Competing risk

[csmith701]

Includes a series of methodologic updates including:

• Fixing indexing error
• Resolve confusing variable naming (Resolve confusing variable names between covariate and non-covariate versions #72)
• Allow user to specify model start time (How to trigger the CRAN submission workflow #77)
• Properly account for competing risks
• Allow for multiple extra-household infection probabilities
• Update package data to be on the natural scale (Re-bake the *_res.rda fixtures on the natural scale #48)

[pearsonca]

Assorted questions below!

[pearsonca]

what's a bit unclear to me here: will this handle the multistrain case, if those strains both have multiple infectious states in their chain? as in, if i have strain A and B, and both have pre-symptomatic infectious states (or maybe just one does), does this indexing machinery work correctly?

[csmith701]

Yep it should be able to handle that scenario just fine.

[pearsonca]

there is a ton of code duplication between this an not-cov - is there someway to have the deduplicated code, and then an import of non-cov vs cov calculations?

[csmith701]

It's something I have been thinking about. The issue is that in the covariate version the infection probabilities become matrices. Still I think some opportunity to deduplicate.

[pearsonca]

this appears to be a fairly complicated "toy" example - possibly to more narrowly / simply test?