Merge pull request diffbot#1 from rschwab68/master

fixed bug in spotting known surface forms
akms17 · Apr 5, 2016 · 9ed6ab8 · 9ed6ab8
2 parents 81c16ee + 80afd47
commit 9ed6ab8
Showing 1 changed file with 1 addition and 1 deletion.
diff --git a/src/main/java/com/diffbot/wikistatsextractor/util/Util.java b/src/main/java/com/diffbot/wikistatsextractor/util/Util.java
@@ -574,7 +574,7 @@ public static List<String> getKnownSurfaceFormsInParagraph(String paragraph, Set
 		 for (int i = 0; i < spans.length; i++) {
 			 for (int j = 0; j < MAX_NB_TOKEN_SF-1; j++) {
 				 if (i+j < spans.length) {
-					 String substr = paragraph.substring(spans[i].getStart(), spans[i + j].getEnd());
+					 String substr = clean_paragraph_text.substring(spans[i].getStart(), spans[i + j].getEnd());
 					 if (allowedSFs.contains(substr)) {
 						 knownSurfaceForms.add(substr);
 					 }