From 2aef4afa6add320a8f0010bcdec45b0b8e42b4e3 Mon Sep 17 00:00:00 2001
From: Simon Cruanes <simon.cruanes.2007@m4x.org>
Date: Fri, 8 Mar 2013 15:11:26 +0100
Subject: [PATCH] sort, uniq, group and sort_uniq combinators implemented

---
 sequence.ml  | 42 ++++++++++++++++++++++++++++++++++++++++++
 sequence.mli | 13 +++++++++++++
 2 files changed, 55 insertions(+)

diff --git a/sequence.ml b/sequence.ml
index f46558e..0da84f1 100644
--- a/sequence.ml
+++ b/sequence.ml
@@ -211,6 +211,48 @@ let persistent (seq : 'a t) : 'a t =
   let l = MList.of_seq seq in
   from_iter (fun k -> MList.iter k l)
 
+(** Sort the sequence. Eager, O(n) ram and O(n ln(n)) time. *)
+let sort ?(cmp=Pervasives.compare) seq =
+  (* use an intermediate list, then sort the list *)
+  let l = fold (fun l x -> x::l) [] seq in
+  let l = List.fast_sort cmp l in
+  fun k -> List.iter k l
+
+(** Group equal consecutive elements. *)
+let group ?(eq=fun x y -> x = y) seq =
+  fun k ->
+    let cur = ref [] in
+    seq (fun x ->
+      match !cur with
+      | [] -> cur := [x]
+      | (y::_) as l when eq x y ->
+        cur := x::l  (* [x] belongs to the group *)
+      | (_::_) as l ->
+        k l; (* yield group, and start another one *)
+        cur := [x]);
+    (* last list *)
+    if !cur <> [] then k !cur
+  
+(** Remove consecutive duplicate elements. Basically this is
+    like [fun seq -> map List.hd (group seq)]. *)
+let uniq ?(eq=fun x y -> x = y) seq =
+  fun k ->
+    let has_prev = ref false
+    and prev = ref (Obj.magic 0) in  (* avoid option type, costly *)
+    seq (fun x ->
+      if !has_prev && eq !prev x
+        then ()  (* duplicate *)
+        else begin
+          has_prev := true;
+          prev := x;
+          k x
+        end)
+
+(** Sort the sequence and remove duplicates. Eager, same as [sort] *)
+let sort_uniq ?(cmp=Pervasives.compare) seq =
+  let seq' = sort ~cmp seq in
+  uniq ~eq:(fun x y -> cmp x y = 0) seq'
+
 (** Cartesian product of the sequences. *)
 let product outer inner =
   let outer = persistent outer in
diff --git a/sequence.mli b/sequence.mli
index a1da7bc..646262f 100644
--- a/sequence.mli
+++ b/sequence.mli
@@ -112,6 +112,19 @@ val persistent : 'a t -> 'a t
   (** Iterate on the sequence, storing elements in a data structure.
       The resulting sequence can be iterated on as many times as needed. *)
 
+val sort : ?cmp:('a -> 'a -> int) -> 'a t -> 'a t
+  (** Sort the sequence. Eager, O(n) ram and O(n ln(n)) time. *)
+
+val sort_uniq : ?cmp:('a -> 'a -> int) -> 'a t -> 'a t
+  (** Sort the sequence and remove duplicates. Eager, same as [sort] *)
+
+val group : ?eq:('a -> 'a -> bool) -> 'a t -> 'a list t
+  (** Group equal consecutive elements. *)
+
+val uniq : ?eq:('a -> 'a -> bool) -> 'a t -> 'a t
+  (** Remove consecutive duplicate elements. Basically this is
+      like [fun seq -> map List.hd (group seq)]. *)
+
 val product : 'a t -> 'b t -> ('a * 'b) t
   (** Cartesian product of the sequences. The first one is transformed
       by calling [persistent] on it, so that it can be traversed