Trotter Costs

[fdmalone]

I'm opening this as a draft at the moment for some feedback, but it may be good to merge and follow up with some issues later:

Issues:

1. The QVR is the naive way which I think can be improved from n log (1/e) to log (1/e).
2. We had some questions whether the potential was suboptimally implemented (c.f. the first quantized block encoding potential doesn't have this eta^2 cost?
3. I started to go down a bit of a rabbit hole with fixed point arithmetic. I would like to revisit the interpolation section in the notebook to use fixed point and a proxy for QROM so that we check the quality of interpolation directly. Could be a nice demonstration of classical_vals
4. I have a huge hack to render the bloq counts for QROM which I think @tanujkhattar had a better solution to.

I think 4 should at least be resolved before merging. But yeah probably good to get some feedback before it gets any bigger.

[fdmalone]

I committed some things accidentally, need to rebase.

[fdmalone]

@ncrubin I restructured a bit if you want to take a look

[fdmalone]

@mpharrigan ready for review

[mpharrigan]

wow! Lots of nice work.

Some nits to address

[fdmalone]

@mpharrigan I addressed everything I think.