1) In sbat.py, the plot functions are decentrally called within the class methods when self.config.plot_results == True
2) The plotting functions pollute the submodules, making them less readable and harder to debug
3) When output == False, plotting fails because there is no output directory defined. However, only plotting when output == True is problematic as sometimes results are not computed because of the tight coupling between plotting and the submodules
At the moment I suggest the following solution, but that can be discussed:
1) We relocate the plotting functions and place them into a separate file, could be for example sbat/plot/plot.py.
2) We create a public class method into the model that is called by the user, that is executed at the end of sbat.main. This method then checks the configuration for present results (as in "Has baseflow been calculated?") and then calls the plot.py.
Bonus) We rename sbat.main into sbat.run or sbat.compute, as main is a pretty meaningless name.
When we are at it, we could do the same for writing CSV output to disk.
Each of the submodules produces results we would like to visualize using
matplotlib. Plotting is invoked whenplot_resultsinsbat.yamlis setTrue.Currently, each submodule has their own plot functions, e.g. sbat.bflow.plot_along_streamlines. These plot functions are partially again wrapped into a higher plot function such as sbat.bflow.plot_bf_results.
This behavior has some downsides:
1) In
sbat.py, the plot functions are decentrally called within the class methods whenself.config.plot_results == True2) The plotting functions pollute the submodules, making them less readable and harder to debug 3) Whenoutput == False, plotting fails because there is no output directory defined. However, only plotting whenoutput == Trueis problematic as sometimes results are not computed because of the tight coupling between plotting and the submodulesAt the moment I suggest the following solution, but that can be discussed:
1) We relocate the plotting functions and place them into a separate file, could be for example
sbat/plot/plot.py. 2) We create a public class method into the model that is called by the user, that is executed at the end ofsbat.main. This method then checks the configuration for present results (as in "Has baseflow been calculated?") and then calls theplot.py. Bonus) We renamesbat.mainintosbat.runorsbat.compute, asmainis a pretty meaningless name.When we are at it, we could do the same for writing CSV output to disk.