As of #4621, summary variables are now evaluated lazily to improve performance. However, it has made getting a data frame of all the summary results a little clunky. E.g. instead of df = pd.DataFrame(sim.solution.summary_variables) you now need to do

summary_variables = sim.solution.summary_variables
all_summary_variables = {
    k: summary_variables[k] for k in summary_variables.all_variables
}
all_summary_variables["Cycle number"] = list(
    summary_variables.cycle_number
)
df = pd.DataFrame(all_summary_variables)

or something equivalent. I propose we add an all_summary_variables attribute that returns the same as all_summary_variables in the above snippet. Or just make it so that summary_variables["Cycle number"] returns list(summary_variables.cycle_number). This would (almost) be consistent with Solution where you can do sol.t or sol["Time [s]"].data.

@pipliggins let me know if this is sensible or if I'm missing something from the new SummaryVariables class.