Proposed affiliated package: PetroFit (pyOpenSci/software-submission#159)

[robelgeda]

EDIT: Superseded by pyOpenSci/software-submission#159

PetroFit is a Python package designed for calculating Petrosian properties, such as radii and concentration indices, as well as fitting galaxy light profiles. In particular, PetroFit includes tools for performing accurate photometry, segmentations, Petrosian profiling, and Sérsic fitting. PetroFit is based on PhotUtils and AstroPy; and we hope to contribute the more generalized aspects of our code upstream for other packages make use of. For more information please visit the PetroFit repo or documentation.

cc @crawfordsm

[dhomeier]

Thank you for your submission and apologies for the long delay in responding to it!
As outlined below, we have been in the middle of migrating our Affiliated Package review process to a new system in partnership with pyOpenSci now described in APE 22, and experienced some unexpected delays in that transition.

What this means for all packages that have applied for Affiliated status before APE 22 was passed is, that we still offer you to continue the package review in the previous mode, but also want to give you the option (and encourage you) to apply through the pyOpenSci page instead.
Some key differences this would bring about are

• Double open review instead of single anonymous
• additional listing as a pyOpenSci package
• automatic submission for publication in JOSS
• review under the pyOpenSci guidelines in addition to Astropy-specific requirements such as integration into the Astropy ecosystem
• submission and review in the pyOpenSci submission space

Please comment here on your preferences, or if you need additional information on this process.
Thanks!

[robelgeda]

@dhomeie Thank you for reaching out, we are a go for the application and have submitted our review request here:
pyOpenSci/software-submission#159

Please let us know here if that request is okay or if we should take any additional steps.

[dhomeier]

Thank you for the renewed submission @robelgeda! I am confident this can proceed through the full submission stage right away from the initial review of your package. I hope to be able to proceed with the search for reviewers soon, once my editor position is confirmed on the pyOS side.

[robelgeda]

Hi @dhomeier , I just wanted to tag you here to ping this PR, not sure if this is still needed. If not, please feel free to close.

[hamogu]

That is correct. Once the pyopenSci review has gone through (which has been a lot slower than we hoped it would be because the number of package submitted is a lot more than we expected them to be - you can argue if that is a good or a bad thing), any package tagged as "astropy" will automatically appear on the astropy-affiliated listing page a few days later.

Thanks for your patience.

[pllim]

For completeness, at the time of writing, there are 4 packages accepted via the new process at https://www.astropy.org/affiliated/#affiliated-package-list . In the future, more will be there, including this one. Thanks, all!

[dhomeier]

Does this take some days in general? I was not sure if some step was still missing to get the Astropy badge added, since the package is already listed on https://www.pyopensci.org/python-packages.html , but then I assume everything will take its due course. I've added the Zenodo links on pyOpenSci/software-submission#159 , that should complete all tasks over there.

[hamogu]

I think the pyOpenSci script that builds the yaml which then in turn is read by the astropy website is run daily or weekly (I just had a PR to change the frequency of the cron job as PyOpenSci, but I forget not what we settled on) and any pyopenSci editor can manually restart the GHA that they use (if you manage to find the right one). However, the scripts are a bit brittle. So, if it doesn't appear in a few days, I would suggest @dhomeier (as the editor) to mention that in the #software-review slack channel of pyopenSci and then someone will look into what broke (e.g. a formatting error in this that the script tries to parse).

[pllim]

Hmm. In particular, Astropy website is trying to mirror https://www.pyopensci.org/communities/astropy.html and I do not see PetroFit there yet.

[pllim]

I see it under https://github.com/pyOpenSci/software-submission/issues?q=is%3Aissue+is%3Aopen+label%3Aastropy . Is it because it is not "accepted" (merged) yet? Maybe @lwasser could help clarify the matter. Because I also see astrodata not merged but it is considered "accepted". 🤷‍♀️

[dhomeier]

Acceptance is manifest by moving to the appropriate category in peer-review-status , which @lwasser did last week. We generally only close the issues after all the post-review cleanup, like blog post etc. is finished.

[pllim]

Huh, weird, I see PetroFit in https://raw.githubusercontent.com/pyOpenSci/pyopensci.github.io/main/_data/packages.yml (which is the file that our site parses) but no "astropy" under "partners" (which is how we know it is affiliated to us).

[pllim]

If you want the nitty gritty details, the JS that do all the parsing starts here:

astropy.github.com/affiliated/index.html

Line 397 in ca8c78c

<script>

[dhomeier]

Hmm, I noticed an error in one of the merge jobs, but that evidently has both labels.
But it is clearly missing on the package listing, so there is nothing our script could do to repair that (claiming without knowing any JS).

[pllim]

In a nutshell, whatever that generates https://raw.githubusercontent.com/pyOpenSci/pyopensci.github.io/main/_data/packages.yml has to be fixed on the pyOpenSci infrastructure. Because that affects both their Astropy listing and our own listing. It is not something we should patch downstream because that would lead to inconsistency.

[pllim]

I will ping them...

[lwasser]

It's fixed on our website, but I still do propose that rather than using the partners list, we use labels. i think that is safer from an editorial point of view, to use. Any submission could check the Astropy box to be considered. But if one of the astropy editors accepts the package as affiliated, they can add (or remove) a label to declare that - no one else can except an editor!

[pllim]

Is label on the yaml now? Can't check on phone right now. We only use basic JS so I'd rather not have to query GH issues on top of the YAML. Thanks for the fix!

[dhomeier]

Yes, it has both labels in the YAML file now – I thought that was what we were using. And it is already on our page as well.
The Astropy label has been set in the issue for over a year anyway.

[lwasser]

Yes!! so no rush on this friends - but i opened this pr for our website to pull astropy from the labels list in the yaml rather than the partners list.

If you want to consider doing the same that would be good for consistency. Then, in this case, @dhomeier you have superpowers to add astropy as a label! so if for some reason, a package is no longer affiliated or doesn't meet the criteria (but the do check the partner box), you can always add or remove them from being "affiliated".

I'm glad we found this, even if it causes a bit of a problem with the timing of Petrofit being cross-listed (sorry about that!)!

[pllim]

I see. Yes, I can see if I can fix the JS when I get a chance. Thanks for the help! #656