Host already in inactive HostSet in call to sim.inner.tree

[ArtPoon]

This seems to be related to superinfection. Here is a small reproducing example:

Parameters:
  simTime:  10.0  # simulation time units before most recent sample date
  lambda: 1e-3  # birth rate (susceptibles only)
  mu: 5e-4  # baseline death rate
  mu2: 0.01  # infected death rate
  beta:  1e-4  # transmission rate
  gamma:  0.1  # clearance rate
  psi:  0.05  # sampling rate
  omega:  0.01  # loss of immunity
Compartments:
  S:
    infected: false
    transmission:
      I: {I: beta*S*I}  # S becomes I due to infection by member of I
    size: 9999  # initial
  I:
    infected: true
    migration:
      I_samp: psi*I
    transmission:
      I: {I: beta*I*I}  # superinfection
    size: 1  
    bottleneck.size: 1 
    coalescent.rate: 1.0
  I_samp:
    infected: true
    size: 0
Sampling:
  mode: compartment 
  targets:
    I_samp: 10  

> setwd("~/git/twt")
> require(twt)
> settings <- read_yaml("iss172.yaml")
> mod <- Model$new(settings)
> set.seed(32)
> dyn <- sim.dynamics(mod)
> outer <- sim.outer.tree(dyn)
> inner <- sim.inner.tree(outer)
Error in recipient$count.pathogens() : attempt to apply non-function
In addition: Warning message:
In active$get.host.by.name(e$to.host) :
  HostSet does not contain Host US_I_8

[ArtPoon]

Something funny seems to be going on with the plot.OuterTree code as well:

[ArtPoon]

Filtered to events that are associated with US_I_8

> outer$get.log()
         time        event from.comp src.comp to.comp from.host to.host
1  5.48312299    migration         I     <NA>  I_samp       I_1    <NA>
5  4.95789376 transmission         S        I       I    US_I_5     I_1
7  4.68270987 transmission         S        I       I    US_I_7  US_I_5
8  4.60949063 transmission         I        I       I    US_I_8  US_I_7
9  4.60839425 transmission         S        I       I    US_I_9  US_I_7
18 3.87393870 transmission         S        I       I   US_I_18  US_I_9
23 3.57004488 transmission         S        I       I   US_I_23  US_I_8
25 3.43283675 transmission         S        I       I   US_I_25 US_I_18
29 3.24422543 transmission         S        I       I   US_I_29 US_I_25
31 3.06146075 transmission         S        I       I   US_I_30 US_I_23

[ArtPoon]

The problem is probably here:

twt/R/simInnerTree.R

Lines 168 to 173 in c84d5f1

	.do.infection <- function(e, inner, envir=baseenv()) {
	active <- inner$get.active()
	inactive <- inner$get.inactive()
	# we are going back in time, so a recipient must already be active
	recipient <- active$get.host.by.name(e$to.host)

[ArtPoon]

Stepping through sim.inner.tree with the outer object breaks at the following iteration:

> row
[1] 23
> e
       time        event from.comp src.comp to.comp from.host to.host
23 3.570045 transmission         S        I       I   US_I_23  US_I_8
> active$get.names()
[1] "I_11"    "US_I_14" "US_I_15" "US_I_16" "US_I_18" "US_I_20" "US_I_21" "US_I_22"
> inactive$get.names()
 [1] "US_I_26" "US_I_25" "US_I_30" "US_I_32" "US_I_24" "US_I_33" "US_I_27" "US_I_28"
 [9] "US_I_34" "US_I_35" "US_I_31" "US_I_29" "US_I_36" "US_I_37" "I_2"     "I_1"    
[17] "US_I_3"  "US_I_5"  "US_I_8"  "US_I_7"  "I_12"    "I_4"     "I_13"    "I_10"   
[25] "US_I_9"  "I_19"    "US_I_17" "US_I_6"

[ArtPoon]

It looks like what is happening is that there is only one Pathogen lineage carried by host US_I_8 and when that lineage is pulled out by a superinfection event, the host is set to 'inactive'. However there is still the first infection event of the host at an earlier point in time, and as we go back in time to that event the simulation loop assumes that the host is still active. We need to add a check of whether the host is inactive and then ignore the first infection event if it is inactive.

[ArtPoon]

Potential fix in dev branch ae79d20.
A warning message came up during further testing:

Warning message:
In inactive$add.host(host) : Host US_I_18 is already in this HostSet!

Probably have a superfluous call to add.host due to superinfection.