Clean up `epi_slide` `f` parameter documentation

[rachlobay]

As discussed with Logan, update the @param documentation for the f parameter in epi_slide. The following are the points to address:

• Instead of “If a function, f must take x, a data frame with the same column names as the original object; followed by any number of named arguments; and ending with ...“, the proper thing may be that f should take x, g/k/key/think-of-a-better-name, and named args (dots don’t seem necessary; I think originally they were just there to absorb the unlisted g).
• Correct "same column names as the original object" --- we might only get the non-grouping columns here. [This raises a separate issue. We are feeding the user an epi_df via f that might not have the geo_value column as one might expect. Hopefully we can get away with it, but maybe this needs to be a proper epi_df or just a tibble instead.]
• If you think it clarifies things, change “named arguments” -> “named arguments (which will be taken from ...)“.
• Make sure that the description of g above matches/resembles how ?group_modify describes such a parameter, if it makes sense.
• In the docs for the f being a formula case, describe .y in the same manner as g, if .y can indeed be used [I'm guessing that it is usable].

Originally mentioned in PR #85.

[brookslogan]

Still see remnants of the old docs somehow. [In slide.Rmd rather than slide.R. I must have used these slide.Rmd text as the basis for another slide.R doc update causing a reversion on a feature branch.]