Add support for parsing intermediate excited state energies

[oliver-s-lee]

For example, in excited state optimisations, where currently only the final energies are parsed.

This will probably have to go into a new attribute, as changing etenergies to support this would involve adding another list layer, which is not backwards compatible.

The other excited state params and metadata['opt_state'] might also need to be looked at to make sure they're pointing to the correct excited state over the course of the optimisation (in case of excited state reordering).