Fix Oyr.osaltdiff discovery with yearly-only aggregation

[rbeucher]

Summary

This PR addresses #199 by allowing Oyr.osaltdiff to be generated from available monthly ACCESS output, while ensuring aggregation behavior only applies to yearly ocean requests.

What changed

• Added frequency-specific per-variable file pattern support in file discovery.
• Added an Oyr-specific override for osaltdiff in ACCESS-ESM1-6 mappings so Oyr can discover monthly salt_vdiffuse_impl files.
• Wired ocean driver/CMORiser resampling controls so ocean yearly tables auto-enable resampling.
• Ensured monthly ocean tables do not auto-enable resampling.
• Added unit tests for:
  • frequency-specific file pattern override behavior and fallback,
  • yearly ocean auto-resampling enablement,
  • monthly ocean non-auto-enable behavior.

Why this resolves #199

Issue #199 reports missing salt_vdiffuse_diff_cbt for Oyr.osaltdiff. In current outputs, salt_vdiffuse_impl is available monthly. This PR enables Oyr.osaltdiff to discover that monthly input and aggregate to yearly only when yearly output is requested.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 76.4%. Comparing base (bf7b1d6) to head (04f183a).
⚠️ Report is 1 commits behind head on main.

Additional details and impacted files

@@          Coverage Diff          @@
##            main    #441   +/-   ##
=====================================
  Coverage   76.4%   76.4%           
=====================================
  Files         31      31           
  Lines       5909    5913    +4     
  Branches    1092    1094    +2     
=====================================
+ Hits        4515    4519    +4     
  Misses      1154    1154           
  Partials     240     240           

Flag	Coverage Δ
unit	76.4% <100.0%> (+<0.1%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[rbeucher]

Scientific caveat — do not merge yet

After opening this PR, further investigation raised a concern about the correctness of the osaltdiff mapping.

Current mapping: salt_vdiffuse_impl (implicit vertical diffusion of Practical Salinity)
CMIP target: tendency_of_sea_water_salinity_expressed_as_salt_content_due_to_parameterized_dianeutral_mixing
Ideal source (per issue #199): salt_vdiffuse_diff_cbt — not present in current outputs

Why these may differ

osaltdiff is defined as a strictly dianeutral (across-density-surface) mixing tendency. salt_vdiffuse_impl is the implicit vertical (z-coordinate) diffusion term from the MOM5 solver. These are related but not equivalent:

• Dianeutral mixing is referenced to neutral/isopycnal surfaces; implicit vertical diffusion is referenced to model z-levels.
• The implicit diffusion term may include contributions from parameterisations that are not purely dianeutral.
• salt_vdiffuse_diff_cbt would isolate the diffusivity-driven (CBT) component that is the intended physical analogue.

Status

This PR provides the infrastructure (frequency-specific file discovery, yearly-only aggregation wiring) which is correct and reusable regardless of which source variable is ultimately used. However, the mapping itself (salt_vdiffuse_impl → osaltdiff) should be validated by an ocean scientist or replaced with salt_vdiffuse_diff_cbt once that diagnostic is added to ACCESS-ESM1.6 output.

Keeping this open for review/discussion but not ready to merge until the source variable choice is confirmed.