bugfix in weighted edit-based distances

markvanderloo · Jul 14, 2015 · 0a960c4 · 0a960c4
1 parent 788cf5d
commit 0a960c4
Show file tree

Hide file tree

Showing 6 changed files with 29 additions and 7 deletions.
diff --git a/pkg/NEWS b/pkg/NEWS
@@ -1,3 +1,6 @@
+version 0.9.3
+- bugfix in dl, lv, osa distance: weights were not taken into account properly (thanks to Zach Price)
+
 version 0.9.2
 - Update  fixing some errors (missing documentation, tests) in the 0.9.1 release.
 - Fixed a few possible memory leaks.

diff --git a/pkg/src/Rstringdist.c b/pkg/src/Rstringdist.c
@@ -184,6 +184,7 @@ SEXP R_amatch(SEXP x, SEXP table, SEXP method
         , p
         , q
     );
+
     double d = R_PosInf, d1 = R_PosInf;
     int index, len_X, len_T;
     unsigned int *str;

diff --git a/pkg/src/dl.c b/pkg/src/dl.c
@@ -135,14 +135,14 @@ double dl_dist(
   /* j = tgt index */
   for(i=1;i<=x;i++){ 
     uniquePush(dict,src[i]);
-    scores[(i+1) * (y + 2) + 1] = i;
+    scores[(i+1) * (y + 2) + 1] = i * weight[0];
     scores[(i+1) * (y + 2) + 0] = score_ceil;
     swapCount = 0;
 
     for(j=1;j<=y;j++){
       if(i == 1) {
         uniquePush(dict,tgt[j]);
-        scores[1 * (y + 2) + (j + 1)] = j;
+        scores[1 * (y + 2) + (j + 1)] = j * weight[0];
         scores[0 * (y + 2) + (j + 1)] = score_ceil;
       }
       targetCharCount = dict->value[which(dict, tgt[j-1])];

diff --git a/pkg/src/lv.c b/pkg/src/lv.c
@@ -43,12 +43,11 @@ double lv_dist(
   int I = na+1, L = na+1, J = nb+1;
   double sub;
 
-
   for ( i = 0; i < I; ++i ){
-    scores[i] = i;
+    scores[i] = i * weight[0];
   }
   for ( j = 1; j < J; ++j, L += I ){
-   scores[L] = j;
+   scores[L] = j * weight[0];
   }
 
   int M;

diff --git a/pkg/src/osa.c b/pkg/src/osa.c
@@ -41,10 +41,10 @@ double osa_dist(unsigned int *a, int na, unsigned int *b, int nb, double *weight
   double sub, tran;
 
    for ( i = 0; i < I; ++i ){
-      scores[i] = i;
+      scores[i] = i * weight[0];
    }
    for ( j = 1; j < J; ++j, L += I ){
-      scores[L] = j;
+      scores[L] = j * weight[0];
    }
 
 for ( i = 1; i <= na; ++i ){

diff --git a/pkg/tests/testthat/testStringdist.R b/pkg/tests/testthat/testStringdist.R
@@ -55,6 +55,25 @@ test_that("weights are handled correctly",{
       stringdist("abc","ac",method='osa',weight=c(0.5,1,1,1)),
       stringdist("ac","abc",method='osa',weight=c(1,0.5,1,1))
    )
+  expect_equal(
+    stringdist("ABC", "BC", method = "lv", weight = c(i=.1, d=.1, s=.1)),.1
+  )
+  expect_equal(
+    stringdist("ABC", "BC", method = "lv", weight = c(i=.1, d=.1, s=1)),.1
+  )
+
+  expect_equal(
+    stringdist("ABC", "BC", method = "osa", weight = c(i=.1, d=.1, s=.1,t=.1)),.1
+  )
+  expect_equal(
+    stringdist("ABC", "BC", method = "osa", weight = c(i=.1, d=.1, s=1,t=.1)),.1
+  )
+  expect_equal(
+    stringdist("ABC", "BC", method = "dl", weight = c(i=.1, d=.1, s=.1,t=.1)),.1
+  )
+  expect_equal(
+    stringdist("ABC", "BC", method = "dl", weight = c(i=.1, d=.1, s=1,t=.1)),.1
+  )
 })
 
 test_that("NA's are handled correctly",{