No energy_tot in the output of the sphinx job

[max-hassani]

During the test of our latest docker images, the notebooks tests_sphinx_sphinx_check_all.ipynb fails with the following error:

if pr.inspect('spx_Fe_ferro')['output/generic/energy_tot'][0]-pr.inspect('spx_Fe_nonmag')['output/generic/energy_tot'][0]>0:
    raise ValueError('BCC Fe erromagnetic state has lower energy than nonmagnetic state')

---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
Input In [12], in <module>
----> 1 if pr.inspect('spx_Fe_ferro')['output/generic/energy_tot'][0]-pr.inspect('spx_Fe_nonmag')['output/generic/energy_tot'][0]>0:
      2     raise ValueError('BCC Fe erromagnetic state has lower energy than nonmagnetic state')

File /opt/conda/lib/python3.9/site-packages/pyiron_base/job/path.py:377, in JobPathBase.__getitem__(self, item)
    375     with open(file_name) as f:
    376         return f.readlines()
--> 377 return self.project_hdf5.__getitem__(item)

File /opt/conda/lib/python3.9/site-packages/pyiron_base/generic/hdfio.py:196, in FileHDFio.__getitem__(self, item)
    194 hdf_object = self.copy()
    195 hdf_object.h5_path = "/".join(item_abs_lst[:-1])
--> 196 return hdf_object[item_abs_lst[-1]]

File /opt/conda/lib/python3.9/site-packages/pyiron_base/generic/hdfio.py:166, in FileHDFio.__getitem__(self, item)
    164                 obj = self._convert_dtype_obj_array(obj)
    165             return obj
--> 166     raise ValueError(
    167         "Unknown item: {} {} {}".format(item, self.file_name, self.h5_path)
    168     )
    169 else:
    170     if (
    171         item_lst[0] == ""
    172     ):  # item starting with '/', thus we have an absoute HDF5 path

ValueError: Unknown item: energy_tot /home/pyiron/SPX_CHECK_ALL/spx_Fe_ferro.h5 /spx_Fe_ferro/output/generic

In the image, the latest pyiron packages are installed:

pyiron_atomistics         0.2.37             pyhd8ed1ab_0    conda-forge
pyiron_base               0.5.4              pyhd8ed1ab_1    conda-forge
sphinxdft                 3.0.3                h1bc63a5_4    conda-forge

[max-hassani]

@samwaseda, is this related to #533?

[samwaseda]

Oh yes it totally is. I’ll fix it today

[niklassiemer]

@samwaseda is this fixed?

[max-hassani]

@niklassiemer, @samwaseda, the problem still persists. The environment includes:

pyiron_atomistics         0.2.38
pyiron_base               0.5.6
sphinxdft                 3.0.3 

You can check the log of the test here

[max-hassani]

I wonder how our notebook test in pyiron_atomistics does not capture this error? Don't we run the same notebooks?

[niklassiemer]

That was the reason I asked for the notebooks used in pyiron in pyiron/pyiron#1397
If we use the same notebooks in different places we should think about fetching them in other places dynamically...

[samwaseda]

At least on the GitHub version the problem has been fixed.

[jan-janssen]

Somehow the tests still fail pyiron/pyiron#1397 and pyiron/docker-stacks#101

[samwaseda]

You just have to replace inspect by load.

[jan-janssen]

Thanks - it finally worked.