A fully functional validation module for tracking the pupil marker

[KamranBinaee]

Make sure you have:

• [not sure ] added the necessary updates to docs/whatsnew.rst
• [ not yet] added documentation for new features
• [ not yet] run bumpversion and reset conda build number if you are preparing a new release

This is an almost complete working marker detector. The parameter passing is not optimal though. I tried using kwargs but it didn't work. It might be due to the fact that in line 124 base.py the kwargs are not being passed correctly? I'm not sure. Ideally the circle detector parameters should be passed through yaml and **kwargs.

[phausamann]

@KamranBinaee I did some refactoring before merging the Validation class. Should have mentioned that earlier, sorry about that. I did a rebase onto devel to incorporate those changes.

[phausamann]

The parameter passing isn't optimal, I agree, but I don't see another way to do it tbh. Keyword arguments need to be written out so that the @process decorator can pick them up for the config mechanism.

[phausamann]

One possibility would be to pass a tracker_params argument as a dictionary to the CircleDetector like this:

@process("circle_detector")
class CircleDetector(BaseProcess):
    """ Circle detector for the world video stream. """

    def __init__(
        self,
        scale=0.5,
        detection_method="pupil",
        tracker_params,
        **kwargs,
    ):
        """ Constructor. """
        super().__init__(**kwargs)

        self.circle_tracker = CircleTracker(
            scale=scale,
            detection_method=detection_method,
            **tracker_params,
            **kwargs,
        )

[KamranBinaee]

Closed this PR in favor of #77