You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
In setplot.py we set plotdata.format to ascii or binary and then all output directories referenced have to have the same format. Often there's only one _output or maybe two similar ones if making comparison plots.
But I'm looking at a case where the I want to compare transects of a 2D solution with a 1D solution in a radially symmetric case, and the 2D output is big so would be best to use binary but there is no binary output option for 1D.
If plotdata.format = "binary" but an outdir has ascii format then it throws a non-informative error
*** Reached EOF in file .../fort.t0000
It took me a long time to figure out that's because for binary output the fort.t files are expected to contain one more line (with num_ghost) than in the ascii case.
We should consider allowing different formats for each outdir.
Also it would help to give a better error message currently.
At some point I had looked into passing down file_format as an argument down the call stack in visclaw but it got a bit too convoluted. Doing that would probably fix this and allow for removing this sort of state from VisClaw. One way around this and to allow more flexibility/modularity I was thinking about was to change the setplot functions so that they initiate the plotting if asked.
In
setplot.py
we setplotdata.format
toascii
orbinary
and then all output directories referenced have to have the same format. Often there's only one_output
or maybe two similar ones if making comparison plots.But I'm looking at a case where the I want to compare transects of a 2D solution with a 1D solution in a radially symmetric case, and the 2D output is big so would be best to use
binary
but there is nobinary
output option for 1D.If
plotdata.format = "binary"
but an outdir has ascii format then it throws a non-informative errorIt took me a long time to figure out that's because for binary output the fort.t files are expected to contain one more line (with
num_ghost
) than in the ascii case.We should consider allowing different formats for each outdir.
Also it would help to give a better error message currently.
See also #222 though this goes beyond that.
The text was updated successfully, but these errors were encountered: