With the code freeze for the next LTS version of astropy coming up soon, I think it would be a good time to discuss the thought of migrating two of the more useful features of ccdproc into the core of astropy, so that they are long term available in the code. So here are some thoughts on adding them as pro's and con's:
Migrating CCDData:
PRO: The API is stable and the object is fairly mature and well tested
PRO: It gives a good use case for NDData and an example of its usefulness
PRO: It probably has more use cases outside of just processing CCDs.
CON: We may have to maintain two code bases until it is fully depreciated in ccdproc or people are no longer using older versions of astropy
CON: There may be NDData changes coming, and this may change the guts of CCDData (although unlikely to change the API)
Migrating ImageFileCollection
PRO: It is probably one of the more useful things in here and applicable to all types of FITS files and not just CCD data
PRO: Generally API is fairly stable and well tested
CON: People are starting to use it more often and it is getting some feedback. It may change in reaction to the format (although if in astropy, likely to get even more feedback)
CON: Likewise, might have to maintain two codes for backward stability
ISSUE: Will need to further improve the documentation
Other things to think about:
ccdproc is not a large dependency, and installation of it is relatively easy.
Having it in the LTS versions, makes it easier for people to know that it will be in there and won't need to update to a later version if not necessary.
With the code freeze for the next LTS version of astropy coming up soon, I think it would be a good time to discuss the thought of migrating two of the more useful features of
ccdproc
into the core of astropy, so that they are long term available in the code. So here are some thoughts on adding them as pro's and con's:Migrating CCDData:
Migrating ImageFileCollection
Other things to think about: