The use of internal address (as in conf/__init__.py) and the use of data not on version control should be avoided
Data are stored and referred to using an internal address: http://emfollow.ldas.cit/data instead of address accessible from outside (e.g. an installation at LHO) https://emfollow.ligo.caltech.edu/data/ Moreover, very large data set are stored as json file (up to 90MByte) or executable load and executed form a file. Meaning changing a file will change the executable. In particular for gstlal
em_bright.source_classification(...)
is a function loaded by a file and not in version control.
em_bright.source_classification(..)
should be defined as a module in gwcelery.
Data should be in gwcelery git using the git-lfs feature of git.ligo.org