Delete get_ensemble_by_name from LocalStorage

[dafeda]

Ensemble names are only unique within a single experiment.

Breaking Change

• Will cause semeio to fail, so this PR will need to be merged: Dont use get_ensemble_by_name from LocalStorage semeio#640
• Will cause `everest to fail

[codecov-commenter]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 90.85%. Comparing base (e33136f) to head (17fe3b4).

Additional details and impacted files

@@           Coverage Diff           @@
##             main    #8623   +/-   ##
=======================================
  Coverage   90.84%   90.85%           
=======================================
  Files         339      339           
  Lines       20897    20894    -3     
=======================================
- Hits        18984    18983    -1     
+ Misses       1913     1911    -2     

Flag	Coverage Δ
gui-tests	75.79% <100.00%> (+0.03%)	⬆️
integration-tests	53.80% <100.00%> (+0.03%)	⬆️
unit-tests	68.66% <100.00%> (+0.01%)	⬆️

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

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

[DanSava]

Are experiment names unique?

[dafeda]

Are experiment names unique?

Yes they are. One reason why want this is that it makes it easier to read plots since we use experiment_name:ensemble_name.

[yngve-sk]

Should this be made unique or is it ok for cases to be migrated (from blockfs in this case so maybe no that many) w/ the same experiment name? Logging statement above is
logger.info(f"Migrating case '{path.name}'")

[dafeda]

Valid point, but this not change behavior as the current code does this when no name is specified:

name = datetime.today().strftime("%Y-%m-%d")

Will leave as is for now. OK?

[DanSava]

Are experiment names unique?

Yes they are. One reason why want this is that it makes it easier to read plots since we use experiment_name:ensemble_name.

I could not find any logic in ERT that enforces the experiment name to be unique, but it might be that I did not look hard enough.

If the experiment name is unique and this is checked when a new experiment is created then to make Everest work with these changes we would need to update:

• ert/src/ert/simulator/batch_simulator.py

experiment = storage.create_experiment(
    parameters=self.ert_config.ensemble_config.parameter_configuration,
    responses=self.ert_config.ensemble_config.response_configuration,
)

to

experiment = storage.create_experiment(
    parameters=self.ert_config.ensemble_config.parameter_configuration,
    responses=self.ert_config.ensemble_config.response_configuration,
    name=f"experiment_{case_name}"
)

[yngve-sk]

Recommend double checking wrt uniqueness of experiment name in some edge cases, otherwise LGTM

[dafeda]

Are experiment names unique?

Yes they are. One reason why want this is that it makes it easier to read plots since we use experiment_name:ensemble_name.

I could not find any logic in ERT that enforces the experiment name to be unique, but it might be that I did not look hard enough.

If the experiment name is unique and this is checked when a new experiment is created then to make Everest work with these changes we would need to update:

• ert/src/ert/simulator/batch_simulator.py

experiment = storage.create_experiment(
    parameters=self.ert_config.ensemble_config.parameter_configuration,
    responses=self.ert_config.ensemble_config.response_configuration,
)

to

experiment = storage.create_experiment(
    parameters=self.ert_config.ensemble_config.parameter_configuration,
    responses=self.ert_config.ensemble_config.response_configuration,
    name=f"experiment_{case_name}"
)

You are right I think, the validation is only in the front-end. @larsevj is working on making sure it's not possible to create experiments with non-unique names I think.