| Title: | Common Forms for Complex Hierarchical and Relational Data Structures |
|---|---|
| Description: | Generate common data forms for complex data suitable for conversions and transmission by decomposition as paths or primitives. Paths are sequentially-linked records, primitives are basic atomic elements and both can model many forms and be grouped into hierarchical structures. The universal models 'SC0' (structural) and 'SC' (labelled, relational) are composed of edges and can represent any hierarchical form. Specialist models 'PATH', 'ARC' and 'TRI' provide the most common intermediate forms used for converting from one form to another. The methods are inspired by the simplicial complex <https://en.wikipedia.org/wiki/Simplicial_complex> and provide intermediate forms that relate spatial data structures to this mathematical construct. |
| Authors: | Michael D. Sumner [aut, cre], John Corbett [ctb] (the original inspiration), Simon Wotherspoon [ctb], Kent Johnson [dtc], Mark Padgham [aut] |
| Maintainer: | Michael D. Sumner <[email protected]> |
| License: | GPL-3 |
| Version: | 0.7.1.9001 |
| Built: | 2024-12-05 04:27:20 UTC |
| Source: | https://github.com/hypertidy/silicate |
Arcs are unique paths that connect nodes. In a polygon layer with shared boundaries, the arcs are the linear features that have no branches.
ARC(x, ...) ## Default S3 method: ARC(x, ...) ## S3 method for class 'PATH' ARC(x, ...)ARC(x, ...) ## Default S3 method: ARC(x, ...) ## S3 method for class 'PATH' ARC(x, ...)
x |
input model |
... |
arguments passed to methods |
Nodes are the vertices where three or more arcs meet. An arc can exist without including any nodes, a path that has no neighbouring relationship with another path.
This is not the same terminology as used by other systems, such as "arc-node". The
arc_link_vertex mapping is inherently ordered, but we don't consider order of arcs.
Duplicated arcs (i.e. complementary turns around neighbouring polygons) are not kept.
The object_link_arc mapping records which arc belongs to the objects, so feature polygons
can in theory be reconstructed within objects by tracing arc_link_vertex start and end point
identity.
ARC model
a <- ARC(minimal_mesh) sc_arc(a) sc_arc(minimal_mesh)a <- ARC(minimal_mesh) sc_arc(a) sc_arc(minimal_mesh)
Filter an SC model, currently only dplyr::filter for SC is available.
## S3 method for class 'SC' filter(.data, ...)## S3 method for class 'SC' filter(.data, ...)
.data |
data object of class SC |
... |
passed to |
Apply expressions as if used on the object table. See sc_object(x) for that form.
Currently all the vertices are still kept, so the model (and any plots) include the filtered edges as well as undifferentiated points. This is likely to change ...
an SC() model, with some parts filtered out
library(dplyr) sc <- SC(inlandwaters) plot(filter(sc, Province == "Tasmania")) plot(filter(sc, Province %in% c("Victoria", "South Australia", "New South Wales"))) plot(filter(SC(minimal_mesh), a == 1))library(dplyr) sc <- SC(inlandwaters) plot(filter(sc, Province == "Tasmania")) plot(filter(sc, Province %in% c("Victoria", "South Australia", "New South Wales"))) plot(filter(SC(minimal_mesh), a == 1))
A 'sfc_GEOMETRYCOLLECTION' of four-cornered triangles ('POLYGON') created by ear cutting the North Carolina polygon from sf.
TRI0(ear_gc)TRI0(ear_gc)
A data set flight tracks in XYZM form, a form of 4D tracks. Primarily to
explore the use of silicate as able to represent this topologically, and to experiment with
auto-time-based plotting in anglr.
Provided by Kent Johnson (kent37) in a github discusion where the data was attached in a zip file.
Original form (in extdata/flight_tracks) is a XYZM LINESTRING shapefile containing 144 flight tracks of aircraft departing runway 33L at Boston Logan airport on January 27, 2017. Data is from an ADS-B recorder. Each point includes lat, lon, altitude in feet and time in North American Eastern Standard Time (EST).
Converted via sf into silicate::PATH normal form, see (data-raw/flight_tracks.R).
The inland waters are lakes and inland waters presenting as holes within the bounded regions of Australian (and New Caledonian) provinces.
This is an extract from the old Manifold DVD. It is in sf format. The features have variables IDandProvince' they are (in order):
"103841"Australian Capital Territory
"103842"New Caledonia
"103843"New South Wales
"103846"South Australia
"103847"Tasmania
"103848"Victoria
There's no good reason that New Caledonia is included and not Queensland (for example) it's just what happened doing a quick crop and extract with the mouse. Lord Howe Island and Macquarie Island are both present, as part of New South Wales and Tasmania respectively.
path <- PATH(inlandwaters) plot(path) obj <- split(path$path_link_vertex, path$path_link_vertex$path_) cl <- grDevices::colors()[-1L] cols <- sample(cl, length(obj), replace = length(obj) > length(cl)) op <- par(mfrow = grDevices::n2mfrow(length(obj)), mar = rep(0, 4)) funplot <- function(ob, vert, col) { vx <- c("x_", "y_") plot(dplyr::inner_join(ob, vert, "vertex_")[vx], col = col, type = "l", axes = FALSE) } junk <- lapply(seq_along(obj), function(a) { funplot(obj[[a]], path$vertex, cols[a]) invisible(NULL) }) par(op)path <- PATH(inlandwaters) plot(path) obj <- split(path$path_link_vertex, path$path_link_vertex$path_) cl <- grDevices::colors()[-1L] cols <- sample(cl, length(obj), replace = length(obj) > length(cl)) op <- par(mfrow = grDevices::n2mfrow(length(obj)), mar = rep(0, 4)) funplot <- function(ob, vert, col) { vx <- c("x_", "y_") plot(dplyr::inner_join(ob, vert, "vertex_")[vx], col = col, type = "l", axes = FALSE) } junk <- lapply(seq_along(obj), function(a) { funplot(obj[[a]], path$vertex, cols[a]) invisible(NULL) }) par(op)
The simplest pairing of two polygons with one shared edge. One polygon contains a hole and a concavity, the other is a simply convex. This is composed of four "arcs", one around each polygon, one for the shared edge, and one for the isolated hole. There are two nodes, the endpoints of the single shared edge.
arc <- ARC(minimal_mesh) plot(arc) sc_arc(arc) sc_node(arc)arc <- ARC(minimal_mesh) plot(arc) sc_arc(arc) sc_node(arc)
This data set is in legacy format and will be removed. A couple of polygons with a single shared edge between them, in PRIMITIVE form.
A PATH model is a direct translation of a simple features-alike object to normal form. This is four tables with the three kinds of entities, "objects" (or "features"), "paths" (or "parts") and "vertices", and a table to link the one-to-many relation between paths and vertices.
PATH(x, ...) ## S3 method for class 'SC' PATH(x, ...) ## S3 method for class 'TRI' PATH(x, ...) ## Default S3 method: PATH(x, ...)PATH(x, ...) ## S3 method for class 'SC' PATH(x, ...) ## S3 method for class 'TRI' PATH(x, ...) ## Default S3 method: PATH(x, ...)
x |
input model |
... |
arguments passed to methods |
In a data set with no parts touching their neighbours, the only normalization of the vertices will be the removal of the duplicated closing coordinate on any polygon ring, and on any self-intersecting case within a single path.
PATH()$path should always have columns object_ path_ subobject ncoords_
a PATH model, with tables 'object', 'path', 'path_link_vertex' and 'vertex'
sc_path, sc_coord
Structural form requires only tables 'object' and 'vertex'.
Minimal columns is x,y but can be grouped by path_ for separate paths, then subobject_ and object_ for full polygon support.
PATH0(x, ...) ## Default S3 method: PATH0(x, ...) ## S3 method for class 'PATH0' PATH0(x, ...) PATH0_from_df( x, ..., path_ = "path_", object_ = "object_", subobject_ = "subobject_", x_ = "x", y_ = "y" )PATH0(x, ...) ## Default S3 method: PATH0(x, ...) ## S3 method for class 'PATH0' PATH0(x, ...) PATH0_from_df( x, ..., path_ = "path_", object_ = "object_", subobject_ = "subobject_", x_ = "x", y_ = "y" )
x |
data frame with at least x, y columns |
... |
ignored |
path_ |
path identifier, these should identify individual paths |
object_ |
object identifier (like group in ggplot) |
subobject_ |
subobject identifier (like polygon_id with multipolygons in sfheaders) |
x_ |
optional name for x column (assumed to be x) |
y_ |
optional name for x column (assumed to be y) |
This function exists as a special-case for non-format input for PATH0(). It's expected there
are columns x, y, and optionally object_, subobject_, and path_. These correspond to
names in sfheaders, multipolygon_id, polygon_id, and linestring_id. (subobject is optional if
not multipolygon).
PATH0 model with tables 'object' and 'vertex'
(p <- PATH0(minimal_mesh)) p$object$topology_ PATH0_from_df(data.frame(x = runif(10), y = runif(10)))(p <- PATH0(minimal_mesh)) p$object$topology_ PATH0_from_df(data.frame(x = runif(10), y = runif(10)))
Basic edge plot, all the standard base graphics facilities for line segments are available.
## S3 method for class 'SC' plot(x, ..., add = FALSE) ## S3 method for class 'SC0' plot(x, ..., add = FALSE)## S3 method for class 'SC' plot(x, ..., add = FALSE) ## S3 method for class 'SC0' plot(x, ..., add = FALSE)
x |
sc object |
... |
arguments passed to lower level plotting functions |
add |
if |
The 'col' argument is passed directly to segments() or polypath() as needed, in the usual
one-to-one or recycling way.
Graphical parameters are not able to be passed to the initial plot setup, but a plot can be set up and then added to with this method.
A simple set of sf neighbouring polygons, with redundant vertices created
from polygonizing a raster.
arc <- ARC(polymesh) plot(arc) sc <- SC(polymesh) plot(sc)arc <- ARC(polymesh) plot(arc) sc <- SC(polymesh) plot(sc)
Print a silicate model.
## S3 method for class 'sc' print(x, ...)## S3 method for class 'sc' print(x, ...)
x |
object inheriting from 'sc' class |
... |
ignore currently |
Simple summary of type and content of a silicate model.
print(TRI(minimal_mesh)) print(SC(minimal_mesh)) print(PATH(minimal_mesh)) print(SC(TRI(minimal_mesh))) print(ARC(minimal_mesh)) print(SC0(minimal_mesh))print(TRI(minimal_mesh)) print(SC(minimal_mesh)) print(PATH(minimal_mesh)) print(SC(TRI(minimal_mesh))) print(ARC(minimal_mesh)) print(SC0(minimal_mesh))
The universal model SC is coordinates and binary relations between
pairs of coordinates. This is purely an edge (or segment) model, with all
higher level structures recorded as groupings of edges.
SC(x, ...) ## Default S3 method: SC(x, ...) ## S3 method for class 'TRI' SC(x, ...) ## S3 method for class 'pslg' SC(x, ...)SC(x, ...) ## Default S3 method: SC(x, ...) ## S3 method for class 'TRI' SC(x, ...) ## S3 method for class 'pslg' SC(x, ...)
x |
input model |
... |
arguments passed to methods |
SC model with tables 'object', 'object_link_edge', 'edge', and 'vertex'
## we can produce a high quality triangulation from a low quality one ## see how the TRI edges are maintained (we can't yet filter out holes from DEL) tri <- TRI(minimal_mesh) plot(tri) plot(SC(tri))## we can produce a high quality triangulation from a low quality one ## see how the TRI edges are maintained (we can't yet filter out holes from DEL) tri <- TRI(minimal_mesh) plot(tri) plot(SC(tri))
Return a label and vertex count of each arc.
sc_arc(x, ...) ## Default S3 method: sc_arc(x, ...) ## S3 method for class 'ARC' sc_arc(x, ...)sc_arc(x, ...) ## Default S3 method: sc_arc(x, ...) ## S3 method for class 'ARC' sc_arc(x, ...)
x |
input object |
... |
arguments for methods |
Arcs are unbranched paths within the line segment graph. Nodes are the vertices where three or more arcs meet.
As with the PATH and SC models the arc id values will only be relevant when
those entities are identified and labelled. Running sc_arc on a simple features model (for example) will
identify them and return a summary, but without having any record of what they refer to. Use ARC(x) first to
work with models that don't included labels.
a data frame with only the identities of the shared boundaries (arcs)
sc_arc(minimal_mesh) ARC(minimal_mesh)[["arc"]] arc <- ARC(minimal_mesh) plot(arc) points(arc$vertex[match(sc_node(arc)$vertex_, arc$vertex$vertex_), c("x_", "y_")]) arc <- ARC(polymesh) plot(arc) title("arcs and nodes") points(arc$vertex[match(sc_node(arc)$vertex_, arc$vertex$vertex_), c("x_", "y_")])sc_arc(minimal_mesh) ARC(minimal_mesh)[["arc"]] arc <- ARC(minimal_mesh) plot(arc) points(arc$vertex[match(sc_node(arc)$vertex_, arc$vertex$vertex_), c("x_", "y_")]) arc <- ARC(polymesh) plot(arc) title("arcs and nodes") points(arc$vertex[match(sc_node(arc)$vertex_, arc$vertex$vertex_), c("x_", "y_")])
Simple set of colours for discrete palette.
sc_colours(x = 16, ..., viridis = FALSE)sc_colours(x = 16, ..., viridis = FALSE)
x |
number of colours to generate |
... |
currently ignored |
viridis |
use viridis, TRUE or FALSE |
vector of colours
sc_colours(10)sc_colours(10)
Collect all instances of all coordinates in one table. This complementary to the sc_path of
an object, since the number of coordinates per path gives a structural mapping into this set.
Collect all coordinates in one table, the path_link_vertex table has the information about the original grouping.
sc_coord(x, ...) ## S3 method for class 'list' sc_coord(x, ...) ## Default S3 method: sc_coord(x, ...) ## S3 method for class 'matrix' sc_coord(x, ...) ## S3 method for class 'ARC' sc_coord(x, ...) ## S3 method for class 'PATH' sc_coord(x, ...) ## S3 method for class 'TRI' sc_coord(x, ...) ## S3 method for class 'PATH0' sc_coord(x, ...) ## S3 method for class 'SC0' sc_coord(x, ...) ## S3 method for class 'SC' sc_coord(x, ...) ## S3 method for class 'sf' sc_coord(x, ...) ## S3 method for class 'sfc' sc_coord(x, ...) ## S3 method for class 'pslg' sc_coord(x, ...) ## S3 method for class 'MULTIPOLYGON' sc_coord(x, ...) ## S3 method for class 'POLYGON' sc_coord(x, ...) ## S3 method for class 'MULTILINESTRING' sc_coord(x, ...) ## S3 method for class 'LINESTRING' sc_coord(x, ...) ## S3 method for class 'MULTIPOINT' sc_coord(x, ...) ## S3 method for class 'POINT' sc_coord(x, ...) ## S3 method for class 'Spatial' sc_coord(x, ...) ## S3 method for class 'Polygons' sc_coord(x, ...) ## S3 method for class 'Lines' sc_coord(x, ...)sc_coord(x, ...) ## S3 method for class 'list' sc_coord(x, ...) ## Default S3 method: sc_coord(x, ...) ## S3 method for class 'matrix' sc_coord(x, ...) ## S3 method for class 'ARC' sc_coord(x, ...) ## S3 method for class 'PATH' sc_coord(x, ...) ## S3 method for class 'TRI' sc_coord(x, ...) ## S3 method for class 'PATH0' sc_coord(x, ...) ## S3 method for class 'SC0' sc_coord(x, ...) ## S3 method for class 'SC' sc_coord(x, ...) ## S3 method for class 'sf' sc_coord(x, ...) ## S3 method for class 'sfc' sc_coord(x, ...) ## S3 method for class 'pslg' sc_coord(x, ...) ## S3 method for class 'MULTIPOLYGON' sc_coord(x, ...) ## S3 method for class 'POLYGON' sc_coord(x, ...) ## S3 method for class 'MULTILINESTRING' sc_coord(x, ...) ## S3 method for class 'LINESTRING' sc_coord(x, ...) ## S3 method for class 'MULTIPOINT' sc_coord(x, ...) ## S3 method for class 'POINT' sc_coord(x, ...) ## S3 method for class 'Spatial' sc_coord(x, ...) ## S3 method for class 'Polygons' sc_coord(x, ...) ## S3 method for class 'Lines' sc_coord(x, ...)
x |
input model |
... |
arguments passed to methods |
data frame of all the coordinates in the order they occur
sc_path for the central part of the model, sc_object for
the features, and PATH for the full model.
sc_path for the central part of the model, sc_object for
the features, and PATH for the full model.
sc_coord(minimal_mesh) sc_coord(SC(minimal_mesh))sc_coord(minimal_mesh) sc_coord(SC(minimal_mesh))
Simple binary relationships, a primitive composed of two vertices.
sc_edge(x, ...) ## Default S3 method: sc_edge(x, ...) ## S3 method for class 'PATH' sc_edge(x, ...) sc_start(x, ...) ## S3 method for class 'SC' sc_start(x, ...) ## S3 method for class 'SC0' sc_start(x, ...) ## S3 method for class 'PATH' sc_start(x, ...) ## S3 method for class 'PATH' sc_end(x, ...) ## S3 method for class 'PATH0' sc_start(x, ...) ## S3 method for class 'PATH0' sc_end(x, ...) ## S3 method for class 'ARC' sc_start(x, ...) ## S3 method for class 'TRI' sc_start(x, ...) sc_end(x, ...) ## S3 method for class 'SC' sc_end(x, ...) ## S3 method for class 'SC0' sc_end(x, ...) ## S3 method for class 'ARC' sc_end(x, ...) ## S3 method for class 'TRI' sc_end(x, ...)sc_edge(x, ...) ## Default S3 method: sc_edge(x, ...) ## S3 method for class 'PATH' sc_edge(x, ...) sc_start(x, ...) ## S3 method for class 'SC' sc_start(x, ...) ## S3 method for class 'SC0' sc_start(x, ...) ## S3 method for class 'PATH' sc_start(x, ...) ## S3 method for class 'PATH' sc_end(x, ...) ## S3 method for class 'PATH0' sc_start(x, ...) ## S3 method for class 'PATH0' sc_end(x, ...) ## S3 method for class 'ARC' sc_start(x, ...) ## S3 method for class 'TRI' sc_start(x, ...) sc_end(x, ...) ## S3 method for class 'SC' sc_end(x, ...) ## S3 method for class 'SC0' sc_end(x, ...) ## S3 method for class 'ARC' sc_end(x, ...) ## S3 method for class 'TRI' sc_end(x, ...)
x |
input object |
... |
arguments for methods |
Edges are unique, undirected line segments. Compare to sc_segment which refers to all
instances of edges.
sc_start and sc_end are convenience functions that provide the obvious
start and end coordinates by joining on the appropriate edge vertex label, .vx0
or .vx1. Currently this returns the ordered segments, along with their unique (unordered) edge_, as
well as unique segment, a object_ labels.
data frame of edge identity, or start/end coordinates
Nodes are the vertices in the graph that are shared by "arcs".
sc_node(x, ...) ## S3 method for class 'SC' sc_node(x, ...) ## S3 method for class 'SC0' sc_node(x, ...) ## Default S3 method: sc_node(x, ...) ## S3 method for class 'PATH' sc_node(x, ...) ## S3 method for class 'ARC' sc_node(x, ...)sc_node(x, ...) ## S3 method for class 'SC' sc_node(x, ...) ## S3 method for class 'SC0' sc_node(x, ...) ## Default S3 method: sc_node(x, ...) ## S3 method for class 'PATH' sc_node(x, ...) ## S3 method for class 'ARC' sc_node(x, ...)
x |
input object |
... |
arguments for methods |
data frame of the nodes
sc_node(ARC(minimal_mesh)) sc <- SC(routes) library(dplyr) plot(sc) sc_node(sc) %>% inner_join(sc$vertex) %>% select(x_, y_) %>% points()sc_node(ARC(minimal_mesh)) sc <- SC(routes) library(dplyr) plot(sc) sc_node(sc) %>% inner_join(sc$vertex) %>% select(x_, y_) %>% points()
The objects are the front end entities, the usual "GIS contract" objects, or features.
The objects are the front end entities, the usual "GIS contract" objects, the features.
sc_object(x, ...) ## Default S3 method: sc_object(x, ...) ## S3 method for class 'sf' sc_object(x, ...) ## S3 method for class 'sfc' sc_object(x, ...) ## S3 method for class 'TRI' sc_object(x, ...)sc_object(x, ...) ## Default S3 method: sc_object(x, ...) ## S3 method for class 'sf' sc_object(x, ...) ## S3 method for class 'sfc' sc_object(x, ...) ## S3 method for class 'TRI' sc_object(x, ...)
x |
input object |
... |
arguments passed to methods |
data frame of the object values
sc_coord for the coordinates part of the model, sc_path for
the central part of the model, and PATH for the full model.
sc_coord for the coordinates part of the model, sc_path for
the central part of the model, and PATH for the full model.
sc_object(minimal_mesh) sc_object(SC0(minimal_mesh))sc_object(minimal_mesh) sc_object(SC0(minimal_mesh))
Start in the middle, and build the 'path-link-vertex' table.
Paths.
sc_path(x, ...) ## S3 method for class 'list' sc_path(x, ids = NULL, ...) ## Default S3 method: sc_path(x, ...) ## S3 method for class 'PATH' sc_path(x, ...) ## S3 method for class 'sfc_TIN' sc_path(x, ...) ## S3 method for class 'PATH0' sc_path(x, ...) ## S3 method for class 'ARC' sc_path(x, ...) ## S3 method for class 'SC' sc_path(x, ...) ## S3 method for class 'SC0' sc_path(x, ...) ## S3 method for class 'matrix' sc_path(x, ...) ## S3 method for class 'sf' sc_path(x, ids = NULL, ...) ## S3 method for class 'sfc' sc_path(x, ids = NULL, ...) ## S3 method for class 'MULTIPOLYGON' sc_path(x, ...) ## S3 method for class 'POLYGON' sc_path(x, ...) ## S3 method for class 'LINESTRING' sc_path(x, ...) ## S3 method for class 'MULTILINESTRING' sc_path(x, ...) ## S3 method for class 'POINT' sc_path(x, ...) ## S3 method for class 'MULTIPOINT' sc_path(x, ...) ## S3 method for class 'GEOMETRYCOLLECTION' sc_path(x, ...) ## S3 method for class 'Spatial' sc_path(x, ids = NULL, ...)sc_path(x, ...) ## S3 method for class 'list' sc_path(x, ids = NULL, ...) ## Default S3 method: sc_path(x, ...) ## S3 method for class 'PATH' sc_path(x, ...) ## S3 method for class 'sfc_TIN' sc_path(x, ...) ## S3 method for class 'PATH0' sc_path(x, ...) ## S3 method for class 'ARC' sc_path(x, ...) ## S3 method for class 'SC' sc_path(x, ...) ## S3 method for class 'SC0' sc_path(x, ...) ## S3 method for class 'matrix' sc_path(x, ...) ## S3 method for class 'sf' sc_path(x, ids = NULL, ...) ## S3 method for class 'sfc' sc_path(x, ids = NULL, ...) ## S3 method for class 'MULTIPOLYGON' sc_path(x, ...) ## S3 method for class 'POLYGON' sc_path(x, ...) ## S3 method for class 'LINESTRING' sc_path(x, ...) ## S3 method for class 'MULTILINESTRING' sc_path(x, ...) ## S3 method for class 'POINT' sc_path(x, ...) ## S3 method for class 'MULTIPOINT' sc_path(x, ...) ## S3 method for class 'GEOMETRYCOLLECTION' sc_path(x, ...) ## S3 method for class 'Spatial' sc_path(x, ids = NULL, ...)
x |
input object |
... |
arguments passed to methods |
ids |
object id, one for each object in the |
Paths have properties of their type, their number of vertices, their geometric dimension and which object they occur in.
data frame of path identity and properties
sc_coord for the coordinates part of the model, sc_object for
the features, and PATH for the full model.
sc_path(minimal_mesh) sc_path(PATH(minimal_mesh)) sc_path(sfzoo$multipolygon) sc_path(sfzoo$polygon) sc_path(sfzoo$linestring) sc_path(sfzoo$multilinestring) sc_path(sfzoo$point) sc_path(sfzoo$multipoint) sc_path(sfzoo$multipoint)sc_path(minimal_mesh) sc_path(PATH(minimal_mesh)) sc_path(sfzoo$multipolygon) sc_path(sfzoo$polygon) sc_path(sfzoo$linestring) sc_path(sfzoo$multilinestring) sc_path(sfzoo$point) sc_path(sfzoo$multipoint) sc_path(sfzoo$multipoint)
Given a 'PATH“ model decompose to 1-dimensional primitives (or 0-dimensional).
sc_segment(x, ...) ## Default S3 method: sc_segment(x, ...) ## S3 method for class 'PATH' sc_segment(x, ...)sc_segment(x, ...) ## Default S3 method: sc_segment(x, ...) ## S3 method for class 'PATH' sc_segment(x, ...)
x |
input object |
... |
arguments passed to methods |
data frame of the segments, each occurence of an edge and its order
sc_segment(SC(minimal_mesh))sc_segment(SC(minimal_mesh))
Find unique labels for entities, or create them if not present.
sc_uid(x, ..., uid_nchar = NULL)sc_uid(x, ..., uid_nchar = NULL)
x |
number of unique IDs to generate |
... |
reserved for future use |
uid_nchar |
number of raw characters to paste as a uuid, default is 6 (only if silicate.uid.type is "uuid", see Details) |
If 'integers' default we generate sequential integers, it's assumed that all IDs are created at one time, we are not adding to an existing set. Code that adds IDs should find the largest existing ID and offset these by that value.
Using 'silicate.uid.type="uuid"' is the default. Using 'silicate.uid.type="integer"' is considered experimental.
By default UIDs are a mix of letters, LETTERS and digits of length getOption("silicate.uid.size") which defaults to 6.
See ids package for random_id used if option 'silicate.uid.type="uuid"'.
vector of unique id values for elements in the input
sc_uid(data.frame(1:10))sc_uid(data.frame(1:10))
Extract unique vertices
sc_vertex(x, ...) ## Default S3 method: sc_vertex(x, ...) ## S3 method for class 'SC' sc_vertex(x, ...) ## S3 method for class 'SC0' sc_vertex(x, ...) ## S3 method for class 'ARC' sc_vertex(x, ...) ## S3 method for class 'TRI' sc_vertex(x, ...) ## S3 method for class 'TRI0' sc_vertex(x, ...) ## S3 method for class 'PATH' sc_vertex(x, ...) ## S3 method for class 'PATH0' sc_vertex(x, ...) ## S3 method for class 'pslg' sc_vertex(x, ...)sc_vertex(x, ...) ## Default S3 method: sc_vertex(x, ...) ## S3 method for class 'SC' sc_vertex(x, ...) ## S3 method for class 'SC0' sc_vertex(x, ...) ## S3 method for class 'ARC' sc_vertex(x, ...) ## S3 method for class 'TRI' sc_vertex(x, ...) ## S3 method for class 'TRI0' sc_vertex(x, ...) ## S3 method for class 'PATH' sc_vertex(x, ...) ## S3 method for class 'PATH0' sc_vertex(x, ...) ## S3 method for class 'pslg' sc_vertex(x, ...)
x |
model |
... |
passed to methods |
data frame of only the unique coordinates
sc_vertex(minimal_mesh) sc_vertex(SC0(minimal_mesh))sc_vertex(minimal_mesh) sc_vertex(SC0(minimal_mesh))
SC0 is the simplest and most general of all silicate models. Composed of
an object and vertex table linked by nested vertex-index pairs.
SC0(x, ...) ## Default S3 method: SC0(x, ...) ## S3 method for class 'pslg' SC0(x, ...)SC0(x, ...) ## Default S3 method: SC0(x, ...) ## S3 method for class 'pslg' SC0(x, ...)
x |
an object understood by silicate |
... |
reserved for methods |
SC0 model with tables 'object' and 'vertex'
SC0(minimal_mesh) SC0(minimal_mesh)SC0(minimal_mesh) SC0(minimal_mesh)
Basic examples of each type of simple feature geometry. sfzoo is a list
with each of point, multipoin, linestring, multilinestring, polygon and
multipolygon. sfgc is a GEOMETRYCOLLECTION of all the types in sfzoo.
lapply(sfzoo, sc_coord) lapply(sfzoo, sc_path) ## unsure how useful this is ... sc_path(sfgc)lapply(sfzoo, sc_coord) lapply(sfzoo, sc_path) ## unsure how useful this is ... sc_path(sfgc)
Decomposes spatial data (of various formats) into simpler forms, including
paths, triangles or segments. A development tool for exploring the underlying
structures of spatial data, and for converting it to something else. The models
PATH(), TRI(), SC() and ARC() provide relational tables of all underlying
entities, and more specialist versions PATH0(), TRI0() and SC0() provide more
efficient topological representations of polygons or lines.
TRI creates a constrained triangulation using 'ear-cutting', or 'ear-clipping' of
polygons. The model is a 'relational' form in that the underlying tables are
linked implicitly by unique identifiers.
Ear-cutting is inherently path-based, so this model is only available for
path-based structures, like simple features, PATH(), PATH0() and ARC().
TRI(x, ...) ## S3 method for class 'sfc_GEOMETRYCOLLECTION' TRI(x, ...) ## S3 method for class 'TRI' plot(x, ..., add = FALSE)TRI(x, ...) ## S3 method for class 'sfc_GEOMETRYCOLLECTION' TRI(x, ...) ## S3 method for class 'TRI' plot(x, ..., add = FALSE)
x |
object understood by silicate (sf, sp, a silicate model, etc.) |
... |
current unused |
add |
logical create new plot (default), or add to existing |
TRI model with tables 'object', 'triangle', 'vertex'
tri <- TRI(minimal_mesh) plot(tri)tri <- TRI(minimal_mesh) plot(tri)
Input is x,y matrix in triplets, with 3 rows per triangle vertex.
tri_area(x)tri_area(x)
x |
matrix of triangle coordinates |
numeric, area of triangles
pts <- structure(c(5L, 3L, 1L, 4L, 4L, 8L, 6L, 9L), .Dim = c(4L, 2L)) tri <- c(2, 1, 3, 2, 4, 1) (a <- tri_area(pts[tri, ])) plot(pts) polygon(pts[head(as.vector(rbind(matrix(tri, nrow = 3), NA)), -1), ]) text(tapply(pts[tri,1], rep(1:2, each = 3), mean), tapply(pts[tri,2], rep(1:2, each = 3), mean), labels = sprintf("area: %0.1f", a))pts <- structure(c(5L, 3L, 1L, 4L, 4L, 8L, 6L, 9L), .Dim = c(4L, 2L)) tri <- c(2, 1, 3, 2, 4, 1) (a <- tri_area(pts[tri, ])) plot(pts) polygon(pts[head(as.vector(rbind(matrix(tri, nrow = 3), NA)), -1), ]) text(tapply(pts[tri,1], rep(1:2, each = 3), mean), tapply(pts[tri,2], rep(1:2, each = 3), mean), labels = sprintf("area: %0.1f", a))
TRI0 creates a constrained triangulation using 'ear-cutting', or 'ear-clipping' of
polygons. It is a 'structural' form, a denser storage mode than 'relational'
as used by TRI(), we trade some generality for size and speed.
TRI0(x, ...) ## Default S3 method: TRI0(x, ...) ## S3 method for class 'mesh3d' TRI0(x, ...) ## S3 method for class 'TRI0' TRI0(x, ...) ## S3 method for class 'sfc_TIN' TRI0(x, ...) ## S3 method for class 'TRI' TRI0(x, ...) ## S3 method for class 'PATH0' TRI0(x, ...) ## S3 method for class 'PATH' TRI0(x, ...) ## S3 method for class 'sf' TRI0(x, ...) ## S3 method for class 'sfc_GEOMETRYCOLLECTION' TRI0(x, ...)TRI0(x, ...) ## Default S3 method: TRI0(x, ...) ## S3 method for class 'mesh3d' TRI0(x, ...) ## S3 method for class 'TRI0' TRI0(x, ...) ## S3 method for class 'sfc_TIN' TRI0(x, ...) ## S3 method for class 'TRI' TRI0(x, ...) ## S3 method for class 'PATH0' TRI0(x, ...) ## S3 method for class 'PATH' TRI0(x, ...) ## S3 method for class 'sf' TRI0(x, ...) ## S3 method for class 'sfc_GEOMETRYCOLLECTION' TRI0(x, ...)
x |
object understood by silicate (sf, sp, a silicate model, etc.) |
... |
currently unused |
TRI0 is suitable for simple conversion to other mesh forms. See the examples for plotting and (in commented code) conversion to rgl's 'mesh3d'.
'Structural' means that the model does not store relational
IDs between tables, the vertex indexing is stored as a nested
list of data frames in the 'object' table. Unlike TRI() we
cannot arbitrarily rearrange the order or remove content
of the underlying tables, without updating the vertex indexes
stored for each object.
Ear-cutting is inherently path-based, so this model is only available for
path-based structures, like simple features, PATH(), PATH0() and ARC().
There is limited support for simple features GEOMETRYCOLLECTION, in short if the GC is composed purely of POLYGON type with 4 coordinates each this is assumed to be a collection of triangles and is converted directly without any triangulation performed. GEOMETRYCOLLECTION of any other form is not supported.
TRI0 model with tables 'object', 'vertex'
TRI
tri <- TRI0(minimal_mesh) print(tri) plot(tri) # obtain the vertices and indices in raw form ## idx is the triplets of row numbers in tri$vertex idx <- do.call(rbind, sc_object(tri)$topology_) idx <- as.matrix(idx[c(".vx0", ".vx1", ".vx2")]) ## vert is the vertices x_, y_, ... vert <- as.matrix(sc_vertex(tri)) ## now we can plot with generic tools plot(vert) polygon(vert[t(cbind(idx, NA)), ]) ## or create other structures like rgl's mesh3d ## (see hypertidy/anglr for in-dev helpers) ## rgl::tmesh3d(t(cbind(vert, 1, 1)), t(idx), ## material = list(color = c("firebrick", "black", "grey", "blue")), ## meshColor = "faces")tri <- TRI0(minimal_mesh) print(tri) plot(tri) # obtain the vertices and indices in raw form ## idx is the triplets of row numbers in tri$vertex idx <- do.call(rbind, sc_object(tri)$topology_) idx <- as.matrix(idx[c(".vx0", ".vx1", ".vx2")]) ## vert is the vertices x_, y_, ... vert <- as.matrix(sc_vertex(tri)) ## now we can plot with generic tools plot(vert) polygon(vert[t(cbind(idx, NA)), ]) ## or create other structures like rgl's mesh3d ## (see hypertidy/anglr for in-dev helpers) ## rgl::tmesh3d(t(cbind(vert, 1, 1)), t(idx), ## material = list(color = c("firebrick", "black", "grey", "blue")), ## meshColor = "faces")