Implement the ranked automata evaluation to guarantee a O(|D|x|Q|)

[tatoo.git] / src / ata.mli

diff --git a/src/ata.mli b/src/ata.mli
index eada8d0..2aa27b3 100644 (file)
--- a/src/ata.mli
+++ b/src/ata.mli
@@ -21,6 +21,21 @@ type move = [ `First_child
             | `Parent
             | `Previous_sibling
             | `Stay ]
+
+module Move :
+  sig
+    type t = move
+    type 'a table
+    val create_table : 'a -> 'a table
+    val get : 'a table -> t -> 'a
+    val set : 'a table -> t -> 'a -> unit
+    val iter : (t -> 'a -> unit) -> 'a table -> unit
+    val fold : (t -> 'a -> 'b -> 'b) -> 'a table -> 'b -> 'b
+    val for_all : (t -> 'a -> bool) -> 'a table -> bool
+    val for_all2 : (t -> 'a -> 'b -> bool) -> 'a table -> 'b table -> bool
+    val exists : (t -> 'a -> bool) -> 'a table -> bool
+  end
+
 (** Type of moves an automaton can perform *)
 
 type predicate =
@@ -34,7 +49,7 @@ type predicate =
 
 module Atom : sig
   include Hcons.S with type data = predicate
-  include Common_sig.Printable with type t:= t
+  include Common_sig.Printable with type t := t
 end
 (** Module representing atoms of Boolean formulae, which are simply hashconsed [predicate]s *)
 
@@ -65,6 +80,7 @@ module Formula :
           particular kind *)
     val get_states : t -> StateSet.t
       (** [get_state f] retrieves all the states occuring in [move] predicates in [f] *)
+    val get_states_by_move : t -> StateSet.t Move.table
   end
 (** Modules representing the Boolean formulae used in transitions *)
 
@@ -97,12 +113,26 @@ val get_starting_states : t -> StateSet.t
 val get_selecting_states : t -> StateSet.t
 (** return the set of selecting states of the automaton *)
 
+val get_states_by_rank : t -> StateSet.t array
+(** return an array of states ordered by ranks.
+*)
+
+val get_max_rank : t -> int
+(** return the maximal rank of a state in the automaton, that is the
+    maximum number of runs needed to fuly evaluate the automaton.
+*)
+
 val get_trans : t -> QNameSet.elt -> StateSet.t -> TransList.t
-(** [get_trans auto l q] return the list of transitions taken by [auto]
+(** [get_trans auto l q] returns the list of transitions taken by [auto]
     for label [l] in state [q]. Takes time proportional to the number of
     transitions in the automaton.
  *)
 
+val get_form : t -> QNameSet.elt -> State.t -> Formula.t
+(** [get_form auto l q] returns a single formula for label [l] in state [q].
+    Takes time proportional to the number of transitions in the automaton.
+ *)
+
 val print : Format.formatter -> t -> unit
 (** Pretty printing of the automaton *)
 
@@ -121,6 +151,26 @@ val merge : t -> t -> t
     in parallel
 *)
 
+val union : t -> t -> t
+(** [union a a'] creates a new automaton [a''] that selects node
+    selected by either [a] or [a']
+*)
+
+val inter : t -> t -> t
+(** [inter a a'] creates a new automaton [a''] that selects node
+    selected by both [a] and [a']
+*)
+
+val neg : t -> t
+(** [neg a] creates a new automaton [a'] that selects the nodes not
+    selected by [a]
+*)
+
+val diff : t -> t -> t
+(** [diff a a'] creates a new automaton [a''] that select nodes selected
+    by [a] but not selected by [a']
+*)
+
 module Builder :
 sig
   type auto = t