Split up create_stan_data()?

[jamesmbaazam]

I've been reviewing #871 to bring myself up to date with developments while I was away and I realise that create_stan_data() is getting quite long and difficult to parse. For one, it's combining two concerns: what I'd call static data creation and dynamic modelling of the initial growth. I would suggest pulling out the modeling of initial growth part (https://github.com/epiforecasts/EpiNow2/blob/main/R/create.R#L504-L524) as it will be easier to maintain and improve if need be and to point to. It only depends on data, seeding_time, and horizon and returns the prior_infections and prior_growth, making it quite focused. This is more of a maintenance issue than an enhancement and can be categorised as low priority.

Thoughts?

[sbfnk]

Sounds good to me