search

scipy-conference / scipy_proceedings

Tools used to generate the SciPy conference proceedings
Other
228 stars 530 forks source link

DOI Checks Failing on Papers #935

Closed rowanc1 closed 3 months ago

rowanc1 commented 4 months ago

On some articles, the DOI checks that we are running for the authors are failing. We are still investigating and haven’t come up with a solution to the DOIs failing checks.

DOI had a service outage a while back that had them change their cloudflare config: https://doi.statuspage.io/incidents/f4hrmd88c94l

I believe that some programatic access (e.g. a basic request without user-agents coming from github actions) to doi.org has been cut off, resulting in a 403 Forbidden when we ask if the DOI exists. We will add a fallback but that is going to take a bit of time. At this stage, some of the failing checks on scipy papers are because of this. I am hoping we will fix this in the coming few days.

cc @fwkoch @stevejpurves

rowanc1 commented 4 months ago

This DOI failing checks are currently impacting #925, #918.

cc @TahiriNadia, @ameyxd, @hongsupshin

ameyxd commented 4 months ago

Also impacted: #931

fwkoch commented 4 months ago

I have updated the checks to no longer fail when doi.org gives a bad response. Some of the PRs listed above are still failing for unrelated reasons.

For papers where all citations are defined in a .bib file, this will be enough - the checks will pass and the build will be fine. Only in cases where citations are built from DOIs referenced directly in the text will the build possibly have errors (there examples of this in the MyST paper). I'm continuing to work on a fix for those cases.

rowanc1 commented 3 months ago

See some updates here:

fwkoch commented 3 months ago

Another fix here: https://github.com/executablebooks/mystmd/pull/1293

There are no longer any more DOI-related check failures 🥳

ameyxd commented 3 months ago

Good job folks! @rowanc1 @fwkoch 👏🏽