THC prepare

[fdmalone]

Add THC prepare bloqs: uniform superposition, and prepare.

Some issues which arose:

1. We need more comparators (less than constant, equals a constant, greater than a constant, ...) I added these as placeholders.
2. For tcomplexity to work I had to add t_complexity methods to several single qubit gates TComplexity is currently broken, XGate doesn't support decomposition
3. relies on Add OnEach bloq #346 Multi controlled pauli #349 and Less than equal bloqs. #348
4. Prepare t complexity is currently not working because controlled bloq (CZ) throws an error during decomposition (ZGate doesn't support decomposition).
5. The complexity of the uniform superposition bloq is higher than the paper because inverting the inequalities should be free and the rotation cost is much higher due to the THC paper using a phased gradient gate rather than trying to synthesize the rotation. Similarly the prepare step is larger because of difference in QROAM Many Variations of QROAM #368

[fdmalone]

cc @mpharrigan

[mpharrigan]

Is this ready for review? can you resolve conflicts?

[fdmalone]

I would look tomorrow if you have a chance. There were some dirty hacks in the mean time to do some accounting.

[fdmalone]

@mpharrigan PTAL

[fdmalone]

weird merging #419 closed this somehow.

[mpharrigan]

Any time you use the word "fixes" and an issue, apparently it will try to close the thing that it "fixes". Makes sense for issues (but clearly not for PRs)

[fdmalone]

So the comparator issue is now resolved since #394, which removes the need for my hacky code. I'll incorporate the changes and any CirqGateAsBloq deprecations as necessary before merging.

[mpharrigan]

some more notebook things

[fdmalone]

I've opened #425 to track the notebook_utils clean up. This is good to go I think @mpharrigan