The code was updated so that confidence_to_alpha can now be a function. As the docstring describes:

confidence_to_alpha: Union[bool, Callable[[float], float], default=False
If False, all keypoints will be plot with alpha=1. Otherwise, this can be defined as a function f: [0, 1] -> [0, 1] such that the alpha value for a keypoint will be set as a function of its score: alpha = f(score). The default function used when True is f(x) = max(0, (x - pcutoff)/(1 - pcutoff)).

As an example of the resulting video obtained with the plotting function, here are the results on a toy project of running create_video_with_all_detections with two different values:

deeplabcut.create_video_with_all_detections(..., confidence_to_alpha=False)

pcutoff = 0.1
def confidence_to_alpha(c):
    return np.clip((c - pcutoff) / (1 - pcutoff), 0, 1)

deeplabcut.create_video_with_all_detections(..., confidence_to_alpha=confidence_to_alpha)

In the second video, the low scores for the red keypoint in the top right corner is very clear, while it cannot be inferred from the first video with confidence_to_alpha=False.