Basic plot integration

[PMeira]

EDIT: For an introduction and a long list of examples of this feature, see https://dss-extensions.org/dss_python/examples/Plotting/

---

Reproduce all plots from OpenDSS using Matplotlib and integrate them into the (DSS) plot command. Ideally plot, show, visualize, dump and export should be integrated somehow. Depending on the difficulty, other GitHub issues will be created per command.

Files

Relevant files from the OpenDSS source tree (items in bold are directly related to plots)

• Execute/ExecHelper.pas: DoDI_PlotCmd, DoVisualizeCmd, DoPropertyDump
• Execute/ExportOptions.pas: DoExportCmd
• Executive/PlotOptions.pas: DoPlotCmd
• Plot/DSSGraph.pas
• Plot/DSSPlot.pas: TDSSPlot. Can implement based on CommandLine/DSSPlot.pas

Related but doesn't seem relevant at the moment:

• Forms/DlgPlotOptions.pas

Brief description

DoPlotCmd processes plot options, setting up a TDSSPlot object and calling its Execute method. This indicates that we can in fact leave DoPlotCmd unchanged for the moment and handle the plot in the TDSSPlot.Execute method.
DoVisualizeCmd is similar, handles the parameters/options and calls TDSSPlot.DoVisualizationPlot. Same goes for ``, which calls TDSSPlot.DoDI_Plot

Alternatives

Two clear alternatives arise:

1. Implement logic and plotting in Python.
2. Use the Pascal and only pass a list of primitives to plot.

Since the first alternative seems both easier and more flexible, it will be investigated first.

[PMeira]

Going with:

Implement logic and plotting in Python.

This allows new useful features and it's also good example code.

With the callback setup from dss-extensions/dss_capi@157dcdb, at least some plots are possible without further API changes:

• Progress (currently tqdm) is shown for Solve mode=dutycycle number=300 stepsize=1s, while the plot command is Plot Monitor object=Mon1 Channels=(1,3,5) Bases=[2400, 2400, 2400]:

• Plot command is Plot Circuit Power Max=5000 dots=n labels=n C1=Blue 1ph=3 -- still missing some details but it's a good proof-of-concept:
  

This will be published when the basic plot variations are implemented. The same base can be used to further integrate with Jupyter (e.g. DSS cells), Spyder, and QGIS.

Initially, the markers from the old DSS plotting tool will be used. Ported back in 2018:

Alternative or customized component markers can be added as options later.

Other callbacks could be added for reports, which could use maybe Arrow DFs down the line.

[PMeira]

On plotting, the initial implementation is available since dss_python 0.12.0. Matplotlib and SciPy are required. Some examples:

from dss import DSS
import dss.plot
dss.plot.enable()

DSS.Text.Command = 'redirect EPRITestCircuits/ckt5/Run_ckt5.dss'
DSS.Text.Command = 'batchedit load..* daily=default'
DSS.Text.Command = 'Solve mode=daily'
DSS.Text.Command = 'calcincmatrix'

DSS.Text.Command = 'plot circuit'
DSS.Text.Command = 'visualize powers Transformer.mdv_sub_1'
DSS.Text.Command = 'visualize voltages Transformer.mdv_sub_1'
DSS.Text.Command = 'visualize currents Transformer.mdv_sub_1'
DSS.Text.Command = 'plot monitor object=ckt5_totalizedmonitor'
DSS.Text.Command = 'Plot Profile'
DSS.Text.Command = 'plot matrix incidence'

plot circuit will be polished for a future version.

[PMeira]

Most commands have at least an initial implementation now. Missing is CompareCases. I think it's safe to close this now. We only need to add some example notebooks to illustrate the feature for v1.0.