Support for 0 day run-in/wash-out parameters

[zimmeee]

Some treatments have fast run-in/wash-out periods. The simulator operates at the day timescale, and the run-in (tc_in) or wash-out (tc_out) parameters must be > 0.

For the cognition study, we have fast acting treatments (caffeine, l-theanine) with run-in/wash-out periods measured in hours, not days. We want the simulator to be able to account for fast acting treatments.

[mrjonesolmedo]

In the interim, it is possible to set decimal values (e.g. 0.1 days) for tc_in and tc_out.

The numerical entry UI in the Shiny app doesn't reflect this well since the increment/decrement is a full day, but it is possible to manually enter a decimal value > 0.

[zimmeee]

It does accept those values, but it still operates at the day timescale. For example, here is what the data look like with these parameter settings:

effect_size = c(0,3,6)
tc_in = c(0.1, 0.1, 0.1)
tc_out = c(.1, .1, .1)

Note how the effect doesn't kick in until the 6th day, and each treatment period has a lag