Encountering "i.npz" issue using Jupyter notebook environment.

[AlienOverlord8769]

Been following this tutorial for using Gospl: https://worldbuildingpasta.blogspot.com/2023/03/an-apple-pie-from-scratch-part-viic.html#usinggospl

After some tweaks, I got to the point of actually running the erosion simulation, when this traceback generates:

File "/opt/conda/envs/gospl/lib/python3.11/site-packages/gospl/tools/inputparser.py", line 172, in _readDomain
    with open(self.meshFile) as meshfile:
         ^^^^^^^^^^^^^^^^^^^
FileNotFoundError: [Errno 2] No such file or directory: 'i.npz'

This stems from inputparser.py trying to read input.yml. Inside input.pyml is this, which defines "npdata" and thus is supposed to direct inputparser.py to detect files matching that case:

domain:
    npdata: 'input_folder/0Ma'
    flowdir: 5
    fast: False
    backward: False
    interp: 1
    #rstep: 5

The main issue here is that "0Ma.npz" is probably read, but for some reason is being overwritten somewhere upstream. So far, I have confirmed the following: The YAML supplementary file is not the problem. The runModel.py supplementary script is not the problem. Gospl's model.py does not appear to be hardcoding "i". FInally, no typos appear to be present. And yet, The actual YAML file being read is not the one it should be and I cannot proceed with the actual simulation. I have not been able to find any information on this particular discrepancy.

If you have the time, I would greatly appreciate your assistance. Let me know also if I should share the extra scrips.

[tristan-salles]

Hi there, not sure which version is installed in the mentioned tutorial. In the latest version of goSPL you will need to modify slightly the input file you are showing here, see this part of the online documentation:

domain:
    npdata: ['input/mesh','v','c','z']
    flowdir: 5

This will require to have the mesh.npz file with the corresponding variables (v, c and z) already defined.