Parameters

on_false: Feature, optional

The feature to resolve if the condition is False. If not provided, the input image remains unchanged.

on_true: Feature, optional

The feature to resolve if the condition is True. If not provided, the input image remains unchanged.

condition: str or bool, optional

The name of the conditional property or a boolean value. If a string is provided, its value is retrieved from kwargs or self.properties. If not found, the default value is True.

**kwargs: dict of str to Any

Additional keyword arguments passed to the parent StructuralFeature.

Attributes

__distributed__: bool

Indicates whether this feature distributes computation across inputs.

Methods

get(image: Any, condition: str | bool, **kwargs: dict[str, Any]) -> Any

Resolves the appropriate feature based on the condition.

Examples

>>> import deeptrack as dt
>>> import numpy as np

Define two Gaussian noise features: >>> true_feature = dt.Gaussian(sigma=0) >>> false_feature = dt.Gaussian(sigma=5)

— Using a boolean condition — Combine the features into a conditional set feature. If not provided explicitely, condition is assumed to be True: >>> conditional_feature = dt.ConditionalSetFeature( … on_true=true_feature, … on_false=false_feature, … )

Define an image: >>> image = np.ones((512, 512))

Resolve based on the condition: >>> clean_image = conditional_feature(image) # If not specified, default is True >>> print(clean_image.std()) # Should be 0 0.0

>>> noisy_image = conditional_feature(image, condition=False)
>>> print(noisy_image.std())  # Should be ~5
4.987707046984823

>>> clean_image = conditional_feature(image, condition=True)
>>> print(clean_image.std())  # Should be 0
0.0

— Using a string-based condition — Define condition as a string: >>> conditional_feature = dt.ConditionalSetFeature( … on_true=true_feature, … on_false=false_feature, … condition = “is_noisy”, … )

Resolve based on the conditions: >>> noisy_image = conditional_feature(image, is_noisy=False) >>> print(noisy_image.std()) # Should be ~5 5.006310381139811

>>> clean_image = conditional_feature(image, is_noisy=True)
>>> print(clean_image.std())  # Should be 0
0.0

Methods Summary

get(image, *, condition, **kwargs)

Resolve the appropriate feature based on the condition.

Methods Documentation

get(image: Any, *, condition: str | bool, **kwargs: dict[str, Any])

Resolve the appropriate feature based on the condition.

Parameters

image: Any

The input image to process.

condition: str or bool

The name of the conditional property or a boolean value. If a string is provided, it is looked up in kwargs to get the actual boolean value.

**kwargs:: dict of str to Any

Additional keyword arguments to pass to the resolved feature.

Returns

Any

The processed image after resolving the appropriate feature. If neither on_true nor on_false is provided for the corresponding condition, the input image is returned unchanged.