Compute the adjacency matrix from various types of objects.

If `x` is a tibble or a dataframe, `get_adjacency` assumes that it has been passed an edgelist. If `x` is a matrix, `get_adjacency` assumes that it has been passed an incidence matrix and constructs the corresponding adjacency matrix. If `x` is a different object, `get_adjacency` uses the appropriate method if it exists. Otherwise it tries to coerce `x` to a `data.frame`.

get_adjacency(
  x,
  select_cols = NULL,
  multiedge = NULL,
  aggr_expression = NULL,
  nodes = NULL,
  sparse = TRUE,
  drop_names = FALSE,
  directed = NULL,
  selfloops = NULL,
  edgelist = NULL,
  ...
)

# S3 method for tbl
get_adjacency(
  x,
  select_cols = NULL,
  multiedge = NULL,
  aggr_expression = NULL,
  nodes = NULL,
  sparse = TRUE,
  drop_names = FALSE,
  directed = NULL,
  selfloops = NULL,
  edgelist = NULL,
  ...
)

# S3 method for tbl_graph
get_adjacency(
  x,
  select_cols = NULL,
  multiedge = NULL,
  aggr_expression = NULL,
  nodes = NULL,
  sparse = TRUE,
  drop_names = FALSE,
  directed = NULL,
  selfloops = NULL,
  edgelist = NULL,
  ...
)

# S3 method for data.frame
get_adjacency(
  x,
  select_cols = NULL,
  multiedge = NULL,
  aggr_expression = NULL,
  nodes = NULL,
  sparse = TRUE,
  drop_names = FALSE,
  directed = NULL,
  selfloops = NULL,
  edgelist = NULL,
  ...
)

# S3 method for matrix
get_adjacency(
  x,
  select_cols = NULL,
  multiedge = NULL,
  aggr_expression = NULL,
  nodes = NULL,
  sparse = TRUE,
  drop_names = FALSE,
  directed = NULL,
  selfloops = NULL,
  edgelist = NULL,
  ...
)

# S3 method for dgTMatrix
get_adjacency(
  x,
  select_cols = NULL,
  multiedge = NULL,
  aggr_expression = NULL,
  nodes = NULL,
  sparse = TRUE,
  drop_names = FALSE,
  directed = NULL,
  selfloops = NULL,
  ...
)

# S3 method for dgCMatrix
get_adjacency(
  x,
  select_cols = NULL,
  multiedge = NULL,
  aggr_expression = NULL,
  nodes = NULL,
  sparse = TRUE,
  drop_names = FALSE,
  directed = NULL,
  selfloops = NULL,
  ...
)

# S3 method for sparseMatrix
get_adjacency(
  x,
  select_cols = NULL,
  multiedge = NULL,
  aggr_expression = NULL,
  nodes = NULL,
  sparse = TRUE,
  drop_names = FALSE,
  directed = NULL,
  selfloops = NULL,
  ...
)

# S3 method for default
get_adjacency(
  x,
  select_cols = NULL,
  multiedge = NULL,
  aggr_expression = NULL,
  nodes = NULL,
  sparse = TRUE,
  drop_names = FALSE,
  directed = NULL,
  selfloops = NULL,
  ...
)

# S3 method for igraph
get_adjacency(
  x,
  select_cols = NULL,
  multiedge = NULL,
  aggr_expression = NULL,
  nodes = NULL,
  sparse = TRUE,
  drop_names = FALSE,
  directed = NULL,
  selfloops = NULL,
  ...
)

Arguments

x	object from which generating the adjacency matrix
select_cols	optional vector of 3 (2 for multi-graphs) elements specifying which columns are the source,target, and attributes from which building the graph. Otherwise Column 1 is assumed to be the source, column 2 the target, column 3 the attribute. In the case of multi-graphs, the third element is not needed and the number of edges between each pair of vertices is computed according to 'aggr_expression'.
multiedge	boolean, are there multiple edges? defaults to FALSE.
aggr_expression	string, the expression used to compute the aggregated value in the adjacency matrix in the presence of multiedges. It defaults to 'dplyr::n()'. If `edge.list` contains other columns such as `weight`, examples of `aggr_expression` could be `"mean(weight)"` to report the average of the `weight` column of the multiedges, `"min(weight)"` to report the minimum of the `weight` column of the multiedges.
nodes	optional vector containing all node names in case disconnected nodes should be included.
sparse	boolean, return sparse matrix? default to TRUE
drop_names	boolean, drop names from matrix and return a vector of names together with the matrix. This option saves considerable memory when dealing with graphs with long node names.
directed	boolean, optional parameter, if FALSE, forces to undirected upper triangular adjacency matrix
selfloops	boolean, optional parameter, if ignores selfloops.
edgelist	optional boolean. If `x` is a matrix, should treat `x` as an edgelist? If not provided, tries to infer if `x` is an edgelist or an incidence matrix from its shape.
...	additional parameters to and from the main method. Currently not used.

Value

A transition matrix in the same format of `adj`.

Examples

el <- data.frame(from= c('a','b','b','c','d','d'),
                to  = c('b','c','d','a','b','a'),
                attr= c( 12, 6, 12 , 6 , 6 , 6 ))
adj <- get_adjacency(el)