wk

R build status Codecov test coverage

The goal of wk is to provide lightweight R, C, and C++ infrastructure for a distributed ecosystem of packages that operate on collections of coordinates. First, wk provides vector classes for points, circles, rectangles, well-known text (WKT), and well-known binary (WKB). Second, wk provides a C API and set of S3 generics for event-based iteration over vectors of geometries.

Installation

You can install the released version of wk from CRAN with:

install.packages("wk")

You can install the development version from GitHub with:

# install.packages("remotes")
remotes::install_github("paleolimbot/wk")

If you can load the package, you’re good to go!

library(wk)

Vector classes

Use wkt() to mark a character vector as containing well-known text, or wkb() to mark a vector as well-known binary. Use xy(), xyz(), xym(), and xyzm() to create vectors of points, and rct() to create vectors of rectangles. These classes have full vctrs support and plot()/format() methods to make them as frictionless as possible working in R and RStudio.

wkt("POINT (30 10)")
#> <wk_wkt[1]>
#> [1] POINT (30 10)
as_wkb(wkt("POINT (30 10)"))
#> <wk_wkb[1]>
#> [1] <POINT (30 10)>
xy(1, 2)
#> <wk_xy[1]>
#> [1] (1 2)
rct(1, 2, 3, 4)
#> <wk_rct[1]>
#> [1] [1 2 3 4]
crc(0, 0, 1)
#> <wk_crc[1]>
#> [1] [0 0, r = 1]

Generics

The wk package is made up of readers, handlers, and filters. Readers parse the various formats supported by the wk package, handlers calculate values based on information from the readers (e.g., translating a vector of geometries into another format), and filters transform information from the readers (e.g., transforming coordinates) on the fly. The wk_handle() and wk_translate() generics power operations for many geometry vector formats without having to explicitly support each one.

C API

The distributed nature of the wk framework is powered by a ~100-line header describing the types of information that parsers typically encounter when reading geometries and the order in which that information is typically organized. Detailed information is available in the C and C++ API article.

wk_debug(
  as_wkt("LINESTRING (1 1, 2 2, 3 3)"),
  wkt_format_handler(max_coords = 2)
)
#> initialize (dirty = 0  -> 1)
#> vector_start: <Unknown type / 0>[1] <0x16d75aac0> => WK_CONTINUE
#>   feature_start (1): <0x16d75aac0>  => WK_CONTINUE
#>     geometry_start (<none>): LINESTRING[UNKNOWN] <0x16d75a950> => WK_CONTINUE
#>       coord (1): <0x16d75a950> (1.000000 1.000000)  => WK_CONTINUE
#>       coord (2): <0x16d75a950> (2.000000 2.000000)  => WK_ABORT_FEATURE
#> vector_end: <0x16d75aac0>
#> deinitialize
#> [1] "LINESTRING (1 1, 2 2..."

sf support

The wk package implements a reader and writer for sfc objects so you can use them wherever you’d use an xy(), rct(), crc(), wkb(), or wkt():

wk_debug(
  sf::st_sfc(sf::st_linestring(rbind(c(1, 1), c(2, 2), c(3, 3)))),
  wkt_format_handler(max_coords = 2)
)
#> initialize (dirty = 0  -> 1)
#> vector_start: LINESTRING B[1] <0x16d75dac8> => WK_CONTINUE
#>   feature_start (1): <0x16d75dac8>  => WK_CONTINUE
#>     geometry_start (<none>): LINESTRING[3] <0x16d75da10> => WK_CONTINUE
#>       coord (1): <0x16d75da10> (1.000000 1.000000)  => WK_CONTINUE
#>       coord (2): <0x16d75da10> (2.000000 2.000000)  => WK_ABORT_FEATURE
#> vector_end: <0x16d75dac8>
#> deinitialize
#> [1] "LINESTRING (1 1, 2 2..."

Lightweight

The wk package has zero dependencies and compiles in ~10 seconds.