Adapt LibBi model with all Variables in Output

Some rbi.helper functions require all variables to have outputs. This function modifies a LibBi model so that this is the case.

everything_from_model(model)

Arguments

model	A LibBi model, loaded using bi_model.

Value

A LibBi model.

Examples

library(rbi)
#> 
#> Attaching package: ‘rbi’
#> The following objects are masked from ‘package:ModelTBBCGEngland’:
#> 
#>     read_libbi, save_libbi
#> The following objects are masked from ‘package:stats’:
#> 
#>     filter, optimise, predict, simulate, update
#> The following object is masked from ‘package:utils’:
#> 
#>     fix
#> The following object is masked from ‘package:base’:
#> 
#>     sample

model <- system.file(package="rbi", "SIR.bi") # get full file name from package
model <- bi_model(model) # load model
model <- replace_all(model, "noise n_recovery", "noise n_recovery\\(has_output = 0\\)")

model
#> bi_model:
#> =========
#>  1: model SIR {
#>  2:   const h = 7
#>  3:   const N = 1000
#>  4:   const d_infection = 14
#>  5:   noise n_transmission
#>  6:   noise n_recovery(has_output = 0)
#>  7:   state S
#>  8:   state I
#>  9:   state R
#> 10:   state Z
#> 11:   obs Incidence
#> 12:   param p_rep
#> 13:   param p_R0
#> 14:   sub parameter {
#> 15:     p_rep ~ uniform(0,1)
#> 16:     p_R0 ~ uniform(1,3)
#> 17:   }
#> 18:   sub initial {
#> 19:     S <- N - 1
#> 20:     I <- 1
#> 21:     R <- 0
#> 22:     Z <- 1
#> 23:   }
#> 24:   sub transition {
#> 25:     n_transmission ~ wiener()
#> 26:     n_recovery ~ wiener()
#> 27:     Z <- (t_now % 7 == 0 ? 0 : Z)
#> 28:     inline i_beta = p_R0 / d_infection * exp(n_transmission)
#> 29:     inline i_gamma = 1 / d_infection * exp(n_recovery)
#> 30:     ode (alg='RK4(3)', h=1e-1, atoler=1e-2, rtoler=1e-5) {
#> 31:       dS/dt = - i_beta * S * I / N
#> 32:       dI/dt = i_beta * S * I / N - i_gamma * I
#> 33:       dR/dt = i_gamma * I
#> 34:       dZ/dt = i_beta * S * I / N
#> 35:     }
#> 36:   }
#> 37:   sub observation {
#> 38:     Incidence ~ poisson(p_rep * Z)
#> 39:   }
#> 40:   sub proposal_initial {
#> 41:     S <- N - 1
#> 42:     I <- 1
#> 43:     R <- 0
#> 44:     Z <- 1
#> 45:   }
#> 46:   sub proposal_parameter {
#> 47:     p_rep ~ truncated_gaussian(mean = p_rep, std = 0.03, lower = 0, upper = 1)
#> 48:     p_R0 ~ truncated_gaussian(mean = p_R0, std = 0.2, lower = 1, upper = 3)
#> 49:   }
#> 50: }

everything_from_model(model)
#> bi_model:
#> =========
#>  1: model SIR {
#>  2:   const h = 7
#>  3:   const N = 1000
#>  4:   const d_infection = 14
#>  5:   noise n_transmission
#>  6:   noise n_recovery
#>  7:   state S
#>  8:   state I
#>  9:   state R
#> 10:   state Z
#> 11:   obs Incidence
#> 12:   param p_rep
#> 13:   param p_R0
#> 14:   sub parameter {
#> 15:     p_rep ~ uniform(0,1)
#> 16:     p_R0 ~ uniform(1,3)
#> 17:   }
#> 18:   sub initial {
#> 19:     S <- N - 1
#> 20:     I <- 1
#> 21:     R <- 0
#> 22:     Z <- 1
#> 23:   }
#> 24:   sub transition {
#> 25:     n_transmission ~ wiener()
#> 26:     n_recovery ~ wiener()
#> 27:     Z <- (t_now % 7 == 0 ? 0 : Z)
#> 28:     inline i_beta = p_R0 / d_infection * exp(n_transmission)
#> 29:     inline i_gamma = 1 / d_infection * exp(n_recovery)
#> 30:     ode (alg='RK4(3)', h=1e-1, atoler=1e-2, rtoler=1e-5) {
#> 31:       dS/dt = - i_beta * S * I / N
#> 32:       dI/dt = i_beta * S * I / N - i_gamma * I
#> 33:       dR/dt = i_gamma * I
#> 34:       dZ/dt = i_beta * S * I / N
#> 35:     }
#> 36:   }
#> 37:   sub observation {
#> 38:     Incidence ~ poisson(p_rep * Z)
#> 39:   }
#> 40:   sub proposal_initial {
#> 41:     S <- N - 1
#> 42:     I <- 1
#> 43:     R <- 0
#> 44:     Z <- 1
#> 45:   }
#> 46:   sub proposal_parameter {
#> 47:     p_rep ~ truncated_gaussian(mean = p_rep, std = 0.03, lower = 0, upper = 1)
#> 48:     p_R0 ~ truncated_gaussian(mean = p_R0, std = 0.2, lower = 1, upper = 3)
#> 49:   }
#> 50: }