Repository navigation
WCS with lookuptables (tabular) #2362
Description
It appears that lookup table WCS cannot be created with astropy.wcs at present.
http://astropy.readthedocs.org/en/latest/api/astropy.wcs.Tabprm.html#astropy.wcs.Tabprm says a Tabprm table cannot be directly instantiated.
This is a feature request or a request for documentation if the functionality already exists. I'll help with implementation, but I need some direction. Does anything need to be modified in the C wrapper? Or is it just the python frontend that needs to be made?
Does astropy.wcs presently know how to write a tabular WCS, or do we need to implement that too?
Activity
astrofrog commentedon Apr 23, 2014
cc @mdboom @nden
mdboom commentedon May 9, 2014
Sorry to just see this now.
The support that is there is read-only, since it looked rather complicated to make an API that would make it easy to create these things, and there was no one clamoring for the feature at the time.
I recently spoke to Mark Calabretta about tabprm and he says it's a very rarely used feature.
I'm not against doing this if there's a good use case, though. It will require spelunking down into the C wrappers.
keflavich commentedon May 9, 2014
I don't have a pressing need for this, but it is one of the fundamental underlying systems that stopped
specutilsfrom going anywhere; the first thing that Wolfgang implemented was a lookuptable spectral axis. I think that makes a lot more sense attached to a WCS axis.If we want to support optical/IR spectrographs and IFUs, this is likely to be necessary, but as a radio astronomer I'm willing to put it off.
astrofrog commentedon Jun 5, 2014
I think there is a chicken-and-egg issue here - I think the reason no-one uses it is because it is hard to create them currently. If we allowed an easy way to create this in astropy, I think there would be wide applicability to spectra/spectral cubes especially.
mdboom commentedon Jun 5, 2014
I'd be happy (given the time) to collaborate with someone on implementing this. It shouldn't be too bad -- the actual calculation and saving to/from FITS is already there as part of wcslib -- the main work here would be providing the wrappers to convert Numpy arrays to/from the tables that wcslib wants.
astrofrog commentedon Jun 6, 2014
Thanks! It sounds like @keflavich mentioned he was interested in helping with the implementation? Maybe it would also be worth posting a message to astropy-dev to see if anyone else is interested in helping (though too many cooks, etc. etc.).
keflavich commentedon Jun 6, 2014
I was interested but time has gotten very thin... don't rely on me.
astrofrog commentedon Jun 7, 2014
Ok, then maybe we should see if someone on astropy-dev is interested.
tepickering commentedon Oct 23, 2014
i just bumped into a use case for this in my own work. has there been any movement on this?
cdeil commentedon Jun 29, 2015
In Gammapy having a way to construct a lookup table WCS from Python would also be very useful.
Currently we have to handle our spectral log energy axis ourselves, without using
astropy.wcs... see SpectralCube and LogEnergyAxis.parejkoj commentedon Oct 4, 2019
This is likely related to #7883 and I think is still an open issue? @nden ?
mcara commentedon Nov 21, 2019
In PR #9641 I implemented support for the
-TABalgorithm in FITS WCS defined in WCS Paper III.mcara commentedon Jan 16, 2020
To those (still) interested: #9641 has been merged and
WCS-TABshould now be supported inastropy.wcs. Please give it a try.pllim commentedon Jan 16, 2020
Closing as solved. Thanks, @mcara !