REFACTOR: make multiband likelihood call Interferometer.get_detector_response

[ColmTalbot]

See https://git.ligo.org/lscsoft/bilby/-/merge_requests/1379

I think this has a conflict with #842.

Based on ongoing discussion about implementing different detector responses, it would be nice for the likelihoods to all access the response functions uniformly. The MB likelihood currently has an almost verbatim copy of the Inteferometer.get_detector_response method. This MR provides a minimal change to be able to directly call that method.…response

[mj-will]

#842 is merged now, so hopefully the conflicts can be resolved.

[ColmTalbot]

This will actually need more extensive changes. I think it would require moving the beam_pattern_reference_time to the Interferometer. I think this would make sense as a change, possibly as another optional argument or an attribute that needs to be manually set, e.g., by the likelihood classes. I'll mark this as draft for now.

[ColmTalbot]

@SMorisaki I can't request a review from you, but please can you review this?

[mj-will]

@ColmTalbot @SMorisaki is this in a state where we can get it into the next release?

[mj-will]

This looks good to me. I've checked and AFAICT it's apply the exact same operations.

This has reminded me we don't have great tests for the mutliband likelihood so it would be great if we could add more in future.

Also, my might want to flag this change to reviewers since it will impact a part of the code that is used for some production analyses.

[GregoryAshton]

This also looks good to me.