feat(plot): add exclude_keys to filter trace plots (closes #730)

[christianescamilla15-cell]

Summary

Addresses #730 by adding an opt-in way to filter nuisance parameters out of the checkpoint trace plots produced by `bilby_mcmc`. The motivation in the issue is that with ~20 calibration parameters per detector, trace plots can easily hit 60+ axes and blow up memory on long runs.

Closes #730.

Design

Rather than hard-removing calibration parameters (which @ColmTalbot correctly pointed out can still be diagnostically useful), this is opt-in with glob-style patterns, matching the spirit of arviz's `var_names` / `filter_vars` selection.

```python
bilby.run_sampler(
sampler="bilby_mcmc",
plot_exclude_keys=["recalib_*"],
...
)
```

Changes

`bilby/bilby_mcmc/chain.py`

• Added `exclude_keys` parameter to `Chain.plot()`
• Uses `fnmatch.fnmatchcase` for glob-style pattern matching
• Filters `self.keys` before computing the figure layout (so the figure height shrinks with the filtered count)
• Applied to both trace axes and histogram axes
• Added comprehensive docstring with the full parameter list

`bilby/bilby_mcmc/sampler.py`

• Added `plot_exclude_keys` kwarg to `Bilby_MCMC.init`
• Stored as `self.plot_exclude_keys`
• Threaded through `print_long_progress` → `plot_progress` → `chain.plot`
• `plot_progress` signature extended with `exclude_keys=None`

Backwards compatibility

• Default behaviour is unchanged (`exclude_keys=None` plots everything)
• Existing users see identical output unless they opt in

Future work (not in this PR)

The dynesty sampler's trace plotting is generated via dynesty's own machinery and would need a separate adapter to accept `exclude_keys`. Happy to follow up in a separate PR once this approach is approved.

Test plan

• `py_compile` passes for both modified files
• `fnmatch` patterns work for typical cases (`recalib_*`, `spin_?`)
• Full test suite in CI (reviewer to confirm)

References

• Issue Remove calibration plots from trace (bilby_mcmc, dynesty) #730 — original report by @gregory.ashton
• arviz trace plot API: https://python.arviz.org/en/stable/api/generated/arviz.plot_trace.html