Model Execution

ravenpy provides two mechanisms to execute a model configuration. The first one described is the ravenpy.run function, which expects to find a directory with RV files and returns the path to the output directory. The second one is the ravenpy.Emulator class which exposes more functionalities and deserves some explanations.

The Emulator class

The Emulator class minimally expects a fully configured Config instance, which it immediately writes to disk. Other optional arguments are workdir, the path to the work directory where configuration files are written, and modelname, the string used to create configuration file names. By default, workdir will be set to a temporary directory, so make sure to specify another location if you want to hold on to the results.

When the run method of an Emulator instance is called, it launches the model which writes simulations results in the output folder, then returns an instance of the OutputReader class. Beyond the write_rv and run methods, Emulator has a number of read-only properties:

• config: the model configuration Config instance;

• workdir: the path to the work directory where configuration files are written;

• output_path: the path to simulation results;

• modelname: the name given to configuration files written to disk;

• output: the OutputReader instance created out of the simulation results.

The Emulator has another method called resume, which returns a new Config instance where the initial states for HRUs and sub-basins are set from the solution.rvc file storing the states at the end of the run. Note that by default, the StartDate of this new configuration will be set to the date following the end of the run. Set timestamp=False to ignore the time stamp information of the solution file.