forty-spots.Rmd

This image has a drawing that bugs me a little as it's not a triangulation, but nearly is. 

Here we download it, read as a 3-layer raster brick, plot with mapview and then digitize on the points. 


```R
u <- "https://pbs.twimg.com/media/DF5UuWsVYAAR7sX.jpg"
download.file(u, basename(u), mode = "wb")

library(raster)
r <- brick(basename(u))
#m <- mapview::plainview(r)
## we can't plainview it so let's pretend it's mercator
extent(r) <- extent(0, ncol(r)*1000, 0, nrow(r)*1000)
projection(r) <- "+init=epsg:3857"
library(mapedit)
library(mapview)

## enter the interactive drawing session
pt <- mapview(r)  %>% editMap()

## note that we obtain the "drawn" sub-object, and have to transform it to web mercator
## and also union it to obtain a single point, and triangulate that to get the mesh that I wanted
tri <- st_triangulate(st_union(st_transform(pt$drawn, 3857)))
## now a trad plot
plot(tri)
plotRGB(r, add = TRUE)
plot(tri, col = NA, border = "dodgerblue", lty = 2, lwd = 2, add = TRUE)

```


Finally, the drawn points (in raw longitude-latitude matrix form). 

```R
#dput(matrix(unlist(st_geometry(pt$drawn)), ncol  = 2, byrow = TRUE))
structure(c(0.3296, 0.3516, 0.8679, 1.0437, 2.7576, 2.5269, 2.2412, 
2.2083, 1.9116, 1.9666, 2.384, 2.7795, 2.2522, 1.604, 1.3184, 
1.1755, 1.3513, 1.615, 1.604, 2.0874, 2.7246, 2.1533, 1.4172, 
1.1426, 1.2964, 1.1096, 1.4612, 1.593, 1.8127, 2.0325, 2.2083, 
2.2742, 2.5378, 3.0103, 2.9663, 3.219, 3.219, 3.3728, 3.1201, 
2.8455, 3.4257, 3.5792, 3.5902, 3.8643, 6.7301, 6.3153, 6.5118, 
6.1187, 6.2389, 5.6269, 5.8128, 5.3098, 5.4629, 5.7581, 5.6051, 
5.2332, 5.2004, 5.3863, 5.1128, 5.2113, 4.6969, 4.9268, 4.883, 
4.8611, 4.6202, 4.5874, 4.193, 3.645, 3.9191, 4.4121, 4.0615, 
3.5902, 4.1492, 4.7188, 4.5874, 4.5764, 4.3793, 4.1602, 4.1821, 
3.8643), .Dim = c(40L, 2L))
```

Here's the final picture, note how the black lines between the nodes (the pardalote's 40 spots) are not (all) Delaunay (but also not convex ...)