UFloat Warnings with ComputedEntry.energy

[kavanase]

Python version

Python 3.10-3.12

Pymatgen version

2025.5.2

Operating system version

No response

Current behavior

From running our doped tests with the latest pymatgen release, we get many of these warnings (which breaks tests):

Using UFloat objects with std_dev==0 may give unexpected results.

From setting the warning level to error, the traceback shows:

File ~/Packages/doped/doped/core.py:1428, in DefectEntry.sc_entry_energy(self)
   1425     return self._sc_entry_energy
   1427 self._sc_entry_hash = hash(self.sc_entry)
-> 1428 self._sc_entry_energy = self.sc_entry.energy
   1430 return self._sc_entry_energy

File ~/Packages/pymatgen/src/pymatgen/entries/computed_entries.py:350, in ComputedEntry.energy(self)
    347 @property
    348 def energy(self) -> float:
    349     """The *corrected* energy of the entry."""
--> 350     return self.uncorrected_energy + self.correction

File ~/Packages/pymatgen/src/pymatgen/entries/computed_entries.py:367, in ComputedEntry.correction(self)
    362 """
    363 Returns:
    364     float: the total energy correction / adjustment applied to the entry in eV.
    365 """
    366 # adds to ufloat(0.0, 0.0) to ensure that no corrections still result in ufloat object
--> 367 corr = ufloat(0.0, 0.0) + sum(ufloat(ea.value, ea.uncertainty) for ea in self.energy_adjustments)
    368 return corr.nominal_value

File ~/miniconda3/lib/python3.10/site-packages/uncertainties/core.py:1024, in ufloat(nominal_value, std_dev, tag)
    992 """
    993 Create an uncertainties Variable
    994 
   (...)
   1021    and `tag` which match the input values.
   1022 """
   1023 if std_dev == 0:
-> 1024     warn("Using UFloat objects with std_dev==0 may give unexpected results.")
   1025 return Variable(nominal_value, std_dev, tag=tag)

UserWarning: Using UFloat objects with std_dev==0 may give unexpected results.

I think this is due to a recent update in uncertainties, which now includes this warning, which occurs when there are no energy adjustments with uncertainties in ComputedEntry.energy_adjustments. Some discussion here: lmfit/lmfit-py#999

Expected Behavior

Not totally sure, but it seems like that warning is there to stay, so might make sense to catch/handle it here somehow?
Only use ufloat when there is an energy adjustment with non-zero uncertainty, otherwise just normal floats for correction values?

Minimal example

from pymatgen.entries.computed_entries import ComputedEntry

ce = ComputedEntry(composition="H", energy=1)
ce.energy  # shows warning

Relevant files to reproduce this bug

No response