BUG: ensure infinite ACT estimates are properly handled

[ColmTalbot]

@lalit-pathak reported the following error. This happens when no points are accepted during the MCMC while using ACT-walk. This is most likely a sign of some underlying pathology in the likelihood/sampling space, but we should handle the case smoothly.

Traceback (most recent call last):
  File "/home/lalit.pathak/.conda/envs/gwsim_env/lib/python3.10/multiprocessing/pool.py", line 125, in worker
    result = (True, func(*args, **kwds))
  File "/home/lalit.pathak/.conda/envs/gwsim_env/lib/python3.10/multiprocessing/pool.py", line 48, in mapstar
    return list(map(*args))
  File "/home/lalit.pathak/.conda/envs/gwsim_env/lib/python3.10/site-packages/bilby/core/sampler/dynesty_utils.py", line 198, in __call__
    while self.cache[0][2] < args.loglstar:
  File "/home/lalit.pathak/.conda/envs/gwsim_env/lib/python3.10/site-packages/bilby/core/sampler/dynesty_utils.py", line 207, in cache
    self.build_cache()
  File "/home/lalit.pathak/.conda/envs/gwsim_env/lib/python3.10/site-packages/bilby/core/sampler/dynesty_utils.py", line 223, in build_cache
    check_interval = int(np.ceil(self.act))
OverflowError: cannot convert float infinity to integer

I also noticed a second case where an infinity can sneak in that will break things, if there are fewer than ten live points, the acceptance walk method will fail on current master.

Traceback (most recent call last):
  File "/Users/colm/modules/bilby/examples/core_examples/linear_regression.py", line 57, in <module>
    result = bilby.run_sampler(
  File "/Users/colm/modules/bilby/bilby/core/sampler/__init__.py", line 234, in run_sampler
    result = sampler.run_sampler()
  File "/Users/colm/modules/bilby/bilby/core/sampler/base_sampler.py", line 97, in wrapped
    output = method(self, *args, **kwargs)
  File "/Users/colm/modules/bilby/bilby/core/sampler/dynesty.py", line 517, in run_sampler
    out = self._run_external_sampler_with_checkpointing()
  File "/Users/colm/modules/bilby/bilby/core/sampler/dynesty.py", line 637, in _run_external_sampler_with_checkpointing
    self.sampler.run_nested(**sampler_kwargs)
  File "/Users/colm/modules/dynesty/py/dynesty/sampler.py", line 1057, in run_nested
    for it, results in enumerate(
  File "/Users/colm/modules/dynesty/py/dynesty/sampler.py", line 877, in sample
    u, v, logl, nc = self._new_point(loglstar_new)
  File "/Users/colm/modules/dynesty/py/dynesty/sampler.py", line 462, in _new_point
    self.update_proposal(blob, update=self.nqueue <= 0)
  File "/Users/colm/modules/bilby/bilby/core/sampler/dynesty_utils.py", line 56, in update_user
    self.walks = (self.walks * delay + n_target / accept_prob) / (delay + 1)
ZeroDivisionError: float division by zero

I'm opening this as a test of migrating open MRs from gitlab to here. https://git.ligo.org/lscsoft/bilby/-/merge_requests/1383 is the original MR.

[GregoryAshton]

Thanks @ColmTalbot this cleans things up nicely. Happy to approve.