[26.1] Add namespace stripping to CachedExplicitSingularityContainerResolver

[mvdbeek]

Add an optional namespace config parameter that strips docker://quay.io/{namespace}/ from image identifiers when constructing the local cache path, producing flat filenames compatible with CVMFS-hosted biocontainer caches (e.g. /cvmfs/singularity.galaxyproject.org/all/).

Mirrors the same stripping already done in pull_mulled_singularity_command.

How to test the changes?

(Select all options that apply)

• I've included appropriate automated tests.
• This is a refactoring of components with existing test coverage.
• Instructions for manual testing are as follows:
  1. [add testing steps and prerequisites here if you didn't write automated tests covering all your changes]

License

• I agree to license these and all my past contributions to the core galaxy codebase under the MIT license.

[natefoo]

Will this not break existing caches?

[mvdbeek]

The core of the logic is

if self.namespace:
    prefix = f"docker://quay.io/{self.namespace}/"
    if image_id.startswith(prefix):
        image_id = image_id[len(prefix):]

namespace is an option you can configure (so you could put non-biocontainers on a different path).
It should not have an effect on any other caches and even the stripping is not enforced.

[natefoo]

Ah yeah, no existing caches would have namespace configured.

[natefoo]

So this also doesn't use configurable cachers (i.e. CacheDirectory from mulled) meaning there's going to be a dir listing every time - this is prohibitively expensive, hence why the dir_mtime cacher exists. Could we refactor and use it here? Otherwise I don't think I can configure this in production.

FWIW I already did that in #20447.

[natefoo]

As an aside, very hard to tell what changes with each iteration. For someone watching and reviewing it'd be easier if Claude pushed new commits and then just squashed at the end when we're happy.

[github-actions]

This PR was merged without a "kind/" label, please correct.