anglr breaks for sf geometry with no attributes

[MilesMcBain]

Not sure if this is an anglr or silicate bug:

# Create an extent to triangulate
library(sf)
#> Linking to GEOS 3.5.1, GDAL 2.2.1, proj.4 4.9.3
library(anglr)
library(raster)
#> Loading required package: sp
my_extent <- st_as_sf(as(extent(c(153.185183093, 153.19443135, -27.705328446, -27.6967222119999)), 
                             "SpatialPolygons"))
anglr(my_extent, max_area = 0.008)
#> Error in .subset2(x, i, exact = exact): subscript out of bounds

my_extent$z <- 0
anglr(my_extent, max_area = 0.008)
#> $o
#> # A tibble: 1 x 2
#>       z  object_
#>   <dbl>    <chr>
#> 1     0 ea325999
#> 
#> $v
#> # A tibble: 4 x 3
#>         x_        y_    vertex_
#>      <dbl>     <dbl>      <chr>
#> 1 153.1852 -27.69672 biNoCfx5Fh
#> 2 153.1852 -27.70533 hetHjZASSt
#> 3 153.1944 -27.70533 UuxkFDs2E3
#> 4 153.1944 -27.69672 a66xCRKwTD
#> 
#> $meta
#> # A tibble: 1 x 2
#>    proj               ctime
#>   <chr>               <chr>
#> 1  <NA> 2017-12-14 04:43:32
#> 
#> $t
#> # A tibble: 2 x 2
#>    triangle_  object_
#>        <chr>    <chr>
#> 1 HHrECKtVDM ea325999
#> 2 auYuKm0Fbw ea325999
#> 
#> $tXv
#> # A tibble: 6 x 2
#>    triangle_    vertex_
#>        <chr>      <chr>
#> 1 HHrECKtVDM biNoCfx5Fh
#> 2 HHrECKtVDM hetHjZASSt
#> 3 HHrECKtVDM UuxkFDs2E3
#> 4 auYuKm0Fbw UuxkFDs2E3
#> 5 auYuKm0Fbw a66xCRKwTD
#> 6 auYuKm0Fbw biNoCfx5Fh
#> 
#> attr(,"class")
#> [1] "trimesh"

I traced the error to silicate:::sc_object.sf() which doesn't handle this case, it creates a data.frame with 0 columns.

[mdsumner]

Thanks! An important silicate bug. It also needs more careful treatment because I use this faster_as_tibble hack in a few projects now. There's probably a better solution as per hypertidy/silicate#39

I haven't forced anglr to require a later version, but I'll try to synch these two soonish.

[mdsumner]

And now really fixed.

[mdsumner]

It's fixed but also has changed:

library(sf)
library(anglr)
library(raster)
my_extent <- st_as_sf(as(extent(c(153.185183093, 153.19443135, -27.705328446, -27.6967222119999)), 
                         "SpatialPolygons"))
DEL(my_extent, max_area = 0.008)
## to pass z down use copy_down
copy_down(DEL(my_extent, max_area = 0.008), z = 1)

Also fwiw spex has this, which saves juggling a bit

spex::spex(extent(0, 1, 0, 1), crs = "+proj=laea")

I'll close this but feel free to re-open or pursue otherwise if you have problems!