fixed issue hankcs#22

MeteorLzz · May 27, 2015 · d56c9b6 · d56c9b6
1 parent 90cd0a5
commit d56c9b6
Show file tree

Hide file tree

Showing 2 changed files with 25 additions and 10 deletions.
diff --git a/src/main/java/com/hankcs/hanlp/collection/trie/DoubleArrayTrie.java b/src/main/java/com/hankcs/hanlp/collection/trie/DoubleArrayTrie.java
@@ -1108,25 +1108,27 @@ public Searcher(int offset, char[] charArray)
 
         public boolean next()
         {
-            if (i == arrayLength)
-            {
-                ++begin;
-                i = begin;
-                last = base[0];
-            }
             int b = last;
             int n;
             int p;
 
-            for (; i < arrayLength; ++i)
+            for (; ; ++i)
             {
-                p = b + (int) (charArray[i]) + 1;    // 状态转移 p = base[char[i-1]] + char[i] + 1
+                if (i == arrayLength)               // 指针到头了，将起点往前挪一个，重新开始，状态归零
+                {
+                    ++begin;
+                    if (begin == arrayLength) break;
+                    i = begin;
+                    b = base[0];
+                }
+                p = b + (int) (charArray[i]) + 1;   // 状态转移 p = base[char[i-1]] + char[i] + 1
                 if (b == check[p])                  // base[char[i-1]] == check[base[char[i-1]] + char[i] + 1]
-                    b = base[p];
+                    b = base[p];                    // 转移成功
                 else
                 {
-                    i = begin;
+                    i = begin;                      // 转移失败，也将起点往前挪一个，重新开始，状态归零
                     ++begin;
+                    if (begin == arrayLength) break;
                     b = base[0];
                     continue;
                 }

diff --git a/src/test/java/com/hankcs/test/seg/TestSegment.java b/src/test/java/com/hankcs/test/seg/TestSegment.java
@@ -13,6 +13,7 @@
 
 import com.hankcs.hanlp.HanLP;
 import com.hankcs.hanlp.dictionary.CoreBiGramTableDictionary;
+import com.hankcs.hanlp.dictionary.CoreDictionary;
 import com.hankcs.hanlp.dictionary.CustomDictionary;
 import com.hankcs.hanlp.dictionary.other.CharTable;
 import com.hankcs.hanlp.dictionary.other.CharType;
@@ -278,4 +279,16 @@ public void testIssuse17() throws Exception
         HanLP.Config.Normalization = true;
         System.out.println(StandardTokenizer.segment("号 "));
     }
+
+    public void testIssue22() throws Exception
+    {
+        CoreDictionary.Attribute attribute = CoreDictionary.get("年");
+        System.out.println(attribute);
+        List<Term> termList = StandardTokenizer.segment("三年");
+        System.out.println(termList);
+        assertEquals(attribute.nature[0], termList.get(1).nature);
+        System.out.println(StandardTokenizer.segment("三元"));
+        StandardTokenizer.SEGMENT.enableNumberQuantifierRecognize(true);
+        System.out.println(StandardTokenizer.segment("三年"));
+    }
 }