Update epidist_db() to give more fine-grained control

[joshwlambert]

This PR addresses #163 (and #91 which was superseded by #163).

The new implementation uses subset argument to allow the use to define subsetting criteria to select the entry from the database they want. It also includes a single_epidist argument to allow uses to specify if they only want a single <epidist> returned.

It also changes is_parameterised() to make it an S3 generic which has methods for <epiparam> and <epidist> objects.

[pratikunterwegs]

Thanks - just flagging that the new implementation would lead to this code

epidist_db(
  disease = "COVID-19",
  epi_dist = "onset_to_death",
  author = "Linton_etal"
)

returning multiple <epidists> after this change, rather than one, if I'm correct. Will file an issue in {cfr} once this PR is merged, as this would require some code in vignettes and possibly examples to be updated to either use single_epidist or manually subset later.

[pratikunterwegs]

Looks good to me - discussion continues on Slack so I'd be happy to take another look once the PR stabilises.

[pratikunterwegs]

Thanks for making these changes @joshwlambert - looks okay to me other than some minor points flagged here.

[pratikunterwegs]

Looks good to me, thanks for making the changes. Just the one issue of unused eparam in test-checkers.R, but otherwise feel free to merge.

[Bisaloo]

Since this function is exported, would it be possible to add in the description or in the examples a typical use case for users? At the moment, I can only think about internal use cases.