Coupling diagnostics

[alexandrebouchard]

• track, for each replica, if that replica touched the prior at least once.
• Once all replica have done that, give a little signal, e.g. switch the colour of the standard out (maybe dark red to dark blue? dark red to black? should be clr blind friendly) and/or an emoji at the end of the PT diagnostic row to denote that "if ELE held, the system would have fully coupled".
• a new column showing the fraction of index processes that have coupled
• some classical convergence checks on the energy statistics so that we do all the above only once the energies have stabilized
• documentation linking this to the upcoming arxiv on the theory we are developing...