Add support data caching to CI

[ReeceHumphreys]

• Review: By commit
• Merge strategy: Merge (no squash)

Description

This PR adds a new environment variable that allows for overriding the default Pooch cache location. We can now use this in CI to set a persistent cache for the support data files so they do not need to be fetched every time. This cache has been set with a biweekly expiration policy to avoid the data from becoming too stale.

Verification

CI

Documentation

Add a section explaining how to set the env var.

Future work

N/A

[schaubh]

Thanks for adding this. This will hopefully ease file server access issues.

[schaubh]

On your branch caching the data downloads, Codex tells me:

In each workflow, ./.github/actions/data-cache is executed after ./.github/actions/build (pull-request.yml, pull-request.yml, pull-request.yml, pull-request.yml, merge.yml, canary.yml).
But the build action itself runs bskLargeData during build (build/action.yml).
Because BSK_SUPPORT_DATA_CACHE is only set inside data-cache (data-cache/action.yml), the initial download happens in the default cache path, not .bsk-cache. Result: the new cache is bypassed for the expensive download path and can even cause duplicate downloads later in the same job.

Can you comment on this? It might be on to something.

[ReeceHumphreys]

Can you comment on this? It might be on to something.

Fixed as suggested, Codex was correct and it should have been moved a few lines earlier!