RunPython

[llaniewski]

This PR extends the RunR capabilities with Python execution through the reticulate. You can now execute Python code with <RunPython/> element, where Solver is exposed in the same way as in R.

Example

<?xml version="1.0"?>
<CLBConfig version="2.0" output="output/">
	<Geometry nx="1024" ny="100">
		<MRT><Box/></MRT>
		<WVelocity><Box nx="1"/></WVelocity>
		<EPressure><Box dx="-1"/></EPressure>
	</Geometry>
	<Model>
		<Param name="VelocityX" value="0.05"/>
		<Param name="Viscosity" value="0.01"/>
	</Model>
	<RunR Iterations="100">
		print(sum(Solver$Quantities$Rho))
	</RunR>
	<RunPython Iterations="100">
print(Solver.Quantities.Rho.sum())
	</RunPython>
	<Solve Iterations="1000"/>
</CLBConfig>

FYI: @TravisMitchell @shkodm @mdzik
I think with this present in the code we can get rid of the native Python support.

[TravisMitchell]

I was already okay with removing native Python support but this also looks nice :).

[codecov-commenter]

Codecov Report

All modified and coverable lines are covered by tests ✅

Comparison is base (4e28d4e) 44.30% compared to head (e096d7d) 44.30%.

❗ Your organization needs to install the Codecov GitHub app to enable full functionality.

Additional details and impacted files

@@           Coverage Diff           @@
##           master     #468   +/-   ##
=======================================
  Coverage   44.30%   44.30%           
=======================================
  Files         169      169           
  Lines        7896     7896           
=======================================
  Hits         3498     3498           
  Misses       4398     4398           

Flag	Coverage Δ
d2q9	30.39% <100.00%> (ø)
d2q9_bc	26.68% <100.00%> (ø)
d2q9_kuper	27.46% <100.00%> (ø)
d3q27_PSM_NEBB	36.00% <100.00%> (ø)
d3q27_pf_velocity	30.46% <100.00%> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[shkodm]

Very exciting, thanks @llaniewski ! Do you want me to test it using the two-phase fracture setup I was working on?

[TravisMitchell]

@shkodm I think I would like you to test on the fracture setup, and maybe add to the multiphase tools repo as a demo for using it ;)

[llaniewski]

@shkodm That would be great. Hope this will simplify your setup a bit. edit: Of course, after you come back to the office.

@TravisMitchell @shkodm Some notes:

1. Seems that the read-only part works quite well, but setting stuff back in the Solver object doesn't really work well (eg. Solver.Geometry.BOUNDARY[0,0] = "None"). I think this would mean much more effort.
2. R objects are auto converted to python. Matrices become numpy arrays, so that's good. But the factor arrays seem to be converted to vectors of strings.

[shkodm]

@llaniewski I tested, and it seems to work for my purposes now (plotting phase field on the aperture plot). We can probably merge this one directly now and then create follow up PRs if I find some issues.

Regarding notes:

1. I don't think I need it right now, but setting stuff back would be cool. Maybe it is possible to do it in the follow up PR?
2. Factor array converted to strings is actually good, because of Inconsistency between RunR export and VTK #470