Add implementation for SelectTHC bloq

[arettig]

Fixes #386. This adds the underlying QROAM and RealGivensRotationByPhaseGradient
bloqs to implement the SelectTHC bloq. Differences in Toffoli costs vs. the paper were
documented in a test.

[review-notebook-app]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB

[arettig]

This adds the implementation for the THC select bloq laid out by @fdmalone. This caused some slight changes in the Toffoli counts for this bloq. The differences as best I can tell are the following:

• QROAMClean - QROAMClean adjoint automatically uses the optimal block size, ignoring the kr1/kr2 parameters, which can change gate count. There are also some small optimizations for loading the data.
• RealGivensRotationByPhaseGradient - We have to prepare the phase gradient state, adding some gates (a small number asymptotically).
• CSWAP - The paper does not account for the swapping of the $\mu$ and $\nu$ registers as it could be combined with the PREPARE step, but it is included explicitly in the bloq.

[gemini-code-assist]

Code Review

This pull request implements the THCRotations bloq, which rotates into the THC basis using a Givens rotation network. It adds the helper _leaf_tensor_to_givens_rotations to compute rotation angles from a THC leaf tensor, updates SelectTHC to integrate these rotations, and adds comprehensive tests and documentation. The reviewer feedback highlights critical issues regarding the handling of one-body terms when two_body_only is False. Specifically, num_terms in THCRotations should account for num_spatial one-body terms, and the bloq needs to accept nu_eq_mp1 to correctly offset the selection index. Additionally, SelectTHC currently lacks a mechanism to accept or compute one-body rotation angles when initialized with a leaf tensor eta for the full Hamiltonian.