as_list

deeptrack.utils.as_list(obj: Any) → list[Any]

Ensure that the input is a list.

Converts the input to a list if it is iterable and not a string or bytes; otherwise, it wraps it in a list.

Note: If obj is a PyTorch Tensor, this function will return a list of its elements along the first dimension (e.g., for a 2D tensor, the result will be a list of 1D tensors). If you want to wrap the entire tensor in a list, use [obj] explicitly.

Parameters

obj: Any

The object to be converted or wrapped in a list.

Returns

list[Any]

The input object as a list.

Examples

from deeptrack.utils import as_list

Wrap a scalar in a list:

>>> as_list(5)
[5]

>>> as_list(None)
[None]

Pass through a list unchanged:

>>> as_list([1, 2, 3])
[1, 2, 3]

Convert a tuple or set to a list:

>>> as_list((1, 2, 3))
[1, 2, 3]

>>> sorted(as_list({3, 2, 1}))
[1, 2, 3]

Convert a generator to a list:

>>> generator = (x * 2 for x in range(3))
>>> as_list(generator)
[0, 2, 4]

Strings and bytes are treated as atomic (not split):

>>> as_list("abc")
['abc']

>>> as_list(b"xyz")
[b'xyz']

NumPy arrays become lists of elements:

>>> import numpy as np
>>> as_list(np.array([1, 2, 3]))
[1, 2, 3]

PyTorch tensors become lists of elements along the first dimension (if PyTorch is available):

>>> import torch
>>> t = torch.tensor([[1, 2], [3, 4]])
>>> as_list(t)
[tensor([1, 2]), tensor([3, 4])]