perf: vectorize generalized Pauli-Z construction

[happykawayigt]

Description

Vectorizes the roots-of-unity construction in gen_pauli_z with NumPy operations while preserving the function's signature and bit-for-bit output. Addresses #1887.

Changes

• Replace the Python generator loop with np.exp(np.arange(dim) * c_var).
• Preserve existing behavior and exact output.

Benchmark

Measured on Windows, Python 3.12, NumPy 2.5.1:

Dimension	Iterations	Before	After	Speedup
64	10,000	23.26 us	13.34 us	1.74x
256	3,000	311.78 us	243.97 us	1.28x
1024	300	2391.03 us	2233.95 us	1.07x

Validation

• python -m pytest toqito/matrices/tests/test_gen_pauli_z.py -q (19 passed)
• python -m ruff check toqito/matrices/gen_pauli_z.py
• python -m ruff format --check toqito/matrices/gen_pauli_z.py
• exact equality against the previous implementation for dimensions 1, 2, 3, 4, 5, 8, 16, 31, 64, and 256

Checklist

• Use ruff for errors related to code style and formatting.
• Verify all relevant previous unit tests pass in pytest.
• Check the documentation build does not lead to failures (not run; this code-only change does not affect documentation).

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.

📢 Thoughts on this report? Let us know!

[vprusso]

Thanks for this, and welcome to toqito! I verified the vectorized output is bit-for-bit identical to the previous implementation across a wide range of dimensions, and the existing tests pass. Clean change, and good call dropping the now-unused cmath import. Merging.