controller

smif: The controller coordinates the model_runner, data-layer and decision_layer

The controller coordinates the model_runner, data-layer and decision_layer.

Stories

Add smif prepare command to cli to convert data to use binary filestore

Labels:

• controller
• smif

smif prepare --from=csv --to=binary <model_run> imports data provided using the csv data store and writes out using the binary data store

This will allow 'pure' implementations of the DataStore classes to retrieve data for a model run.

CLI run_server should use configured Store

Labels:

• controller
• smif

Currently hard-coded to DatafileInterface - see _run_server in cli/__init__.py

Archiving a ModelRun packages input data and configuration with results from the run.

Labels:

• controller
• smif

A ModelRun relates to a SosModel, SectorModels, scenario data, strategy and decision data and narrative parameter data.

See also #142125185

To recreate a model run, we need to persist all input data, model versions and the smif version.

An archiving feature allows the input data to be packaged with the model configuration and results, for verification and replication.

Choose data layer

Labels:

• controller
• smif

When configuring the system, choose whether to use yaml files for data input/output or the database api