Clarify how pre-processing steps are related to model_default equations

[amanda-minter]

Add comments to {epidemics} vignette to explain how pre-processing steps are related to how transmission is modelled in model_default. Specifically, this issue arose from the tutorials, the model code appears to be density dependent transmission, in the contact matrix preparation there is a division by population size meaning frequency dependent transmission.

Originally posted by @amanda-minter in epiverse-trace/tutorials-late#63 (comment)

[adamkucharski]

To highlight specific code, we should explain why the transpose is done here in the vignettes (it is so the internal function that normalises to ensure correct R0, then divides through by demography vector, produces a matrix that matches the per-capita beta terms are correct):

contact_matrix <- t(contact_data$matrix)