add zbox and mbox attributes

Request

If an sfc object has XYZ(M) dimensions, can you add a zbox (mbox) attribute to the object?

Example

In this contrived example I'm setting the Z to range in [0,2].

library(sf)
l <- sf::st_linestring(matrix(c(0,0,0,1,1,1,2,2,1,3,3,2), ncol =2, byrow = T))
sfc <- sf::st_sfc( l )

Can a zbox be calculated and added, so it would take the form

zbox <- c(0,2)
attr(sfc, "zbox") <- zbox

attributes(sfc)

# $class
# [1] "sfc_LINESTRING" "sfc"           
# 
# $precision
# [1] 0
# 
# $bbox
# xmin ymin xmax ymax 
# 0    0    3    3 
# 
# $crs
# Coordinate Reference System: NA
# 
# $n_empty
# [1] 0
# 
# $zbox
# minz maxz 
# 0    2

Use case

I'm starting to add a lot of elevation and time animations into madpeck, so it would be useful to have the ranges of these values easily accessible so they can be read directly, rather than calculated on-the-fly.

[edzer]

I think that is not an unreasonable request, but I'm afraid I won't have much time in the near future to work on this.

About the naming: I'd suggest to have two attributes, zbox and mbox, that may be present in case of Z or M data, and if present contain named elements zmin and zmax, resp. mmin and mmax, to keep in line with how bbox does this. These values should never be NA, and absence of an attribute does not indicate that there is no Z or M.

No worries, I'll see if I can make some time in the coming days to implement this and submit a PR.

I've started working on this; would zbounds and mbounds be better attribute names, rather than "box", because they are not actually boxes?

[edzer]

I suggest to call them z_range and m_range.

[dcooley]

Here's what It currently looks like

pt <- sf::st_sfc( sf::st_point(x = c(0,0)) )
st_z_range( pt )
# NULL

st_m_range( pt )
# NULL

pt <- sf::st_sfc( sf::st_point(x = c(0,0,1)) )
pt
# Geometry set for 1 feature 
# geometry type:  POINT
# dimension:      XYZ
# bbox:           xmin: 0 ymin: 0 xmax: 0 ymax: 0
# z_range:           zmin: 1 zmax: 1
# epsg (SRID):    NA
# proj4string:    NA
# POINT Z (0 0 1)

st_z_range( pt )
# zmin zmax 
# 1    1

pt <- sf::st_sfc( sf::st_point(x = c(0,0,1,2)) )
pt
# Geometry set for 1 feature 
# geometry type:  POINT
# dimension:      XYZM
# bbox:           xmin: 0 ymin: 0 xmax: 0 ymax: 0
# z_range:           zmin: 1 zmax: 1
# m_range:           mmin: 2 mmax: 2
# epsg (SRID):    NA
# proj4string:    NA
# POINT ZM (0 0 1 2)

st_m_range( pt )
# mmin mmax 
# 2    2

pt <- st_sf( geometry = pt )
pt
# Simple feature collection with 1 feature and 0 fields
# geometry type:  POINT
# dimension:      XYZM
# bbox:           xmin: 0 ymin: 0 xmax: 0 ymax: 0
# z_range:           zmin: 1 zmax: 1
# m_range:           mmin: 2 mmax: 2
# epsg (SRID):    NA
# proj4string:    NA
# geometry
# 1 POINT ZM (0 0 1 2)

st_z_range( pt )
# zmin zmax 
# 1    1

st_m_range( pt )
# mmin mmax 
# 2    2

I think I've covered all cases I can think of. Are you happy if I make a PR at this stage?

Another more complicated example

n <- 100
lst <- list()
min_z <- numeric(n)
max_z <- numeric(n)

set.seed(123)

for(i in 1:n) {
	m <- matrix(rnorm(sample(seq(3,300,by=3), size = 1)), ncol = 3)
	min_z[i] <- min(m[,3])
	max_z[i] <- max(m[,3])
	lst[[i]] <- sf::st_linestring( m )
}

sfc <- sf::st_sfc( lst )

# Geometry set for 100 features 
# geometry type:  LINESTRING
# dimension:      XYZ
# bbox:           xmin: -3.402024 ymin: -3.487803 xmax: 3.455241 ymax: 3.038443
# z_range:        zmin: -4.055104 zmax: 3.715721
# epsg (SRID):    NA
# proj4string:    NA
# First 5 geometries:
# LINESTRING Z (0.8005543 -0.5996083 2.149919, 1....
# LINESTRING Z (1.360652 -0.3475426 -1.617883, -0...
# LINESTRING Z (-0.4905574 1.909104 -0.1194526, -...
# LINESTRING Z (-1.177967 0.1602164 -0.3460009, -...
# LINESTRING Z (-2.01421 -0.9454088 1.633568, 0.2...

sf::st_z_range( sfc ) == c(min(min_z), max(max_z))
# zmin zmax 
# TRUE TRUE

[edzer]

This can now be closed, right?