Adds two new arguments to points_to_od()

max_dist lets you set the maximum distance e.g. 10,000m useful if you wanted to make a national OD matrix but only for local trips

max_dest lets you set the maximum number of destinations regardless of distance. E..g only to the 10 nearest destinations for each OD.

Both method are very useful for large OD matrices as faster and use less memory.

Benchmark

p = pct::get_centroids_ew()
p = p[1:1000,]

system.time(r1 <- od::points_to_od(p))
user  system elapsed 
   3.31    0.01    3.32 
nrow(r1)
[1] 1000000


system.time(r2 <- points_to_od(p, max_dist = 10000))
user  system elapsed 
   0.22    0.00    0.22 
nrow(r2)
[1] 33930

This scales exponentially so for large ODs e.g. all LSOAs it makes a massive difference.

I've also added st_centriod to p & pd so non-point geometries can be used