Closed zkbt closed 1 year ago
@catrionamurray , this should now at least ignore values that are not ok
in the normalization. It will still happily normalize by a negative number and therefore produce negative uncertainties, but at least it warns you about it now. I think for the vast majority of applications you could probably consider negative median flux as making a wavelength effectively useless.
This is great, thank you! I think that maybe a statement in the init file is missing? Should there be from .fluxlike import *
in the /get/init.py?
Oh, yeah, I guess so. I think it works for me without it because the only place it's needed is as a direct import into the Rainbow
class definition...
...but it's possible that including in the get/__init__.py
helps for some other versions of Python?
(Just updated this directly in the develop
branch).
get_ok_data
as (possibly pointless) wrapper to get flux with bad data replaced withnan
.normalize
to now useok
.normalize
if any of the normalization factors are negative, with suggestion for fix.normalize
. Closes #222 . Passes tests.