Fix a bug in converting units for List[System]

[GardevoirX]

In _convert_systems_units, we calculate the unit conversion factor, conversion, for length, and then we enter a for-loop, to apply the conversion to systems:

metatomic/python/metatomic_torch/metatomic/torch/model.py

Lines 933 to 949 in 475fab2

	if model_length_unit == "" or system_length_unit == "":
	# no conversion for positions/cell/NL
	conversion = 1.0
	else:
	conversion = unit_conversion_factor(
	quantity="length",
	from_unit=system_length_unit,
	to_unit=model_length_unit,
	)
	new_systems: List[System] = []
	for system in systems:
	new_system = System(
	types=system.types,
	positions=conversion * system.positions,
	cell=conversion * system.cell,
	pbc=system.pbc,

However, when there are additional inputs attached to a system, to convert them to our requested unit, we calculate the conversion factor, and also store it in conversion:

metatomic/python/metatomic_torch/metatomic/torch/model.py

Lines 976 to 983 in 475fab2

	if requested.quantity != "" and unit is not None:
	conversion = unit_conversion_factor(
	quantity=requested.quantity,
	from_unit=unit,
	to_unit=requested.unit,
	)
	else:
	conversion = 1.0

This covers the original conversion factor, and in the next round of for-loop, we are using a wrong length conversion factor.

This is fixed by renaming the conversion factor used for length to length_conversion, but I am not sure if it's the best solution, and if we need to add some tests to prevent this from happening again in future

Contributor (creator of pull-request) checklist

• Tests updated (for new features and bugfixes)?
• Documentation updated (for new features)?
• Issue referenced (for PRs that solve an issue)?

Reviewer checklist

• CHANGELOG updated with public API or any other important changes?

[HaoZeke]

LGTM, thanks @GardevoirX. The fix is correct and minimal. One suggestion: it might be worth adding a regression test to prevent this from happening again. Something that tests:

• Multiple systems with unit conversion
• Additional inputs that also need unit conversion

Another thing is that in a follow up we can maybe hoist these into external functions but for now I think a test would be enouh.