[Bug]: Update CRPIX to STARCEN in calculate_centers

[JarronL]

Check Existing Issues

• Yes, I have checked existing issues to ensure this problem hasn't already been reported.

Instrument or Category

Image Tools

Description

CRPIX should inherently be the position of the star in order for the WCS to work correctly. I believe this should be handled in the calculate_centers() function. I believe the CRPIX values should 1-indexed (e.g., DS9 convention) rather than 0-indexed (numpy convention).

Error traceback output

No response

What operating system are you using?

No response

What version of Python are you running?

No response

What Python packages do you have installed?

No response

Additional context or information

No response

[wbalmer]

Would any of this be changed / updated by any of what @strampelligiovanni or @semaphoreP just did for pyklip v2.8.4? If so, we can close. If not, I will try to hack a fix.

[semaphoreP]

Even though pyklip now ignores CRPIX, I think we should still update CRPIX so that spaceklip products are consistent in e.g., ds9.

[JarronL]

@wbalmer No, this has been a persistent issue with regards to proper WCS implemenation in the headers. Something that needs to occur earlier in the pipeline processing once the star location is determined. This was discussed with Aarynn, Giovanni, and Jason last week and Aarynn asked me to make the bug report so we could implement. I can sort it out in the next day or two.

[AarynnCarter]

@JarronL Just looping back on this, did you happen to put any code tweaks together?

[JarronL]

Yes, but as a larger series of improvements to the alignment procedure. However, I can push a quick fix that addresses just this issues more immediately.