From bd7a9ce0709c09e53cf1986f6cf2e3b4363dd2fb Mon Sep 17 00:00:00 2001
From: Simon Cruanes <simon.cruanes.2007@m4x.org>
Date: Thu, 3 Nov 2016 20:27:26 +0100
Subject: [PATCH] add `CCString.edit_distance`

---
 src/core/CCString.cppo.ml | 32 ++++++++++++++++++++++++++++++++
 src/core/CCString.mli     | 34 ++++++++++++++++++++++++++++++++++
 2 files changed, 66 insertions(+)

diff --git a/src/core/CCString.cppo.ml b/src/core/CCString.cppo.ml
index 55018995..719de108 100644
--- a/src/core/CCString.cppo.ml
+++ b/src/core/CCString.cppo.ml
@@ -407,6 +407,38 @@ let compare_versions a b =
   in
   cmp_rec (Split.gen_cpy ~by:"." a) (Split.gen_cpy ~by:"." b)
 
+let edit_distance s1 s2 =
+  if length s1 = 0
+  then length s2
+  else if length s2 = 0
+  then length s1
+  else if s1 = s2
+  then 0
+  else begin
+    (* distance vectors (v0=previous, v1=current) *)
+    let v0 = Array.make (length s2 + 1) 0 in
+    let v1 = Array.make (length s2 + 1) 0 in
+    (* initialize v0: v0(i) = A(0)(i) = delete i chars from t *)
+    for i = 0 to length s2 do
+      v0.(i) <- i
+    done;
+    (* main loop for the bottom up dynamic algorithm *)
+    for i = 0 to length s1 - 1 do
+      (* first edit distance is the deletion of i+1 elements from s *)
+      v1.(0) <- i+1;
+
+      (* try add/delete/replace operations *)
+      for j = 0 to length s2 - 1 do
+        let cost = if Char.compare (String.get s1 i) (String.get s2 j) = 0 then 0 else 1 in
+        v1.(j+1) <- min (v1.(j) + 1) (min (v0.(j+1) + 1) (v0.(j) + cost));
+      done;
+
+      (* copy v1 into v0 for next iteration *)
+      Array.blit v1 0 v0 0 (length s2 + 1);
+    done;
+    v1.(length s2)
+  end
+
 let repeat s n =
   assert (n>=0);
   let len = String.length s in
diff --git a/src/core/CCString.mli b/src/core/CCString.mli
index a7058e43..624eefce 100644
--- a/src/core/CCString.mli
+++ b/src/core/CCString.mli
@@ -517,6 +517,40 @@ val compare_versions : string -> string -> int
 *)
 
 
+val edit_distance : string -> string -> int
+(** Edition distance between two strings. This satisfies the classical
+    distance axioms: it is always positive, symmetric, and satisfies
+    the formula [distance a b + distance b c >= distance a c] *)
+
+(*$Q
+  Q.(string_of_size Gen.(0 -- 30)) (fun s -> \
+    edit_distance s s = 0)
+*)
+
+(* test that building a from s, and mutating one char of s, yields
+   a string s' that is accepted by a.
+
+   --> generate triples (s, i, c) where c is a char, s a non empty string
+   and i a valid index in s
+*)
+
+(*$QR
+  (
+    let gen = Q.Gen.(
+      3 -- 10 >>= fun len ->
+      0 -- (len-1) >>= fun i ->
+      string_size (return len) >>= fun s ->
+      char >|= fun c -> (s,i,c)
+    ) in
+    let small (s,_,_) = String.length s in
+    Q.make ~small gen
+  )
+  (fun (s,i,c) ->
+    let s' = Bytes.of_string s in
+    Bytes.set s' i c;
+    edit_distance s (Bytes.to_string s') <= 1)
+*)
+
 (** {2 Slices} A contiguous part of a string *)
 
 module Sub : sig