Closed JanStreffing closed 3 years ago
@JanStreffing This looks really nice, thanks a lot for adding it!
Looks like the failure is at the level of the notebook, and in your output notebook I found the warning:
sh: convert: command not found
mv: cannot stat './results/ice_vol_test/figures/ice_vol_test_ice_integrals_combined_nocomparison_icevol_combined.png_trimmed.png': No such file or directory
I hope:
module load GCCcore/.10.3.0
module load ImageMagick/7.0.10-25
will fix the issue. Relying on ImageMagick for conversion is not the best idea, although I thought it is installed everywhere, but I am wrong :)) Probably we have to change it to pythonic way of trimming.
I found and loaded ImageMagick/7.0.10-25
, but not the GCCcore/.10.3.0:
module spider GCCcore
Lmod has detected the following error: Unable to find: "GCCcore".
I tried like this, since I do have a gcc compiler by default, without loading a module. I had to make one more small change to the path in drivers.py, then it worked.
Hey Nikolay, I added a diagnostic where the ice area, extent and volume for March, September and for SH & NH are added in just one plot each. The plots look fine, e.g. but adding them to the html file did not quite work:
In particular this part:
_combined.png_trimmed.png
seems wrong. Maybe you can have a quick look at my work here: https://github.com/FESOM/fdiag/tree/combined_ice_integrals I'm guessing this is easy to fix.p.s. I tired to have the feature such that there is one color per experiment and different linestyles per season and hemisphere, but I found the result much less clear than using different colors for each season and hemisphere. Therefore I deviate fromt he default blue/orange color scheme.