Specify rate functions in input.json
#62
Conversation
| @@ -4,6 +4,12 @@ use std::path::PathBuf; | |||
| use ixa::{define_global_property, ContextGlobalPropertiesExt, IxaError}; | |||
| use serde::{Deserialize, Serialize}; | |||
|
|
|||
| #[derive(Debug, Clone, Serialize, Deserialize, PartialEq)] | |||
There was a problem hiding this comment.
@k88hudson-cfa what do you think of this approach for specifying in different rate functions?
| /// Instantiate the rate functions specified in the global parameter `rate_of_infection` as actual | ||
| /// rate functions for the simulation that are assigned randomly to agents when they are infected. | ||
| pub fn instantiate_rate_fns(context: &mut Context) -> Result<(), IxaError> { | ||
| let rate_of_infection = context.get_params().rate_of_infection.clone(); |
There was a problem hiding this comment.
Really don't love that this structure basically duplicates the info in rate_of_infection. I think there are two potential solutions: (a) implement InfectiousnessRateFn on enum Rates instead, or (b) actually move the info in rate_of_infection rather than copying it. I don't think (a) is a possible option because in implementing the methods for InfectiousnessRateFn on EmpiricalRate, you need to create new vectors that you store in the initialization process, and I don't see a way to do this if we just implement the method on the type as it is fed in from deserialization. I think (b) is an easy option, but I am open to other thoughts and ways of thinking about this.
I think we have to move the info (at least) because in EmpiricalRate we want to create
| Rates::Constant(rate) => Box::new(ConstantRate::new(rate, infection_duration)?), | ||
| Rates::Empirical(rate_fn) => { | ||
| let (t, r): (Vec<f64>, Vec<f64>) = rate_fn.into_iter().unzip(); | ||
| Box::new(EmpiricalRate::new(t, r)?) |
There was a problem hiding this comment.
Also, doesn't using Box create an overhead here/should we be concerned about that? I could imagine us having a number of infectiousness curves on the order of the number of agents in the simulation.
WIP