Add dea_versioned naming conventions

[jeremyh]

Adds options conventions="dea_versioned" and conventions="dea_s2_versioned".

It would be nice to combine them into one naming convention, but that would require a bigger refactor of eodatasets: they could no longer be static, as the conventions would have to change slightly depending on satellite.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 84.76%. Comparing base (8eca8e4) to head (c5e1bce).

Additional details and impacted files

@@             Coverage Diff             @@
##           develop     #490      +/-   ##
===========================================
+ Coverage    84.73%   84.76%   +0.03%     
===========================================
  Files           26       26              
  Lines         4527     4536       +9     
===========================================
+ Hits          3836     3845       +9     
  Misses         691      691              

☔ View full report in Codecov by Sentry.
📢 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.

[jeremyh]

@robbibt I'm currently calling them dea_versioned and dea_s2_versioned but other suggestions are welcome :)

[sixy6e]

It would be nice to combine them into one naming convention, but that would require a bigger refactor of eodatasets: they could no longer be static, as the conventions would have to change slightly depending on satellite.

Yeah, it would be nice, and the only reason S2 is different is to account for the insertion of the datatake sensing time that already exists in GAs data products.

In a future state, consolidation would be ideal.

But for now, as it's explicit for the S2 case, I'd suggest the same for Landsat, i.e dea_ls_versioned. I'd recommend dropping the a in dea though. How much is there in the existing codebase that refers to dea?

[sixy6e]

Is there any benefit to being explicit with a DEALandsatVersionedNamingConventions as what is done for DEAS2VersionedNamingConventions? Or simply that this would be generic that items other and Landsat could use it?

[jeremyh]

Yes, in theory they were designed to be multi-sensor originally, it's just s2 that broke it. They're still intended to be the general defaults when a platform doesn't need its own unique conventions.

It's also for consistency with the other naming conventions in the codebase. We could change them to "de" as mentioned above, though it might sound odd for there to only be one set called "de"