Fix I3HighestEparticleExtractor

[sevmag]

This PR addresses the bugs explained in #821.

Solving Issues

1. Negative visible length

set visible length to 0 if it would be negative

2. Track filtering only for top-level primaries

Look, if a track particle is in the subtree of any primary, not if the head-level primary is in the primaries list

3. Wrong return in get_bundle_HEP

return the right tuple

Additionally, I introduced the get_descendants function, which accelerates this code snippet:

graphnet/src/graphnet/data/extractors/icecube/i3highesteparticleextractor.py

Lines 399 to 414 in f3faff1

	e_p = np.array(
	[
	np.array([p.energy, p])
	for p in particles
	if (
	(
	(p.energy > min_e)
	& (
	frame[self.mctree].get_primary(p.id)
	in primaries
	)
	)
	& (not p.is_track)
	)
	]
	).T

by not iterating through all particles, but only the descendants of the primaries

[Aske-Rosted]

Looks good to me! Nice contribution, thanks.

[sevmag]

I made some more commits. They address further issues I explained in #821. Let me know if you agree, @Aske-Rosted.

[Aske-Rosted]

I made some more commits. They address further issues I explained in #821. Let me know if you agree, @Aske-Rosted.

I agree with the new changes as well. I also ran into the muongun issues a while back, but managed to remove them with the logic statements I had in play, guess some of the changes to the logic made them reappear...