BUG: The QHA makers should probably allow cell shape relaxations

When calculating the EOS in the QHA makers, it is done at a fixed cell volume. Of course, this makes sense otherwise the structure will lose the strain applied. But I believe it should also be done with the unit cell shape allowed to change (at fixed cell volume). This would be `FrechetCellFilter(constant_volume=True)` or ISIF=4 for VASP. At the very least, this should be a toggle-able option. https://github.com/materialsproject/atomate2/blob/68447d313890113c5eb53e5a41f7c24d325dbc58/src/atomate2/forcefields/flows/qha.py#L128-L133 https://github.com/materialsproject/atomate2/blob/68447d313890113c5eb53e5a41f7c24d325dbc58/src/atomate2/vasp/flows/qha.py#L66-L72