Feature request: Is it possible to access the raw data in an FPACK-compressed HDU?

[dstndstn]

Apologies if I am just missing something, but here's my situation:

• I am reading FPACK-compressed (lossy) images
• I want to compute the quantization error -- specifically, how large a value that is supposed to be zero can be.
• FPACK is implemented by replacing the image HDU by a BINTABLE HDU, where the numbers I need are columns in the table (ZSCALE, ZZERO).
• fitsio, very reasonably, transparently decompresses the data and presents it as an IMAGE HDU
• however, this means I can't get at the underlying TABLE representation to retrieve the ZSCALE and ZZERO columns from the table!

In astropy.io.fits, they provide an hdu.compressed_data field that gives you access to the underlying BINTABLE. It would be awesome if fitsio could provide something similar (eg, fitsio.read(filename, ext=None, decompress=True))!

[dstndstn]

PS, an alternative API / great feature would be a call that retrieved a quantization error map!

[esheldon]

I imagine there is a way to do this from cfitsio. I'm using the "standard" routines for querying the fits file, maybe it can be queried differently.

But note this information would not generally be visible the way the code works now. I think it would make most sense to have a different kind of opening mode.

[beckermr]

Years ago I hit a bug where I read a compressed image as a table and was super confused. This must have been fixed by now since I cannot reproduce it.