After some internal discussion, we have decided to move the documentation of Dask-CUDA from Read the Docs to RAPIDS Docs. This would make it relatively easy to set up a docs CI workflow script like this, and is motivated by the fact that there are a lot of barriers keeping this repo from migrating to Dask org.
What's a clean way to do this without causing confusion? Should we keep both sites up simultaneously for a period of time with some banner notifying the migration? Or have the RTD docs redirect straight to the RAPIDS docs?
cc @quasiben @pentschev
Opened #580 to make a gpuCI docs build script.
Redirecting seems like a simpler way to deal with it.
I agree, just wasn't sure if RTD had the ability to easily manage a redirect but looks like it can.
This issue has been labeled inactive-30d due to no recent activity in the past 30 days. Please close this issue if no further response or action is needed. Otherwise, please respond with a comment indicating any updates or changes to the original issue and/or confirm this issue still needs to be addressed. This issue will be labeled inactive-90d if there is no activity in the next 60 days.
Dask-CUDA's docs just got published to RAPIDS docs:
Now we would just want to figure out redirecting the RTD version of the docs.
In conda-forge we did this like so
Great - thanks for that example! Do you know if we would need to keep that rtd/ directory in our repo if we disable webhooks on RTD's end after the last build?
Peter was mentioning redirecting, which agree probably makes sense (there may be other links to RTD out there).
So the rtd directory is important for supplying a documentation build that creates a webpage on RTD to do the redirecting. This is set in the RTD config file. We wouldn't want that in our usual docs hence the separate directory.
With #615 merged, looks like the redirect is working and the docs migration is complete 馃槃 thanks for the help @jakirkham @pentschev!
Most helpful comment
With #615 merged, looks like the redirect is working and the docs migration is complete 馃槃 thanks for the help @jakirkham @pentschev!