Fix a bug with soft clips in LeftAlignIndels

[davidbenjamin]

Closes #6765. @fleharty can you review this fix?

[fleharty]

@davidbenjamin
I don't see any tests that have leading or trailing softclips. Maybe I missed them?
Also, we are now using hard clipping a lot more frequently too, so could you add a few tests that include hard clipped reads on both ends.

Otherwise I think this looks good.

[davidbenjamin]

@fleharty See lines 789-90 in AlignmentUtilsUnitTest.java for soft clips. I will do the same thing for hard clips.

[jamesemery]

@davidbenjamin Interesting. In #6634 I added a fix similar to this as now HC is holding onto bases that were clipped by the various layers of low edge and low quality end removal code in the form of soft clips. I suspect your fix might be a little more elegant than mine. Could we hold off on merging this for a bbit to save ourselves the trouble when we try to rebase?

[davidbenjamin]

@fleharty It now handles hard clips.

@jamesemery Could you double-check whether this and #6634 actually interfere with one another? This PR doesn't affect HC at all. That is, it neither touches HC code nor affects its output. All it does is make an AlignmentUtils method correct for a case — leading clips adjacent to an indel — that doesn't come up in HC.

[fleharty]

Great thanks!

[davidbenjamin]

@jamesemery Should I merge this PR before Friday's release, or hold off?

[jamesemery]

@davidbenjamin go ahead and merge. I was conflating this code with some other code in the DRAGEN branch that also deals with Indels.