Facilitated visualisation of Uncertainty in Model Ensemble outputs using plots.
For testing try (while in "fumeplot" directory):
python3 PlotNamedStocksByTimestep.py
Uses default, which is "flee". You can also use "facs" or "homecoming" as an argument:
python3 PlotNamedStocksByTimestep.py flee
python3 PlotNamedStocksByTimestep.py facs
python3 PlotNamedStocksByTimestep.py homecoming
This will create plots of quantities evolution with their stadard deviation uncertainties, individual runs trajectories, and histogram animation for each data denomination, as well as a violin plot animation for combined data.
To create a single bar plot with uncertainties for all data denominations run:
python PlotNamedSingleByStep.by flee
FUMEplot acts as an independent plugin to FabSim3, enabling output plotting and analysis for a wide range of its supported codes. To install it, simply run:
fabsim localhost install_plugin:FUMEplot
Next, to use it, run any workflow, and fetch its results to the local machine using
fabsim <machine> fetch_results
At this point, FUMEplot can be invoked as follows for ensembles:
fabsim localhost fplot:<name_of_results_subdirectory>
FUMEplot will automatically detect the underlying solver, and construct the visualisations specified for that specific code.
Because replicated runs have a slightly different directory structure, FUMEplot should be used there as follows:
fabsim localhost fplot:<base_name_of_replica_run_subdirectory>
For example, if your replicas have names ranging from flee_archer2_24_replica_1 to flee_archer2_24_replica_9, then the base_name_of_replica_run_subdirectory should be set to flee_archer2_24_replica_.