Implement Mappers for d-level Operators

[kaelynj]

What should we add?

From the available QubitMappers in the docs, it appears to me that there may be room to include other mappings such as those described in https://arxiv.org/abs/1909.12847 or this PRB paper. The DirectMapper is already implemented (what the authors call the Unary Encoding in the arxiv paper), but others based on the Gray Code or a mixture of Gray Codes separated into "blocks" --- each of which is mapped via the DirectMapper --- may be useful features to implement.

Additionally, these also might be useful once #958 or #795 are finished for systems with mixtures of bosonic and fermionic operators.

Would this be a helpful feature to work on? Or should the issues above be focused on first?

[mrossinek]

Additional QubitMapper implementations are welcome contributions. However, instead of a "catch-all" issue like this I would suggest that on a case by case basis a feature request ought to be opened where details of the implementation can be discussed. This should be done one at a time as someone has interest in working on them.

Some notes:

• the mappers recently underwent a major refactoring: QubitConverter Refactoring #967
• mixed particle-type systems will require a Implement a MixedOperator #795 and corresponding MixedMapper protocol to be added first
• before adding mappers for bosonic cases, the work on Bosonic operators #1085 needs to be finished

[SamD-1998]

Hi, is there something else to be done in this issue or is this just a discussion which has ended?

I see that type: feature request label has been removed.

[kaelynj]

I'm happy to continue the discussion on adding a specific feature instead of a catch-all as the above comment points out. I was mostly waiting for the refactoring and other merges to be completed.