search

kevin218 / Eureka

Eureka! is a data reduction and analysis pipeline intended for time-series observations with JWST.
https://eurekadocs.readthedocs.io/
MIT License
58 stars 45 forks source link

Minor bug fixes #558

Closed taylorbell57 closed 1 year ago

taylorbell57 commented 1 year ago

This PR has some minor bug fixes. Specifically, it fixes some small bugs with S3 photometry code, and it fixed a bug where setting fill_value to 'boxcar' in Stage 4 still resulted in points being masked while they should've been left unmasked and simply replaced with the value from the boxcar smoothed data.

taylorbell57 commented 1 year ago
taylorbell57 commented 1 year ago

Alright, @kevin218 hopefully that latest push will do the job of fixing the GitHub pytest testing (we'll see shortly). I've tested installing the code four different ways on my Intel Mac (pip with/without pymc3, conda environment yml files with/without pymc3), and they all worked well.

taylorbell57 commented 1 year ago

Sigh, well I removed some of GitHub's issues, but now we're back to some gcc error that I don't get locally (and I do remember seeing with some recent PRs until the new numpy blas errors starting popping up even more recently). I'll try to read through the error messages closely and see if I can figure out what the issue is

kevin218 commented 1 year ago

I see that Pytest Testing fails with a new error, possibly relating to gcc.

taylorbell57 commented 1 year ago

Yeah, I'm actively working on that gcc issue - I'll re-request your review once I've figured out how to get gcc working again (my best guess is that a gcc version number changed and broke GitHub's testing)

taylorbell57 commented 1 year ago

FYI, I had no luck with changing the gcc version, so I reverted those commits and my next best guess is that a cython version is to blame. I'll have to come back to that later (perhaps in a few days). In the meantime, you could approve this PR if you want to and leave the resolution to the GitHub-exclusive gcc error for another PR