search

amanzi / ats-demos

Demonstration problems of ATS capability (look here first!)
Other
12 stars 13 forks source link

Update demos to use new ats_xdmf #3

Open ecoon opened 4 years ago

ecoon commented 4 years ago

A bunch of the jupyter notebooks refer to ats/tool/utils scripts that need to go away (see https://github.com/amanzi/ats/issues/41 ) and need to be replaced with ats_xdmf, which should simplify a lot of the code for plotting output in these notebooks.

banesullivan commented 4 years ago

I may be able to help (perhaps just for a demo or two) on this. Is all of the needed code to reproduce the routines currently called in the notebooks present in the ats_xdmf module (I imagine some matplotlib routines will still need to be written)?

What exactly will it look like to do this conversion?

ecoon commented 4 years ago

I believe so, and yes, matplotlib still required -- ats_xdmf will get you the data in a sane format for most problems. That file is better documented than most of the rest of the python codebase.

Effectively you'd create a VisFile instance (one for each domain, so for the problems you're looking at, probably 2, surface + the "default" domain that is the subsurface. Then call loadMesh(), after which getArray() will return the output data in a sane shape/format numpy array. Visualizing that array is up to you (and likely done in the .ipynb files anyway).

ecoon commented 3 years ago

Most of these are done! Need to do a quick check and will close.