`Rtree`

Type of regular tree with nodes labelled by values of type 'a The implementation uses de Bruijn indices, so binding capture is avoided by the lift operator (see example below).

Note that it differs from standard regular trees by accepting vectors of vectors in nodes, which is useful for encoding disjunctive-conjunctive recursive trees such as inductive types. Standard regular trees can however easily be simulated by using singletons of vectors

Building trees

Build a node given a label and a vector of vectors of sons

`val mk_rec_calls : int -> 'a t array`

Build mutually recursive trees: X_1 = f_1(X_1,..,X_n) ... X_n = f_n(X_1,..,X_n) is obtained by the following pseudo-code let vx = mk_rec_calls n in let `|x_1;..;x_n|`

= mk_rec`|f_1(vx.(0),..,vx.(n-1);..;f_n(vx.(0),..,vx.(n-1))|`

First example: build rec X = a(X,Y) and Y = b(X,Y,Y) let `|vx;vy|`

= mk_rec_calls 2 in let `|x;y|`

= mk_rec `|mk_node a [|[|vx;vy|]|]; mk_node b [|[|vx;vy;vy|]|]|`

Another example: nested recursive trees rec Y = b(rec X = a(X,Y),Y,Y) let `|vy|`

= mk_rec_calls 1 in let `|vx|`

= mk_rec_calls 1 in let `|x|`

= mk_rec`|mk_node a [|[|vx;lift 1 vy|]|]|`

let `|y|`

= mk_rec`|mk_node b [|[|x;vy;vy|]|]|`

(note the lift so that Y links to the "rec Y" skipping the "rec X")

`lift k t`

increases of `k`

the free parameters of `t`

. Needed to avoid captures when a tree appears under `mk_rec`

`val is_node : 'a t -> bool`

`val dest_var : 'a t -> int * int`

dest_var is not needed for closed trees (i.e. with no free variable)

`val is_infinite : ('a -> 'a -> bool) -> 'a t -> bool`

Tells if a tree has an infinite branch. The first arg is a comparison used to detect already seen elements, hence loops

`Rtree.equiv eq eqlab t1 t2`

compares t1 t2 (top-down). If t1 and t2 are both nodes, `eqlab`

is called on their labels, in case of success deeper nodes are examined. In case of loop (detected via structural equality parametrized by `eq`

), then the comparison is successful.

`Rtree.equal eq t1 t2`

compares t1 and t2, first via physical equality, then by structural equality (using `eq`

on elements), then by logical equivalence `Rtree.equiv eq eq`

Iterators

`module Smart : sig ... end`