Visualise pose tracks as a napari Tracks layer

[niksirbi]

Plot body part trajectory. See example implementation here

[niksirbi]

Could also be part of the napari frontend, coercing the tracks layer (see #31)

[niksirbi]

Current plan

We will first try to plot the Trajectory in napari.
The idea is to display individual keypoint tracks via the napari Tracks layer.
The user should be able to:

• color the tracks by various attributes (individual, keypoint, confidence, time)
• select a subset of the data to display (individuals, keypoints, time window, confidence threshold)

This will most likely necessitate a custom widget to supplement the native capabilities of the Tracks layer.

Potential blockers

Issues with handling of NaN values in napari tracks and their properties

[niksirbi]

NaN values continue being a problem, see also https://forum.image.sc/t/tracks-layer-with-blank-nan-values-in-properties/81834

We are considering some hacky workarounds until this issue is fixed in napari. For example, we can set NaNs to a placeholder value (99999? out of canvas? zero?) and make them transparent.

[niksirbi]

After discussion with @sfmig today, we decided to make this issue more concrete, the initial name and scope were too broad and ambiguous. From now on, we take this issue to mean the visualisation of movement poses datasets as a napari Tracks layes, enabling us to plot a "tail" of variable length for each keypoint. This is a saparate issue to visualising the keypoints themselves as a napari Points layer.

For "trajectory plots" and the like in matplotlib, see issue #282