Reduce package file size

[polettif]

This PR reduces the package size below the 5 MB CRAN policy threshold. To achieve this, I removed some routes from the NYC sample feed and simplified the shapes. The routes are not explicitly used in the vignettes or examples. Given the fact that the feed is already six years old at this point I don't think there is an issue with transforming it to a pure "example feed".

Notable changes:

• Renamed google_transit_nyc_subway.zip to nyc_subway.zip throughout the package
• Removed the following routes from nyc_subway.zip: A, SI, H, FS, J, Z and F.
• Minor changes in the vignette output, mainly in the servicepattern vignette but the gist of them hasn't changed.
• Removed the leaflet map in the introduction vignette that showed the bounding boxes for feeds. It increases the package size by almost 1 MB and it's not a core application for tidytransit. The code is still there.
• Reduced the figure size of some plots

The code to create the reduced feed is pasted below, for posterity.

library(dplyr)
gtfs_nyc = read_gtfs("google_transit_nyc_subway.zip")
gtfs = gtfs_nyc

routes_to_remove = c("A", "SI", "H", "FS", "J", "Z", "F")

# Remove constant columns ####
gtfs[!names(gtfs) %in% c(".", "agency", "transfers", "routes")] <- gtfs[!names(gtfs) %in% c(".", "agency", "transfers", "routes")] |>
  lapply(janitor::remove_constant, quiet = FALSE)

# Minimize shapes ####
shapes_sf = gtfs$shapes |> 
  shapes_as_sf() |> 
  rmapshaper::ms_simplify(keep = 0.1)
gtfs$shapes <- dplyr::as_tibble(sf_lines_to_df(shapes_sf))

# Remove select routes and trips ####
gtfs$routes |> select(route_id, route_desc) |> print(n = 30)
trips_to_keep = gtfs$trips |> 
  filter(!route_id %in% routes_to_remove) |> 
  distinct(trip_id)

# Remove unused service_ids ####
gtfs$trips <- gtfs$trips |> semi_join(trips_to_keep, "trip_id")
gtfs$stop_times <- gtfs$stop_times |> semi_join(gtfs$trips, "trip_id")

gtfs$calendar <- gtfs$calendar |> filter(service_id %in% gtfs$trips$service_id)
gtfs$calendar_dates <- gtfs$calendar_dates |> filter(service_id %in% gtfs$trips$service_id)

# Remove stops ####
stops_to_keep = gtfs$stops |> semi_join(gtfs$stop_times, "stop_id")
stops_to_keep <- c(stops_to_keep$stop_id, stops_to_keep$parent_station)

stops_to_keep <- c(stops_to_keep,
                   gtfs$transfers |> 
                     filter(from_stop_id %in% stops_to_keep | to_stop_id %in% stops_to_keep) |>
                     select(from_stop_id, to_stop_id) |> 
                     unlist(use.names = F) |> unique())

gtfs$stops <- gtfs$stops |> filter(stop_id %in% stops_to_keep)
gtfs$transfers <- gtfs$transfers |> filter(from_stop_id %in% stops_to_keep | to_stop_id %in% stops_to_keep)
gtfs$routes <- gtfs$routes |> semi_join(gtfs$trips, "route_id")

write_gtfs(gtfs, "google_transit_nyc_subway.zip")