Closed matijapretnar closed 2 months ago
Have you used a script to generate the entries? If so, then it may be useful to integrate it into the CI as a sort of "format checker" for future PRs.
Indeed I had, but there were a bunch of one-off scripts that are slowly converging. When I'm done, I can make it available, though I do not want to be too strict with CI so to not discourage submissions.
Also, the script uses a local JSON file as a cache to avoid hitting DBLP with too many requests. Can this be persisted through PRs?
Indeed I had, but there were a bunch of one-off scripts that are slowly converging. When I'm done, I can make it available, though I do not want to be too strict with CI so to not discourage submissions.
Also, the script uses a local JSON file as a cache to avoid hitting DBLP with too many requests. Can this be persisted through PRs?
I don't think we should be strict either. Such a script should simply check our existing style guidelines programmatically, it would be convenient. And yes, we can persist files by having the CI upload an artifact after a successful run.
In case anyone wants to play around with GitHub actions, here is the final script I've used to update the bibliography: https://gist.github.com/matijapretnar/926db074d96b51b8aa079f66cc785e11. I did the arXiv changes manually, so that is not in yet.
If you run it, you get the updated README.md
. The uncommited changes in author names are all regressions, so perhaps it does not even make sense to rely on DBLP for CI. Perhaps you can run it every now and then to do a clean-up.
Thanks! I will experiment with it at a later time.
I've now gone through all the papers in the list and updated all of those that exist in DBLP (add DOI, BibTeX, fix typos, use official title capitalisation). I think we can consider #8 closed.