[FEATURE] heat_coupling: true fails with PyPSA 0.35.1 - "cannot rename 'Bus'"

[he1203-hub]

PyPSA-China version: 1.3.0
PyPSA version: 0.35.1
linopy version: 0.5.8
Error: ValueError: cannot rename 'Bus' because it is not a variable or dimension in this dataset
The error occurs in the define_nodal_balance_constraints function

[0set]

I found that the issue seems to be caused by an unconditional renaming of the "Bus" dimension in PyPSA, while in some cases the dimension has already been renamed (e.g. "Bus_el").

As a temporary workaround, I modified the following file:

/lib/python3.12/site-packages/pypsa/optimization/constraints.py

Around line ~657, I changed the renaming logic to explicitly check whether "Bus" exists in the dataset dimensions before renaming:

`if suffix:
# handle lhs
if 'Bus' in lhs.dims:
lhs = lhs.rename(Bus=f"Bus{suffix}")
elif f'Bus{suffix}' in lhs.dims:
pass # already renamed

# handle rhs
if 'Bus' in rhs.dims:
    rhs = rhs.rename(Bus=f"Bus{suffix}")
elif f'Bus{suffix}' in rhs.dims:
    pass  # already renamed

if mask is not None and 'Bus' in mask.dims:
    mask = mask.rename(Bus=f"Bus{suffix}")`

With this change, the error disappears

[irr-github]

Thanks, we will be updating the back-end to pypsa-v1.0.x soon, which hopefully will solve such issues. I have had issues with the linopy x-array dimensions in the past but not with the released version. You can also change the groupby dims.