[BUG][ENH] Hashing of ESM Datastores to verify that datastores haven't been updated

[charles-turner-1]

Despite our attempts to ensure that datastores are not updated/changed or broken (see #308), we cannot enforce that datastore owners do not make changes to their datastores, and we are not necessarily notified if changes are made. At date of writing (10/02/25), e2e tests are failing due to changes in the cmip5-al33 datastore made yesterday.

• Note: we are only aware of this due to the e2e tests failing. Other datastore, eg. cmip6-oi10, are not included in the end to end tests, and so are liable to change without us noticing.

This represents a reproducibility as well as a package stability issue. Hashing to verify status of esm-datastores was introduced in #317, and would be a way to alert us to potential datastore changes easily (rather than losing an entire morning to bug hunting).

I'm going to add this as a sub-issue to #308.