Include parameters in recipe to let users choose which data to save/plot
This was requested in #27
Opening this issue to keep track of the potentials ways we can handle this.
TODOs (Saving):
-
Anomalies -
Calibration -
Skill -
Probabilities -
Downscaling -
Visualization -
Change unit tests -
Update recipes
TODOs (Visualization):
-
Update recipes
Other tasks:
-
Update recipe checker
Saving Option 1: Including parameters inside the configuration of every module
Workflow:
Anomalies:
compute: yes
cross_validation: yes
Calibration:
method: raw
save_experiment: yes/no # saves hcst/fcst
save_observations: yes/no # saves the corresponding observations
Skill:
metric: RPSS CRPSS EnsCorr
save_metrics: yes/no # save the skill metrics
Probabilities:
percentiles: [[1/3, 2/3], [1/10, 9/10]]
save_probabilities: yes/no
save_percentiles: yes/no
# ... etc
Potential issues: ensuring flexibility and consistency if Anomalies are computed, given that the user has the option to compute them before or after the Calibration step.
Saving Option 2: Adding a Saving section inside 'Workflow':
Workflow:
# (...)
Saving:
outputs: skill # possibilities: skill, experiment, observations, probabilities, percentiles
## Add possible options for significance?
# ... etc
Potential issues: Not as user-friendly and clear as Option 1
Visualization Option 1: Adding a Visualization section inside 'Workflow':
Workflow:
# (...)
Visualization:
plots: most_likely_terciles, skill # possibilities: skill, forecast_ensemble_mean, most_likely_terciles
## Add possible options for significance?
# ... etc