The current implementation of the HOME_DATA_DIR setting variable assumes that users will have local user accounts and home directories on the system running the pipeline. This isn't always the case, e.g. when running the pipeline in a container.
It would be good if we could specify where the home directories are instead of relying solely on os.path.expanduser to find it. For example, if we had a HOME_DATA_ROOT setting that if set would cause the pipeline to search for extra input data in HOME_DATA_ROOT/<username>/HOME_DATA_DIR. If left unset, the current behaviour is maintained, i.e. the pipeline searches ~<username>/HOME_DATA_DIR.
The current implementation of the
HOME_DATA_DIR
setting variable assumes that users will have local user accounts and home directories on the system running the pipeline. This isn't always the case, e.g. when running the pipeline in a container.It would be good if we could specify where the home directories are instead of relying solely on
os.path.expanduser
to find it. For example, if we had aHOME_DATA_ROOT
setting that if set would cause the pipeline to search for extra input data inHOME_DATA_ROOT/<username>/HOME_DATA_DIR
. If left unset, the current behaviour is maintained, i.e. the pipeline searches~<username>/HOME_DATA_DIR
.