Update nowcast beyond -1 week maximum delay

Would be interesting to see if you can get any performance gain from expanding the nowcast method a bit.

My suggestion would be to try out using baselinenowcast on your case data by reference and report date (which I will call df below). X is the number of draws you want.

rep_tri <-baselinenowcast::as_reporting_triangle(df, max_delay = 5)
prob_nowcast_df <- baselinenowcast(rep_tri, draws = X)

prob_nowcast_df will return a dataframe of probabilistic nowcasts of the expected final number of cases by reference date with columns: pred_count, reference_date, draw and output_type.

We just added the wrapper function features so would be curious to hear any feedback you have! We haven't put them into vignettes yet so look to the tests and docs for example usage https://baselinenowcast.epinowcast.org/reference/baselinenowcast.reporting_triangle.html

Start with the tests and documentation for the baselinenowcast wrapper referenced in the issue, then locate the current one-week nowcast limit in NowcastAutoGP. Compare the suggested calls using case data by reference and report date; done means evaluating delays beyond one week and reporting any performance gain or feedback.