Skip to contents

Combine or Split Complex Geometries

Source: R/giotto_structures.R
combine_split_geoms.Rd

Geometries can be either single/simple or multi with multiple closed rings defined as a single record. combineGeom() is used to combine polygons. splitGeom() breaks combined geometries down into constituent parts.
Avoid using the SpatVector methods. They are lower-level and does not deal with IDs like might be expected by Giotto.

Usage

# S4 method for class 'giottoPolygon'
combineGeom(
  x,
  by = NULL,
  dissolve = FALSE,
  fun = "mean",
  ...,
  fmt = "poly_%d",
  previous_id = "source_id"
)

# S4 method for class 'giottoPolygon'
splitGeom(x, fmt = "poly_%d", previous_id = "source_id", ...)

# S4 method for class 'SpatVector'
combineGeom(x, by = NULL, dissolve = FALSE, fun = "mean", ...)

# S4 method for class 'SpatVector'
splitGeom(x, ...)

Arguments

x

geometry class to combine or split.

by

character. Column name of variable used to group the geometries. Will be used as the new poly_ID column. All geometries will be combined if not provided.

dissolve

logical. Should borders between aggregated geometries be dissolved?

fun

function used to aggregate values. Either an actual function, or for the following, their name: "mean", "max", "min", "median", "sum", "modal", "any", "all", "prod", "which.min", "which.max", "table", "sd" (sample standard deviation) and "std" (population standard deviation)

...

additional params to pass to terra::aggregate() (and then to fun, such as na.rm=TRUE) or terra::disagg()

fmt

character. sprintf formatting to use to generate poly_ID column values if no attributes are retained after combining.

previous_id

character. If not NULL, column name to store original poly_ID values under. Note that merged IDs will be NA.

Value

the same class as x

Details

Currently, these are simple wrappers around terra's aggregate(dissolve = FALSE) and disagg() with some additional handling around the poly_ID column and a different name to avoid confusion with spatial feature aggregation.

Functions

  • combineGeom(SpatVector): Avoid using.

  • splitGeom(SpatVector): Avoid using.

Geometry Attributes Handling

  • combineGeom() attributes are only kept if by param is used. For conflicts where more than one value is present for a single variable after combining, they are resolved using fun param. When by=NULL, all attributes are dropped and a default poly_ID column is generated based on the fmt param.

  • splitGeom() attributes are duplicated for each part that was part of the same multipolygon. The poly_ID column will be made unique.

  • overlaps and centroids information will be dropped after either operation.

Examples

dt <- data.table::data.table(
    id = c(
        rep('a', 3), # Triangle (id 'a')
        rep('b', 4), # Square 1 (id 'b')
        rep('c', 4), # Square 2 (id 'c')
        rep('d', 4) # Square 3 (id 'd')
    ),
    x = c(
        0, 1, 0.5,
        2, 5, 5, 2,
        5, 5, 6, 6,
        6, 7, 7, 6
    ),
    y = c(
        0, 0, 1,
        2, 2, 5, 5,
        2, 3, 3, 2,
        5, 5, 6, 6
    )
)
plot_colors <- getRainbowColors(4)
gpoly <- createGiottoPolygon(dt, verbose = FALSE)
plot(gpoly, col = plot_colors)


gpoly$group_id <- sprintf("group_%d", c(1, 2, 2, 3))
gpoly$values <- 1:4
force(gpoly)
#> An object of class giottoPolygon
#> spat_unit : "cell"
#> Spatial Information:
#>  class       : SpatVector 
#>  geometry    : polygons 
#>  dimensions  : 4, 3  (geometries, attributes)
#>  extent      : 0, 7, 0, 6  (xmin, xmax, ymin, ymax)
#>  coord. ref. :  
#>  names       : poly_ID group_id values
#>  type        :   <chr>    <chr>  <int>
#>  values      :       a  group_1      1
#>                      b  group_2      2
#>                      c  group_2      3
#>  centroids   : NULL
#>  overlaps    : NULL

c_all <- combineGeom(gpoly) # combine all
force(c_all)
#> An object of class giottoPolygon
#> spat_unit : "cell"
#> Spatial Information:
#>  class       : SpatVector 
#>  geometry    : polygons 
#>  dimensions  : 1, 1  (geometries, attributes)
#>  extent      : 0, 7, 0, 6  (xmin, xmax, ymin, ymax)
#>  coord. ref. :  
#>  names       : poly_ID
#>  type        :   <chr>
#>  values      :  poly_1
#>  centroids   : NULL
#>  overlaps    : NULL
plot(c_all, col = plot_colors)


c_gid <- combineGeom(gpoly, by = "group_id")
force(c_gid)
#> An object of class giottoPolygon
#> spat_unit : "cell"
#> Spatial Information:
#>  class       : SpatVector 
#>  geometry    : polygons 
#>  dimensions  : 3, 4  (geometries, attributes)
#>  extent      : 0, 7, 0, 6  (xmin, xmax, ymin, ymax)
#>  coord. ref. :  
#>  names       : poly_ID mean_values source_id agg_n
#>  type        :   <chr>       <num>     <chr> <int>
#>  values      : group_1           1         a     1
#>                group_2         2.5        NA     2
#>                group_3           4         d     1
#>  centroids   : NULL
#>  overlaps    : NULL
plot(c_gid, col = plot_colors)

# `dissolve` removes touching boundaries
plot(combineGeom(gpoly, by = "group_id", dissolve = TRUE),
     col = plot_colors)


# split combined geometries
s_cgid <- splitGeom(c_gid)
force(s_cgid)
#> An object of class giottoPolygon
#> spat_unit : "cell"
#> Spatial Information:
#>  class       : SpatVector 
#>  geometry    : polygons 
#>  dimensions  : 4, 4  (geometries, attributes)
#>  extent      : 0, 7, 0, 6  (xmin, xmax, ymin, ymax)
#>  coord. ref. :  
#>  names       : poly_ID mean_values source_id agg_n
#>  type        :   <chr>       <num>     <chr> <int>
#>  values      :  poly_1           1   group_1     1
#>                 poly_2         2.5   group_2     2
#>                 poly_3         2.5   group_2     2
#>  centroids   : NULL
#>  overlaps    : NULL
plot(s_cgid, col = plot_colors)