Add function or prominent documentation about how to get a single group&reftime slide input

[brookslogan]

Relates to: epi_slide, epix_slide, epi_df, epi_archive.

In evalcast, one major user issue with get_predictions was that there was no corresponding get_inputs function, and it took a bit of hacking to see what the input to the forecasters actually looked like. Things seem possible in epiprocess without any environment/global/attribute-storage hacks:

library(epiprocess)
library(dplyr)
example_geo_value = "ca"
example_ref_time_value = as.Date("2020-03-20")
jhu_csse_daily_subset %>% 
  group_by(.data$geo_value) %>% 
  filter(.data$geo_value == .env$example_geo_value) %>%
  epi_slide(~ tibble(data=list(.env$.x)), before = 6, ref_time_values=example_ref_time_value) %>%
  pull(.data$slide_value_data)

but this still seems a bit nontrivial.

It seems worth making this type of operation into a function or part of the epi[x]_slide documentation.

[ryantibs]

@lcbrooks Thanks! I think just documenting this somewhere seems sufficient, no? Perhaps this is worth putting into one of the vignettes.

What do you think about the advanced sliding vignette---in the last example in that vignette, where we talk about forecasting? Or in the archive vignette---in its last example, where we talk about forecasting?

[brookslogan]

I think this would be a fairly common operation, so maybe both vignettes + the help documentation if we include an @examples section.

If used, the code above should be edited so the user can just copy & paste and get some example input for any epi_df they are working with quickly. Another issue... while the code might point out to a user their code that they forgot a group_by(geo_value), maybe it's too easy to overlook (I've already made this mistake while trying out this code again...); this might be an argument for turning this into a function (that can be easily swapped with epi[x]_slide for dev & debugging), or maybe it's more hinting that we need to do something else to help users not forget to group_by(geo_value).