RuntimeWarning using binomial_confidence_interval on Mac Silicon

[taldcroft]

On Mac Silicon:

In [1]: from chandra_aca.star_probs import binomial_confidence_interval

In [2]: binomial_confidence_interval(8, 50)
/Users/aldcroft/miniconda3-arm/envs/ska3/lib/python3.11/site-packages/scipy/stats/_continuous_distns.py:696: RuntimeWarning: invalid value encountered in _beta_isf
  return _boost._beta_isf(x, a, b)
Out[2]: (0.16, 0.11509604953864323, 0.21835391570110868)

On Mac x86, the same command runs with no warning and gives the same answer. I think this is addressed by this issue:

https://github.com/scipy/scipy/issues/20208

It looks like this is fixed in scipy 1.13.1 or 1.14, but I haven't confirmed.

[jeanconn]

For now, do we want to try to upgrade, programmatically ignore the warning in chandra_aca, or just not pay attention to the warning?