Repository navigation
Add support for OGIP unit strings - #377
Conversation
ghost
assigned
|
@mdboom - does it make sense to delay this for 0.4.0? |
|
Yeah -- it's really just here for completeness. As in, "this is a unit format that's out there, we should support it". But no one has actually asked for it and I have no idea how widely used it is. (The FITS and CDS unit formats have had a number of bug reports, though, so I know someone is using them). |
|
Following the discussion in #1911, this adds support for OGIP format units. How we actually use this in the FITS parser (how we automatically determine whether to use this or the FITS unit parser, or somehow try both and check for agreement as @hamogu suggests) should probably be left as a follow-on PR. This merely makes OGIP parsing available to us. Cc: @hamogu, @philhodge |
|
This is ready for review. @embray? |
There was a problem hiding this comment.
One day I'm gonna make a formatter metaclass to handle this sort of boilerplate. But in the meantime it's really not necessary :)
|
Looks fine to me. How this integrates with the FITS module depends entirely on how unit support is added in the first place (something I'd really like to get done for 0.4 though). I think it should try the standard FITS unit format first, and failing that fall back on trying OGIP. There should also be an option to disable OGIP support altogether for applications that know they don't need it and don't want the extra overhead. I might apply that to unit support in general as well. But the default should be the most permissive. |
Add support for OGIP unit strings
Remove unused astropy_helpers.compat
Add "Interim finance committee role"
The OGIP unit standard has some subtle differences to the core FITS unit standard. We should implement a new formatter for this standard.