Dependency Incompatibility with linopy v0.5+

[CLiu0412]

ValueError in optimization with linopy >= 0.5.0 (Strict dimension naming)

Description:
Running the workflow with the latest linopy (tested with v0.5.8) causes a ValueError during the optimization phase. This appears to be due to linopy v0.5 enforcing strict dimension naming (force_dim_names=True), which conflicts with how PyPSA (v0.34.1) or ISPyPSA constructs constraints (specifically nodal balance constraints causing unnamed dimensions like dim_0).

Traceback:
File ".../pypsa/optimization/constraints.py", line 666, in define_nodal_balance_constraints
n.model.add_constraints(lhs, "=", rhs, name=f"Bus{suffix}-nodal_balance", mask=mask)
File ".../linopy/model.py", line 672, in add_constraints
self.check_force_dim_names(data)
ValueError: Added data contains non-customized dimension names. This is not allowed when setting force_dim_names to True.

Steps to Reproduce:
Install ISPyPSA with linopy>=0.5.0.
Run any standard workflow (e.g., test_workflow_2025.py).
Updates fail at network.optimize.solve_model().

Suggested Fix: Either: Pin Dependencies: Update pyproject.toml to restrict linopy<0.5.0 (and xarray<2024.06 if using older linopy).
Update Code: Refactor the constraint generation in ISPyPSA to ensure all xarray objects passed to linopy have named dimensions.

[nick-gorman]

Hey @CLiu0412, I'm struggling to replicate on my machine. Could you provide a uv.lock file so I can see the exact dependencies?

[CLiu0412]

uv lock.txt

Hi Nick, please find the attached file. the error is resolved by the coding agent.

Summary of events:

1. linopy was installed as v0.5.8 by default.
2. This triggered the error message shown above.
3. The coding agent resolved the issue by downgrading the version.
4. I asked the agent generate the attached report to document the fix.

[nick-gorman]

What about a copy of the lock file before the downgrade, when the error was occurring? That would be useful to help me replicate the error.