fixed issue #10

hankcs · May 8, 2015 · 8803d8f · 8803d8f
1 parent 5ce3471
commit 8803d8f
Show file tree

Hide file tree

Showing 3 changed files with 17 additions and 1 deletion.
diff --git a/pom.xml b/pom.xml
@@ -4,7 +4,7 @@
 
     <groupId>com.hankcs</groupId>
     <artifactId>hanlp</artifactId>
-    <version>1.2.0</version>
+    <version>1.2.1</version>
 
     <name>HanLP</name>
     <url>http://www.hankcs.com/</url>

diff --git a/src/main/java/com/hankcs/hanlp/seg/Segment.java b/src/main/java/com/hankcs/hanlp/seg/Segment.java
@@ -275,6 +275,7 @@ protected void mergeNumberQuantifier(List<Vertex> termList, WordNet wordNetAll,
                     sbQuantifier.setLength(0);
                 }
             }
+            sbQuantifier.setLength(0);
             line += pre.realWord.length();
         }
 //        System.out.println(wordNetAll);

diff --git a/src/test/java/com/hankcs/test/seg/TestSegment.java b/src/test/java/com/hankcs/test/seg/TestSegment.java
@@ -27,6 +27,7 @@
 
 import java.io.BufferedReader;
 import java.io.StringReader;
+import java.util.List;
 
 /**
  * @author hankcs
@@ -182,4 +183,18 @@ public void testNumberAndQuantifier() throws Exception
             System.out.println(StandardTokenizer.segment(sentence));
         }
     }
+
+    public void testIssue10() throws Exception
+    {
+        StandardTokenizer.SEGMENT.enableNumberQuantifierRecognize(true);
+        IndexTokenizer.SEGMENT.enableNumberQuantifierRecognize(true);
+        List termList = StandardTokenizer.segment("此帐号有欠费业务是什么");
+        System.out.println(termList);
+        termList = IndexTokenizer.segment("此帐号有欠费业务是什么");
+        System.out.println(termList);
+        termList = StandardTokenizer.segment("15307971214话费还有多少");
+        System.out.println(termList);
+        termList = IndexTokenizer.segment("15307971214话费还有多少");
+        System.out.println(termList);
+    }
 }