BUG: ztf_functions.py should be importable even when there is no access permission to the s3 bucket

[zoghbi-a]

ztf_functions.py accesses a remote bucket during import. This is probably not a good idea as it prone to failure when the permission to the bucket is not setup correctly.

[bsipocz]

This problem has already been captured in #311. Note, the ZTF access issue cannot be remedied yet as it is not yet have an ODR place.

I would go ahead and close this as a duplicate, but if you have an idea for a suitable workaround, feel free to reopen (I don't think it's any particularly useful to skip ZTF and yet create the light curves when there is no access to the bucket as in practice that dataset is heavily used in all the follow-up notebooks).

[zoghbi-a]

I agree that skipping ztf is not an option. I was thinking just move it from the import section. If that is going to be done as part of #311, then feel free to close this.

[bsipocz]

OK, that makes a lot of sense, reopening it for tracking the python script importability rather than the bucket issue.

[troyraen]

@bsipocz is it ok if I fix this? This module uses multiprocessing internally which complicates this a bit, but I think I see how to lazy load that variable in an efficient way.

[bsipocz]

Sure! I was only planning to move the quoted line into a function body, so it's not touched at import time, but anything would work really.