🧰 Task
We currently have multiple failures of docs build in the middle of a long run because of the inaccessibility of some remote resources.
As far as I checked, all downloads are done in a cacheable way (pooch or other lib like this).
So we could add a prefetch steep befor start building html.
Advantages:
- Early fail on inaccessible resources, no more parsing of long logs
- Better ability to cache it.
- If data are too big on cache, we could put them on our server, and prepopulate before tests with source codes pointing to original data.
Disadvantages
- Next list to keep in sychronisation
🧰 Task
We currently have multiple failures of docs build in the middle of a long run because of the inaccessibility of some remote resources.
As far as I checked, all downloads are done in a cacheable way (pooch or other lib like this).
So we could add a prefetch steep befor start building html.
Advantages:
Disadvantages