marx_rl_decon: new script to perform Richardson Lucy deconvolution ... #776
Conversation
kglotfelty
force-pushed
the
marx_rl_deconvolution
branch
from
ea26b60 to
4004914
Compare
kglotfelty
force-pushed
the
marx_rl_deconvolution
branch
from
4004914 to
d90a443
Compare
|
Do we think we want to wait until marx is released with this (so go out with it post the December release), or include it now? |
|
I realized that there is a normalization problem somewhere -- I think I know where but I haven't had time to go about fixing. [Basically I believe I need to filter/clip by the FOV file when combining multi-obi, but it has to happen at the right place.] I'm doubtful it'll be ready for 4.16.0 |
|
So, I think it's safest to not hurry for the 4.16.0 release then. |
Merge in the 4.16 code.
Unfortunately the API we need is not "standardised" so it is a bit annoying to support.
initial mamba/micromamba support for check_ciao_version
Space management
Remove the README_CIAO_scripts file
…es when DS9 preference file sets default region file format to CIAO
added `xpaset` region formatting specification
…eadme update tarball build script with the README_CIAO file dropped in 4.16.0
|
The marx source IMAGE also does not work as it is used here, that's not just a simple bug, it's designed to work in a different way (see discussion in Chandra-MARX/marx#50). I think what's needed is a new marx source ("IMAGE-WCS") which makes full use of the WCS information: Chandra-MARX/marx#57 |
- the contrib VERSION file is no-longer in contrib/ - we appear to only care about the main "CIAO" package and the "contrib" package; the other packages don't have an obvious version string - we shall have to wait for an update to sherpa (say) to see what happens here - the ciao-type file no-longer exists, so we ask the system for what is in use (now that we have an ARM native version is is easier to do)
Merge too small fix
centroid_map: fix issue with NaNs in input image
Update changelog for centroid_map
map2reg: combine regions using CXCRegion instead of dumping to ascii file
add changelog for cxcsds#851
merge_too_small add join parameter
…ing marx to perform the PSF convolutions
…was the trick; still seeing some artifacts, need to try something else
kglotfelty
force-pushed
the
marx_rl_deconvolution
branch
from
d90a443 to
d9ccfd4
Compare
…using marx to perform the PSF convolutions.
This needs the bug fix in https://github.com/kglotfelty/marx/tree/fix_offset_input_image to work.