fixes for #165, #238, #210, #205

spacetelescope / imexam

imexam is a python tool for simple image examination, and plotting, with similar functionality to IRAF's imexamine

http://imexam.readthedocs.io

BSD 3-Clause "New" or "Revised" License

74 stars 45 forks source link

fixes for #165, #238, #210, #205 #241

Open sosey opened 2 years ago

sosey commented 2 years ago

This PR attempts to fix issues with the 1D gaussian fits used in radial plots and line plots. As such it addresses some common issues in the following:

fixes #165 closes #238 closes #210 closes #205

I still need to clean up and there's likely some more minor changes to make before this PR is merged, but FWHM looks more appropriate.

@janerigby checkout this branch and let me know if you see improvement

sosey commented 2 years ago

@larrybradley will you look at the changes I made for the modeling and make sure I'm not doing something stupid?

larrybradley commented 2 years ago

@sosey The Gaussian model (red line) in your figure above looks (by eye) to have a mean/center value of r > 0 (it appears to be sloping downward to the left at r=0). It should be fixed to have a mean/center at exactly r=0.

janerigby commented 2 years ago

Is this where I admit my ignorance, that I don't know how to check out a branch that addresses an issue?

alanwatsonforster commented 2 years ago

I checked out the branch, and it still doesn’t seem to work as it should. As commented above, the Gaussian is not centered on zero. imexam copy

prdurrell commented 2 years ago

Hi - I wanted to check on updates to imexam to get correct radial profile FWHM's from 'r' command. How does one update from 0.9.1 to test?

prdurrell commented 2 years ago

Hi -- I wanted to check back on this ticket -- I see it has been combined with other tickets, but I am still having trouble with imexam. I am now running imexam 0.9.1, and I am still getting erroneous results using 'r'. Using 'j' or 'k'(1D profiles along X, Y axes) seem to work fine. See the following -- this is for a star on a drizzled ACS/WFC image, so the FWHM should be a little over 2. I have no idea what 'r' is fitting to. I have included screenshots of the same object using 'r', 'j' and 'k' for reference.

I am not sure if the centering is off, or what have you.  But the amplitude in rimexam is far above what is suggested in the other subroutines, as is the more obvious FWHM difference.

Pat Durrell