use leafgl to visualise mesh3d

mesh3d_and_leafgl.R

library(ggplot2)
library(leafgl)
library(leaflet)
library(sf)

sfx <- sf::st_transform(sf::st_cast(dplyr::filter(silicate::inlandwaters, Province == "Tasmania"), "POLYGON")[2, ], 4326)
topo <- ceramic::cc_elevation(sfx, zoom = 9)

mesh <- anglr::as.mesh3d(anglr::copy_down(anglr::DEL(sfx, max_area = .0001), 
                                          topo))
st_as_sf.mesh3d <- function(x, ...) {
  idx = if (!is.null(x$it)) x$it else x$ib
  idx = rbind(idx, idx[1, ])
  nc = dim(idx)[2L]
  idx = as.vector(idx)
  
  mat = cbind(x$vb[1L, idx], x$vb[2L, idx])
  grp = rep(seq_len(nc), each = 4L)
  lst = split.data.frame(mat, grp)
  
  sf::st_sf(
    average_z = tapply(x$vb[3L, idx], grp, mean, na.rm = TRUE)
    , geometry = sf::st_sfc(
      lapply(
        lst
        , function(i) sf::st_polygon(list(i))
      )
    )
    , ...
  )
}

mesh_sf = st_as_sf(mesh, crs = 4326)

clr = makeColorMatrix(~average_z, mesh_sf, palette = "inferno")

leaflet() %>%
  addProviderTiles("CartoDB.DarkMatter") %>%
  addGlPolygons(mesh_sf, color = clr, popup = TRUE) %>%
  setView(lng = 146.5, lat = -42.5, zoom = 7)

Yip, it's much faster. I updated the original gist with a much faster version of st_as_sf.mesh3d, Rcpp still eats it for breakfast!

library(microbenchmark)

microbenchmark(
  sf = st_as_sf(mesh)
  , md = mapdeck:::mesh_to_sf(mesh, c("vb", "it"))
  , times = 10
)

Unit: milliseconds
 expr        min         lq      mean     median        uq       max neval
   sf 5517.50783 5825.09536 6232.2307 6232.57181 6524.7249 6998.7204    10
   md   82.27326   94.87767  107.7158   99.13353  105.7012  166.2793    10

@mdsumner maybe the above at_as_sf.mesh3d should be included in anglr for round-tripping purposes? It isn't as performant as @dcooley 's Rcpp version, but it isn't half bad either.

I'll be into this in a few days going riding 👍👍

Have fun!