Make room_history handle a list of room_ids

[GregSutcliffe]

In true R style, R history should be able to handle a vector of rooms.

This PR moves the implementation for a single room to an internal method, and then rewrites room_history to iterate over the ID list and nest the results. Here's an example:

room_ids <- c("!nmvUwBzGOuTedhIPEV:libera.chat","!pMZboYFCScZJfXmOtH:ansible.im")
since <- "2021-12-07"
room_history(room_ids, since)

# A tibble: 2 × 2
# Groups:   room [2]
  room                            events            
  <chr>                           <list>            
1 !nmvUwBzGOuTedhIPEV:libera.chat <tibble [42 × 6]> 
2 !pMZboYFCScZJfXmOtH:ansible.im  <tibble [255 × 6]>

This can then be expanded, eg by tidyr::unnest(events)

I've also included a small change to add id to the schema, as that's useful when storing and then re-running the script - events should not change, but at least we can check :)

[johrpan]

I added some minor style fixes that you wouldn't disagree with (I think). Before merging, I would merge those with the actual commit adding the changes.

[johrpan]

Otherwise, it looks good to me and seems to work as expected. I would suggest, to not have the nested tibble, but just include the room column to the result. This also facilitates the case, where you only fetch one room and unnesting is especially redundant.

[johrpan]

Also, one more thing: In my opinion, the internal function should be named differently (i.e. not with the prepended dot). For example room_history_internal() or single_room_history(). But I leave that to your preferences.

[GregSutcliffe]

Very happy with those style fixes (seriously, how did that glue get left behind?) - much thanks for that. I'll squash on merge.

I think you're right on the naming - I got the idea from meetupr but looking at other tidyverse packages I don't see it so much in use. I'll think of a name that works.

On the nesting - I think I'll merge as-is for now, as I don't have a lot of time today. I agree nesting for a single room is un-needed, but we can address that in a new PR - I'm thinking an optional parameter to the function, and the default is "nest if rooms >1, unnest if ==1)" ...