Obtaining lists of supernovae of given type

[rbiswas4]

I want to look at all builtin sncosmo sources that are of a particular type of SN. For example, I may want to obtain a list of supernovae which have type Ib. This might be complicated as templates are often called Ib Ib/c etc. So perhaps, an easier and more meaningful grouping beyond just Ia and core collapse is :

• group of IIP (along with L and N)
• Ib, Ic, Ibc, Ib/c, IIb

Happy to hear other suggestions on how I should group supernovae, but we do not have a given way of obtaining such lists, though it is not hard to write a couple of lines to group them. I think it would be nice to provide a method to actually give users this.

Code to make the above grouping, but we can change it for other suggestions is:

sources = sncosmo.builtins._SOURCES.get_loaders_metadata()  

iips  = list(source['name'] for source in sources if 'sn ii' in source['type'].lower() and not source['type'].endswith('b'))
stripped_envelopes = list(source['name'] for source in sources if not (('sn ii' in source['type'].lower() and not source['type'].endswith('b')) or  ('Ia' in source['type'])))

While doing this, I was also thinking that it would be nice to be able to interrogate a source and find out the information in the builtins like name, type, version etc. A reason not to have this, is that a source may not be a builtin. But, as we move to thinking more about all kinds of transients, it would be good to have a way of obtaining the type of a source. Could we add this to attributes of the Source, even though the user may easily put None as they can with the version?

[kbarbary]

I think an important part of this would be standardizing the names. For example, use either Ib/c or Ibc, but not both.

If we provide a function, it could support multiple levels of groupings:

source_names('Ib')   # return list of all names that have type Ib
source_names('core collapse')  # return list of all names that have any core-collapse SN type

We may want to actually use names like SN Ib and SN CC to distinguish from non-SN transients.

[rbiswas4]

think an important part of this would be standardizing the names. For example, use either Ib/c or Ibc, but not both.
Sure, we can standardize these. It would be a backward incompatible change though. I would vote for Ibc rather than Ib/c
but don't have a strong preference.

We may also want to take the opportunity to encode transient type, subtype, subsubtype, model in the standardized names :
SNIaNormSALT_2.4 ?

We may want to actually use names like SN Ib and SN CC to distinguish from non-SN transients.
Yes! But can we avoid spaces, and use underscores or hyphens instead?