Package 'polymer'

Simple overlapping layers

Description

A, B and C are three partially overlapping polygon layers with different attributes: A is four squares in a 2 by 2 grid over the unit square, B is a single triangle and C a single thin rectangle.

Format

'sf' data frames with no coordinate reference system:

• A: 4 rows, layer, a and geometry.

• B: 1 row, b and geometry.

• C: 1 row, c and geometry.

Examples

polymer(A, B, C)

---

Field and soil polygons

Description

Two polygon layers used to show overlay: soil has 12 polygons with a made-up soil type, field has two rectangular fields that cross them. The example comes from a question on the GIS Stack Exchange (https://gis.stackexchange.com/a/230161), using the Luxembourg boundaries shipped with the 'raster' package.

Format

'sf' data frames in longitude and latitude (WGS84):

• soil: 12 rows, soil (character) and geometry.

• field: 2 rows, field (character) and geoms.

Examples

polymer(soil, field)

---

Fuse a polymer into overlay pieces

Description

Dissolve the triangles of a polymer into polygons, one for each distinct set of covering input features. This is the n-ary intersection of all layers (the same partition that sf::st_intersection() computes for a single layer), derived from the mesh rather than from pairwise overlay.

Usage

fuse(x, ...)

layer_n(x, n = 2, ...)

Arguments

x	A polymer (see polymer()).
...	Ignored.
n	Minimum number of covering features, for layer_n().

Details

layer_n() keeps the pieces covered by at least n features.

Pieces are dissolved along mesh edges, so they share vertices exactly where they touch. Outer rings are counter-clockwise and holes clockwise. Every piece is a MULTIPOLYGON, with one polygon for each connected part.

Value

A data frame with one row per piece and columns

• n: the number of input features covering the piece.

• origins: a list column, each element a data frame of layer and feature (row within its layer) for the features covering the piece.

• geometry: a wk::wkb() vector with the polymer's coordinate reference system. Convert with sf::st_as_sf() if you work with 'sf'.

Examples

pool <- polymer(A, B, C)
pieces <- fuse(pool)
pieces[c("n", "geometry")]

## which features made the deepest piece
pieces$origins[[which.max(pieces$n)]]

## pieces covered two or more times
plot(pool)
wk::wk_plot(layer_n(pool, 2)$geometry, add = TRUE, col = "grey")
wk::wk_plot(layer_n(pool, 3)$geometry, add = TRUE, col = "dodgerblue")

layer_n(pool, 3)

---

Plot a polymer

Description

Draw the triangles of the mesh. With show_intersection = TRUE the triangles covered by two or more input features are filled.

Usage

## S3 method for class 'polymer'
plot(x, ..., show_intersection = FALSE)

Arguments

x	A polymer.
...	Arguments passed to graphics::polygon(), such as col and border.
show_intersection	Logical, fill the triangles covered more than once (default FALSE).

Value

x, invisibly.

Examples

plot(polymer(A, B, C))
plot(polymer(A, B, C), show_intersection = TRUE, col = "firebrick")

---

Build a polymer mesh from polygon layers

Description

Combine one or more polygon layers into a single triangle mesh, where every input edge from every layer is kept as a chain of mesh edges, and record which input features cover each triangle.

Usage

polymer(..., engine = "cdtr")

Arguments

...	Polygon layers. Each can be anything the 'wk' package can read (an 'sf' data frame or sfc, a wk::wkb() or wk::wkt() vector, and so on). Data frames keep their non-geometry columns as attributes.
engine	The triangulation engine. Only "cdtr" is available.

Details

All the inputs' edges are triangulated together in one constrained Delaunay triangulation. Where edges from different features cross, the crossing is inserted as a new vertex, so the mesh is a partition of the plane that respects every input boundary. Vertices are shared on exact coordinate equality before triangulation.

Each triangle is then labelled with the features that contain it: a triangle is inside a feature when its centroid is inside an odd number of the feature's rings, so holes are subtracted. Triangles covered by no feature (gaps enclosed by the inputs) are dropped.

The result is the mesh and its index, not a set of polygons. Use fuse() or layer_n() to get polygons back.

Value

A polymer object, a list with

• mesh: a wkpool_tri (see wkpool::new_wkpool_tri()) holding the triangles on a shared vertex pool. The vertex table has an origin column: "input" for input vertices and "crossing" for vertices inserted where input edges cross.

• index: a data frame with one row per triangle and covering feature: triangle (row in mesh), layer (position of the layer in ...) and feature (row within that layer).

• layers: a list with one data frame of attributes per layer (no geometry).

• crs: the coordinate reference system shared by the inputs.

Examples

pool <- polymer(A, B, C)
pool
head(pool$index)

## the mesh is a 'wkpool' triangle pool
pool$mesh

---

Print a polymer

Description

Print a short summary of a polymer mesh: the number of layers, input features, triangles, crossing vertices and the deepest overlap.

Usage

## S3 method for class 'polymer'
print(x, ...)

Arguments

x	A polymer.
...	Ignored.

Value

x, invisibly.

Examples

print(polymer(A, B, C))

---