[wsttiger]

Adds a host-side lattice tutorial comparing JW, BK, standard superfast, and binary-tree generalized superfast mappings of one four-mode interacting spinless ring. An independent occupation-bit Hamiltonian verifies the tensor coefficients and fermionic signs. Each superfast parity sector is projected onto the public loop stabilizers’ joint +1 eigenspace before comparing all eight physical eigenvalues, including multiplicities. The example also checks stabilizer algebra and reports Pauli weights, code dimensions, and qubit counts without assuming an asymptotic winner.

Validation on CUDA-Q 0.16.0.post1, Python 3.12.3:

• Standalone example passed in 0.01 s; four declared qubits, one independent stabilizer and code dimension eight for each superfast parity. Maximum spectral error was about 3.55e-15.
• python3 -m pytest tests/python/test_examples.py -k fermion_mappings_lattice -q: 1 passed in 1.90 s.
• Sphinx HTML build with -W --keep-going: passed independently.
• YAPF 0.43.0 and whitespace checks passed.

Independent plan 07 PR against main; uses public mapping APIs with no library change or chemistry/phase-generation dependency.

Development is hosted in wsttiger/cudaq-algorithms. This replaces #86 after relocating its branch from upstream to the fork. Existing review discussion remains on that linked PR. The latest commit history is preserved unchanged; this relocation does not rerun scientific validation.

[amezzacapo1]

I would add another example with high number of qubits (e.g. up to ~100 qubits) where one can appreciate the difference between the mappings more.