Fix some link issues #544
Fix some link issues #544
Conversation
|
hey @erik-whiting thank you so much for this!! The way this works is that pyosMeta (the package) parses our peer review issues and returns a packages.yml file. The workflow that calls this is below: When that happens, it updates the contributors.yml and the packages.yml file. So this pr will fix some links but it will then get overwritten in two weeks when the above workflow runs again. In the doi side of things, @banesullivan is working on normalizing all of our dois in pyosMeta to ensure that we have consistent returns of these values. then we can do less "cleaning" on the Jekyll side of things. For the partnership side of things, we haven't yet formally partnered with Sunpy, but we are with Astropy. Someone could check that they want to be considered for affiliation. Still, we can't assume the Astropy community accepts them unless we see the Astropy label on the issue, which an Astropy editor will vet. So, I think the fix for astropy is likely due to a change in how we parsed the partner label! if you are game, let's use the partners list key in the packages.yml file. If astropy is there (rather than a boolean), we can add the Astropy affiliation text to the card. this should be a straightforward Jekyll fix. We can merge your link changes, but please know they will be overwritten when the next cron job runs!! Please let me know if you have any questions, and THANK YOU so much for fixing this for us! Welcome to pyOpenSci !! |
| <li><a href="{{ apackage.archive }}" rel="permalink"><i class="fas fa-bookmark fa-fw"></i> JOSS Approved</a></li> | ||
| {% endif %} | ||
| {% if apackage.astropy %} | ||
| {% if apackage.partners contains "astropy" %} |
There was a problem hiding this comment.
@erik-whiting i'm sorry what i asked you to do is exactly what you did. let's go ahead and merge this knowing the other changes will get overwritten in the next contributor update.
2 participants
I am attempting to address #525, #526, and possibly pyOpenSci/pyosMeta#228.
Regarding #525, I am not entirely sure that the problem still exists? You say the
issue_linkisn't rendering, but the badge and text that says "View Review" is wired to the value ofissue_linkand seems to be linking just fine. Can you confirm that this is still a problem and clarify a little bit?Regarding #526, this seems to be happening because there are no
astropy: truelines in thepackages.ymlfile. Likewise, there are nosunpy: truelines in the file, despite the fact that you (@lwasser) added at least one such line in this PR: #207. I cannot figure out when that line was deleted, but I opted instead to check for the presence of "sunpy" or "astropy" in the list ofpartnersfor a package and use that as a boolean. I added sunpy as a community to the package from the above-linked PR.Regarding pyOpenSci/pyosMeta#228, I am a little confused on this. The links are not working because in the
packages.ymlfile, some of them are written out as markdown, but Jekyll is blindly pasting that markdown on top of the base url link. For example, currently, the "JOSS Approved" link for sourmash is pointing tohttps://www.pyopensci.org/[](https://doi.org/10.5281/zenodo.10951577), which I'm guessing is not what you want. You could write something like{{ apackage.joss | markdownify }}in thepackage-grid.htmlfile, but this method "smartly" parses the markdown and turns it into HTML, you would have to do some extra work to get it to keep the "JOSS Review" look it has now and still link to the appropriate URL. Instead, in this PR, I standardized thearchivelinks and pasted them if ajosslink was present. I did not standardize the joss links because the existing code does not use the.jossvalue of a package anyway, so I wanted to maintain the spirit of what was already there.I may have missed something or causes some tests to fail, I'm not sure, and I'm not totally familiar with the architecture of this site, so please let me know if I broke something or if there's something you want me to fix. As of right now, at least locally, things work and look how they're supposed to--at least what I think they're supposed to look like.