CI build:sphinx job periodicatlly failing
Seems to be based on some network failure, and happens about half the time:
https://git.ligo.org/Calibration/pydarm/-/jobs/3329208
We should make the doc build not depend on external resources that can not be dependably retrieved.