Umi format explore fix

[alneberg]

The methods len_after_index and len_before_index on the adaptor parts were not correct for index regions containing UMIs. This should fix it and also:

• Rename the methods to indicate index_region instead of index
• Remove the currently unused AdaptorPart.get_mask

[remiolsen]

I was not happy with this part when I first wrote it, and am not happy with seeing for a second time. Problem might be I wrote it too generic. In reality it's only the first case that is valid, i.e. the INDEX+UMI configuration provided by that company whose name is mostly known by it's three letter abbreviation.

I'm leaning towards allowing this. As I stated in the previous PR, these two adaptor classes needs a proper cleaning up.
Just as a side-note, from purely semantics it's not clear to me what an "adaptorpart" is and what such an object would be a reflection of.

[alneberg]

Yes, I wanted to support the same type as you did in the other code essentially.

For me each adaptor would have two adaptorparts, i.e. i7 and i5. Does that make sense?

[remiolsen]

Don't worry, I'm being nitpicky over the name. It might have been more clear if AdaptorPart was a child class of Adaptor. But that's currently a little bit weird since AdaptorParts are instance variables of Adaptor.

[alneberg]

No worries, but no I think child classes would have been more confusing actually. Their just the two parts of an Adapter to DRY up all the methods and attributes which are identical between i7 and i5.

[remiolsen]

In isolation think it would be less confusing. BUT it doesn't really fit with how the adaptor class is currently used. So it would mean a whole lot of refactoring for the sake of idealism, which I don't advocate for.