Incorrect mass-weighting for centre-of-mass (COM) force

In our documentation we currently state the following:

This form of mass-weighting is dimensionally inconsistent: we equate forces (and hence energies) to mass * forces (and hence mass * energies) when we opt to mass-weight the COM force, rather than correctly re-weight the forces:

- Incorrect mass-weighting: $\mathbf{F}_{i} = \frac{s_{i} m_{i}}{N}\mathbf{F}_{\mathrm{COM}}$, $E = \frac{\sum_{i} m_{i}}{N} E_{\mathrm{COM}}$

- Correct mass-weighting: $\mathbf{F}_{i} = \frac{s_{i} m_{i}}{\sum_{j} m_{j}}\mathbf{F}_{\mathrm{COM}}$, $E = E_{\mathrm{COM}}$

We should address this. Thanks to @ludovicaisa for realising there was an inconsistency here and helping to figure out what was going on!

Things to do to address this issue:
- [ ] Fix the code in NanoVer Server
- [ ] Make sure the docs are consistent

These changes should also not affect the way that the non-mass-weighted COM force works.