separating fake communicator from mpifile

[iparask]

Isolating the FakeCommunicator class. It is a very useful class, but the star import from mpi4py affects mpi4py's behavior I do not need a whole MPI world or the world does not exist (running other types of parallelization).

This does not change the functionality of the mpi.py as the FakeCommunicator class is imported by it. It just makes sure that the FakeCommunicator can be imported in isolation.

PR simonsobs/sotodlib#1096 depends on this

[JBorrow]

Are the test failures unreleated?

[iparask]

I do not think so.

[JBorrow]

Oh you gotta add the new file to the list of files included in the meson distribution (see meson.build)

[codecov]

Codecov Report

Attention: Patch coverage is 64.00000% with 9 lines in your changes missing coverage. Please review.

Project coverage is 39.79%. Comparing base (a61ca1a) to head (d64ad3e).
Report is 4 commits behind head on master.

Files with missing lines	Patch %	Lines
pixell/fake_communicator.py	62.50%	9 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #286      +/-   ##
==========================================
+ Coverage   39.78%   39.79%   +0.01%     
==========================================
  Files          29       30       +1     
  Lines       10436    10440       +4     
==========================================
+ Hits         4152     4155       +3     
- Misses       6284     6285       +1     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[JBorrow]

Looks reasonable to me but I will let the others who are more familiar with the core comment.

[amaurea]

Implemented