R/context.R

Defines functions with_graph .free_graph_context .register_graph_context .E .N .G expect_edges expect_nodes

Documented in .E .G .N .register_graph_context with_graph

#' @importFrom R6 R6Class
ContextBuilder <- R6Class(
  'ContextBuilder',
  public = list(
    set = function(graph) {
      if (!is.tbl_graph(graph)) cli::cli_abort('{.arg graph} must be a {.cls tbl_graph} object')
      private$context <- c(private$context, list(graph))
      invisible(self)
    },
    clear = function() {
      private$context <- private$context[-length(private$context)]
    },
    alive = function() {
      length(private$context) != 0
    },
    graph = function() {
      private$check()
      private$context[[length(private$context)]]
    },
    nodes = function() {
      as_tibble(self$graph(), active = 'nodes')
    },
    edges = function() {
      as_tibble(self$graph(), active = 'edges')
    },
    active = function() {
      private$check()
      active(self$graph())
    },
    free = function() {
      private$FREE != 0 || inherits(self$graph(), 'free_context_tbl_graph')
    },
    force_free = function() {
      private$FREE <- private$FREE + 1
    },
    force_unfree = function() {
      private$FREE <- private$FREE - 1
    }
  ),
  private = list(
    context = list(),
    FREE = 0,
    check = function() {
      if (!self$alive()) {
        cli::cli_abort('This function should not be called directly')
      }
    }
  )
)
#' @export
.graph_context <- ContextBuilder$new()
expect_nodes <- function() {
  if (!.graph_context$free() && .graph_context$active() != 'nodes') {
    cli::cli_abort('This call requires nodes to be active', call. = FALSE)
  }
}
expect_edges <- function() {
  if (!.graph_context$free() && .graph_context$active() != 'edges') {
    cli::cli_abort('This call requires edges to be active', call. = FALSE)
  }
}

#' Access graph, nodes, and edges directly inside verbs
#'
#' These three functions makes it possible to directly access either the node
#' data, the edge data or the graph itself while computing inside verbs. It is
#' e.g. possible to add an attribute from the node data to the edges based on
#' the terminating nodes of the edge, or extract some statistics from the graph
#' itself to use in computations.
#'
#' @return Either a `tbl_graph` (`.G()`) or a `tibble` (`.N()`)
#'
#' @rdname context_accessors
#' @name context_accessors
#'
#' @examples
#'
#' # Get data from the nodes while computing for the edges
#' create_notable('bull') %>%
#'   activate(nodes) %>%
#'   mutate(centrality = centrality_power()) %>%
#'   activate(edges) %>%
#'   mutate(mean_centrality = (.N()$centrality[from] + .N()$centrality[to])/2)
NULL

#' @describeIn context_accessors Get the tbl_graph you're currently working on
#' @export
.G <- function() {
  .graph_context$graph()
}
#' @describeIn context_accessors Get the nodes data from the graph you're currently working on
#' @export
.N <- function() {
  .graph_context$nodes()
}
#' @describeIn context_accessors Get the edges data from the graph you're currently working on
#' @export
.E <- function() {
  .graph_context$edges()
}

#' Register a graph context for the duration of the current frame
#'
#' This function sets the provided graph to be the context for tidygraph
#' algorithms, such as e.g. [node_is_center()], for the duration of the current
#' environment. It automatically removes the graph once the environment exits.
#'
#' @param graph A `tbl_graph` object
#'
#' @param free Should the active state of the graph be ignored?
#'
#' @param env The environment where the context should be active
#'
#' @export
#' @keywords internal
.register_graph_context <- function(graph, free = FALSE, env = parent.frame()) {
  if (!is.tbl_graph(graph)) cli::cli_abort('{.arg graph} must be a {.cls tbl_graph} object')
  if (identical(env, .GlobalEnv)) {
    cli::cli_abort('A context cannot be registered to the global environment')
  }
  if (free) {
    class(graph) <- c('free_context_tbl_graph', class(graph))
  }
  .graph_context$set(graph)
  do.call(on.exit, alist(expr = .graph_context$clear(), add = TRUE), envir = env)
  invisible(NULL)
}
.free_graph_context <- function(env = parent.frame()) {
  if (identical(env, .GlobalEnv)) {
    cli::cli_abort('A context cannot be freed in the global environment')
  }
  .graph_context$force_free()
  do.call(on.exit, alist(expr = .graph_context$force_unfree(), add = TRUE), envir = env)
  invisible(NULL)
}
#' Evaluate a tidygraph algorithm in the context of a graph
#'
#' All tidygraph algorithms are meant to be called inside tidygraph verbs such
#' as `mutate()`, where the graph that is currently being worked on is known and
#' thus not needed as an argument to the function. In the off chance that you
#' want to use an algorithm outside of the tidygraph framework you can use
#' `with_graph()` to set the graph context temporarily while the algorithm is
#' being evaluated.
#'
#' @param graph The `tbl_graph` to use as context
#'
#' @param expr The expression to evaluate
#'
#' @return The value of `expr`
#'
#' @export
#'
#' @examples
#' gr <- play_erdos_renyi(10, 0.3)
#'
#' with_graph(gr, centrality_degree())
#'
with_graph <- function(graph, expr) {
  .register_graph_context(graph, free = TRUE)
  expr
}

Try the tidygraph package in your browser

Any scripts or data that you put into this service are public.

tidygraph documentation built on Feb. 16, 2023, 10:40 p.m.