Astropy recently voted to partner with pyOpenSci – see the recently merged APE 22
As such we are ready to take the next steps towards reviewing packages in partnership with Astropy!
I think we are ready now to take the next steps towards formalizing a review. And the best thing to do would be to find a package to kick off the process and learn together.
Suggested next steps
I’m going to throw out some ideas on next steps but lets discuss and decide what makes the most sense - together!
I think on the astropy side of things we should have 1 to 2 editors from astropy join our editorial board. Yes normally editors start as “guest” editors. However i don’t think we need to worry too much about those labels - especially if an editor from astropy has been working on the astropy process already.
We have an Onboarding guide here that defines the onboarding process and an editor guide that clarifies the steps associated with leading a review
in our system an editor will serve for 1-2 years leading 3-4 reviews a year on average. but there is no specific time frame or requirement. we just don’t want anyone to get burnt out!
On the pyOpenSci side of things i have a TODO in place to update our website and continue work that i started on the astropy landing page !!
I think once we have atleast one editor on board we want to get a review going to test out the process and learn together.
Sounds good to me. Once we’ve updated our editors on the astropy side (~2 weeks), we’ll work with the package authors of packages that recently asked for a review to be astropy-affiliated and see if one of them volunteer to be the test package.
@hamogu welcome to discourse
that sounds perfect. 2 weeks will give me a bit of time to also revisit the website work we started last summer! i think having a package author + editor go through our process is the perfect next step.
please let me know if you need anything from me. AND if the author/maintainer has any questions they can also ask here as well.
Hi @lwasser, as I am going through the next steps for offering our package submissions to be promoted to the pyOpenSci review system, and also need to update the role description for our editors, some questions came up:
will it be our editors’ role to direct Astropy-affiliated applications (at least those that already have a submission through our portal) to either Pre-submission or Submission for Review in Issues · pyOpenSci/software-submission · GitHub
should we directly decide if the pre-submission step is still required?
Related, a follow-up issue lists the tasks of getting our own guidelines updated to (or replaced by?) the pyOpenSci ones.
Would this best be done as e.g. an envelope around your Review Guide, just adding the Astropy-specific criteria?
hi @dhomeier i have been traveling and am just catching up here! This is a good question.
will it be our editors’ role to direct Astropy-affiliated applications (at least those that already have a submission through our portal) to either Pre-submission or Submission for Review in Issues · pyOpenSci/software-submission · GitHub
I think it should be astropy’s role to redirect existing submissions. Given they already submitted to your community, I suspect it would be awkward if one of us suggested they redirect. it should come from the community. But with that said - if it is on github and you have a question or a place where we can help you in that effort, you can ping me as we begin to kick things off.
should we directly decide if the pre-submission step is still required?
We may need to see how this goes. pre-submissions are not a requirement at pyOpenSci. However, when there is question about a package’s scope, an author can chose go that route. This is our scope page that discusses our scope criteria. Normally if a package is out of scope and it’s a full submission our editor in chief makes that call. And if it’s a hard call they may bring it up in our editorial channel (which astropy editors should also be a part of and check). I would imagine that a package that could be affiliated with astropy would most often be in scope for us too as the “Relates to science” component is really what is most important. We’ve had one package that had a very very complex maintenance structure that did not get reviewed because of the extreme technical complexity). As such i’d just let your existing packages submit to us via a full submission and we can evaluate if we get many that have questionable scope as we go.
i opened this pr in the summer to address this exact question! it’s been open for so long i just need to rebase it and clean it up! i’ll do that this week. Can you and potentially Erik and anyone else who needs to approve, review? i had forgotten i started that already.
tangentially related i have a website page pr as well for astropy that will need approval from your community before i’m comfortable with merging it!
So my short answer is i think you can cross-link to our guide book pages in your content. we definitely want to add content on astropy to help promote your work as much as we can on the pyOpenSci side of things.
once that is merged i can then update our software submission template on github with a link to those requirements similar to what we have for pangeo.
The website page review can move a bit more slowly only because we don’t have any astropy packages to list there (yet) but we do have sunpy. so we want to work on that as well but the peer review guide review and merge is my next priority that i need your help with.
