Module Traverse.Dfs

Depth-first search

Classical big-step iterators

Parameters

module G : G

Signature

val iter : ?pre:(G.V.t -> unit) -> ?post:(G.V.t -> unit) -> G.t -> unit

iter pre post g visits all nodes of g in depth-first search, applying pre to each visited node before its successors, and post after them. Each node is visited exactly once. Not tail-recursive.

val prefix : (G.V.t -> unit) -> G.t -> unit

applies only a prefix function; note that this function is more efficient than iter and is tail-recursive.

val postfix : (G.V.t -> unit) -> G.t -> unit

applies only a postfix function. Not tail-recursive.

Same thing, but for a single connected component (only prefix_component is tail-recursive)

val iter_component : ?pre:(G.V.t -> unit) -> ?post:(G.V.t -> unit) -> G.t -> G.V.t -> unit
val prefix_component : (G.V.t -> unit) -> G.t -> G.V.t -> unit
val postfix_component : (G.V.t -> unit) -> G.t -> G.V.t -> unit

Classical folds

val fold : (G.V.t -> 'a -> 'a) -> 'a -> G.t -> 'a

The function is applied each time a node is reached for the first time, before idoterating over its successors. Tail-recursive.

val fold_component : (G.V.t -> 'a -> 'a) -> 'a -> G.t -> G.V.t -> 'a

Idem, but limited to a single root vertex.

Step-by-step iterator

This is a variant of the iterators above where you can move on step by step. The abstract type iterator represents the current state of the iteration. The step function returns the next state. In each state, function get returns the currently visited vertex. On the final state both get and step raises exception Exit.

Note: the iterator type is persistent (i.e. is not modified by the step function) and thus can be used in backtracking algorithms.

type iterator
val start : G.t -> iterator
val step : iterator -> iterator
val get : iterator -> G.V.t

Cycle detection

val has_cycle : G.t -> bool

has_cycle g checks for a cycle in g. Linear in time and space.