search

tomasstolker / species

Toolkit for atmospheric characterization of directly imaged exoplanets
https://species.readthedocs.io
MIT License
22 stars 10 forks source link

Discussion: Use of specutils.Spectrum class? #128

Open kelle opened 10 hours ago

kelle commented 10 hours ago

The issue about fits encoding(#127) got me thinking....would it be worthwhile to implement use of the specutils.Spectrum class? It would increase the dependency on specutils, an astropy coordinated package. I see pros and cons and wonder if you've thought about it?

tomasstolker commented 10 hours ago

Thanks for the suggestion! I think it is an interesting idea as it would be a more generalized approach for working with spectra, both as input and output. It would require quite a lot of rewriting of species though, so I am not fully convinced (yet) if the pros would outweigh the cons. From what I understood, specutils do not handle covariances while these can be (optionally) used in a spectral fit with species.

kelle commented 9 hours ago

Hmmm, yes, looks like that's true. It's been discussed, but not implemented: https://github.com/astropy/astropy/pull/16690