feat: Add JAX acceleration support to z_n_search

[temp-noob]

feat: Add JAX acceleration support to z_n_search

Adds optional JAX-accelerated backend for z_n_search via use_jax parameter.
The JAX implementation computes exact unbinned Z^2_n statistics directly from
event phases, complementing the existing numba-JIT'd binned approach.

Implementation details:

• JAX kernel using vmap for vectorized frequency/fdot grid search
• Conditional @jax.jit with static_argnums for nharm parameter
• Graceful fallback when JAX not installed (HAS_JAX flag)
• Compatible with existing API (1D and 2D search with fdots)

Added comprehensive test suite (16 tests):

• Correctness tests validating peak detection
• Comparison tests (JAX vs standard: >0.95 correlation)
• Benchmark tests showing ~19x speedup on 1D searches (CPU)

Performance on CPU: 1D (no fdot vector) searches gain ~19x speedup. Larger gains expected
on GPU-enabled JAX backends.

---

This change is 

[temp-noob]

@matteobachetti, this is my first time contributing to the repo. Let me know if I can help out with more things or things which would be more crucial.

[matteobachetti]

@temp-noob thanks for the PR.
Could you show this code in action, e.g. through a jupyter notebook modifying one of our tutorials? Also on colab.

[temp-noob]

Sorry for the late reply. And thanks for the review @matteobachetti.
Which tutorial are you referring to? I have this test which contrasts the performance

[temp-noob]

@matteobachetti let me know if I can provide any more clarifications ?

[matteobachetti]

@temp-noob We have the notebooks that we use for documentation here: https://github.com/stingraysoftware/notebooks

If you could modify the notebook with Z searches and make it work with your code, we would have a clearer idea of the actual usability of the code, besides a few tests (which are still necessary and important, of course!)

[temp-noob]

@matteobachetti added this notebook for the jupyter notebook tests: StingraySoftware/notebooks#129
Let me know if I can expand more on the reproducing steps.

[temp-noob]

@matteobachetti let me know if this looks good to you.