Check power for edge cases marginal power 0 or 1

[EeethB]

Results are currently different for calculate_power(g, marginal_power = c(0, 0, 0), sim_seed = 51123, force_closure = TRUE) and calculate_power(g, marginal_power = c(0, 0, 0), sim_seed = 51123).

Should validation exclude 0/1 for marginal power?

[EeethB]

Running graph_calculate_power() regardless of test type performs correctly now regardless of test type when power_marginal is all 0 or all 1. The only issue I've found still is when power_marginal is has a 0 or 1, and alpha is also the corresponding value (0 or 1). Perhaps validation should exclude 0/1 for alpha, since that's a null check.