genotype parsing issues in experiments with reproduction = "sexual"

[mdahirel]

After running a simulation including reproduction = "sexual", there are two issues that arise during post-processing, when one wants to extract genotypic values of traits:

• first, they are stored as lists within Netlogo, which are not extracted in an intuitive way by nlrx:

  > exp_h2$metrics.turtles[[48]]$genotype_disp_slope[2000]
  [1] "[2.6197684135047491.587343036024559]"

  (compare with what happens for a clonal individual:)

> exp_h2$metrics.turtles[[47]]$genotype_disp_slope[2000]
[1] 1.556749

because both our genotypic traits are decimal, we can parse the string in the "sexual" case to get the value back using the decimal sign, but it's a mess

• second and more importantly, it means in an experiment that includes both clonal and sexual replicates, we can't merge them together easily since that would be asking to merge character strings and numbers

Proposed solution:

• find a way to parse the "genotype" string
• only work on clonal experiment for the moment
• in any case, always do and analyse clonal and sexual experiments separately

[mdahirel]

just realised (yes at 3 a.m.) that we don't actually need the values of each allele to be exported to R, we just need the genotypic value for each trait, i.e. the mean of the allelic values So we can perfectly arrange to only export this genotypic mean for disp_slope and logit_disp0, rather than the two allele values, side-stepping the list parsing issue

update 3:10 a.m. : it looks we can even do the averaging from R, like by asking to export "mean metric" rather than "metric" in the metric.turtles part of the experiment design. That would make things wayyy easier. Will double check tomorrow

[mdahirel]

solved as of cbad77d4af5a73929b328499eaa388f75af647d1 but probably moot due to developments in #16 anyway (we'll need to only export population-level metrics "pre-cooked" on the NetLogo side, so no more parsing issues)