[patch] Incorrect result of org-babel-edit-distance

[patch] Incorrect result of org-babel-edit-distance

[Nicolas Richard]

[-- Attachment #1: Type: text/plain, Size: 322 bytes --]

Hello,

(org-babel-edit-distance "foo" "ffoo") returns 0, whereas 1 seems
appropriate. I don't know much about computing the levenshtein distance,
but it seems that part of the algorithm (which i found explained on
fr.wikipedia) is missing from the code. Please find a patch below trying
to address the issue.

-- 
Nico.


[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #2: 0001-ob-Fix-org-babel-edit-distance-for-insertion-deletio.patch --]
[-- Type: text/x-patch, Size: 1686 bytes --]

From 448cecaf3d6618274719fc6fa96f278b7202b784 Mon Sep 17 00:00:00 2001
From: "nrichard (geodiff-mac3)" <nrichard@ulb.ac.be>
Date: Thu, 6 Dec 2012 16:38:44 +0100
Subject: [PATCH] ob: Fix org-babel-edit-distance for insertion/deletion

* lisp/ob.el (org-babel-edit-distance): When insertion or deletion are
  needed, make sure the distance is incremented.  In addition, the now
  obsolete mmin function was removed.
---
 lisp/ob.el | 15 +++++++++------
 1 file changed, 9 insertions(+), 6 deletions(-)

diff --git a/lisp/ob.el b/lisp/ob.el
index c030a7f..0aba523 100644
--- a/lisp/ob.el
+++ b/lisp/ob.el
@@ -629,16 +629,19 @@ arguments and pop open the results in a preview buffer."
 	 (l2 (length s2))
 	 (dist (vconcat (mapcar (lambda (_) (make-vector (1+ l2) nil))
 				(number-sequence 1 (1+ l1)))))
-	 (in (lambda (i j) (aref (aref dist i) j)))
-	 (mmin (lambda (&rest lst) (apply #'min (remove nil lst)))))
+	 (in (lambda (i j) (aref (aref dist i) j))))
     (setf (aref (aref dist 0) 0) 0)
+    (dolist (j (number-sequence 1 l2))
+      (setf (aref (aref dist 0) j) j))
     (dolist (i (number-sequence 1 l1))
+      (setf (aref (aref dist i) 0) i)
       (dolist (j (number-sequence 1 l2))
 	(setf (aref (aref dist i) j)
-	      (+ (if (equal (aref s1 (1- i)) (aref s2 (1- j))) 0 1)
-		 (funcall mmin (funcall in (1- i) j)
-			  (funcall in i (1- j))
-			  (funcall in (1- i) (1- j)))))))
+	      (min
+	       (1+ (funcall in (1- i) j))
+	       (1+ (funcall in i (1- j)))
+	       (+ (if (equal (aref s1 (1- i)) (aref s2 (1- j))) 0 1)
+		  (funcall in (1- i) (1- j)))))))
     (funcall in l1 l2)))
 
 (defun org-babel-combine-header-arg-lists (original &rest others)
-- 
1.8.0

Re: [patch] Incorrect result of org-babel-edit-distance

[Eric Schulte]

"Nicolas Richard" <theonewiththeevillook@yahoo.fr> writes:

> Hello,
>
> (org-babel-edit-distance "foo" "ffoo") returns 0, whereas 1 seems
> appropriate. I don't know much about computing the levenshtein distance,
> but it seems that part of the algorithm (which i found explained on
> fr.wikipedia) is missing from the code. Please find a patch below trying
> to address the issue.
>

Applied. Thanks!,

-- 
Eric Schulte
http://cs.unm.edu/~eschulte