Add new configuration generator into `ve_run`

[davidorme]

Add the use of the new generator from #1087 into ve_run alongside the existing one - models won't use it, but we can iron out issues from tests on loading to that new style. It should still take dict or TOML inputs, so should be fine.