add forecast_opts()

[sbfnk]

Description

This PR closes #867 and part of #898 whilst enabling addressing #640.

It introduces a forecast argument to estimate_infections() that is set via forecast_opts() (preferred over add_horizon() as it allows us to identify the last date in the data - though the add_horizon() function exists internally so we could make the switch at a later time if we want).

Initial submission checklist

• My PR is based on a package issue and I have explicitly linked it.
• I have tested my changes locally (using devtools::test() and devtools::check()).
• I have added or updated unit tests where necessary.
• I have updated the documentation if required and rebuilt docs if yes (using devtools::document()).
• I have followed the established coding standards (and checked using lintr::lint_package()).
• I have added a news item linked to this PR.

After the initial Pull Request

• I have reviewed Checks for this PR and addressed any issues as far as I am able.

[github-actions]

This is how benchmark results would change (along with a 95% confidence interval in relative change) if 8003d06 is merged into main:

• ✔️default: 23.5s -> 28.9s [-38.67%, +84.58%]
• ✔️no_delays: 26.6s -> 29.9s [-0.55%, +24.78%]
• ✔️random_walk: 9.68s -> 10.2s [-1.38%, +11.95%]
• ✔️stationary: 13.2s -> 13.8s [-15.75%, +24.17%]
• ✔️uncertain: 38.6s -> 59.1s [-69.85%, +175.99%]
  Further explanation regarding interpretation and methodology can be found in the documentation.

[jamesmbaazam]

Thanks for implementing this, Seb.

I made a few comments and suggestions.

[sbfnk]

Thanks, @jamesmbaazam. Did the new identification of gaps look OK to you (I think you weren't convinced of the idea in the corresponding Issue)?

EpiNow2/R/create.R

Line 447 in 8003d06

if (length(gaps) == 1 && gaps > 1) { ## all gaps are the same

[jamesmbaazam]

I agree but - should we then also rename cases_time to case_times?

That would be better too.

[jamesmbaazam]

Thanks, @jamesmbaazam. Did the new identification of gaps look OK to you (I think you weren't convinced of the idea in the corresponding Issue)?

I like the current implementation.

[github-actions]

This is how benchmark results would change (along with a 95% confidence interval in relative change) if f0f453e is merged into main:

• ✔️default: 19.3s -> 30.7s [-64.27%, +183.1%]
• ✔️no_delays: 23.4s -> 30.8s [-35.06%, +97.92%]
• ✔️random_walk: 9.56s -> 9.16s [-12.76%, +4.51%]
• ✔️stationary: 12.2s -> 11.8s [-16.96%, +8.97%]
• ✔️uncertain: 33s -> 34.1s [-8.22%, +14.96%]
  Further explanation regarding interpretation and methodology can be found in the documentation.