进一步改进原子切分 fix hankcs#1421 (comment)

hankcs · hankcs · commit b62db0de5df2 · 2020-02-14T21:06:57.000-05:00
diff --git a/src/main/java/com/hankcs/hanlp/seg/Segment.java b/src/main/java/com/hankcs/hanlp/seg/Segment.java
@@ -161,7 +161,7 @@ protected static List<AtomNode> quickAtomSegment(char[] charArray, int start, in
         int offsetAtom = start;
         int preType = CharType.get(charArray[offsetAtom]);
         int curType;
-        while (++offsetAtom < end || (atomNodeList.isEmpty() && offsetAtom < charArray.length))
+        while (++offsetAtom < end)
         {
             curType = CharType.get(charArray[offsetAtom]);
             if (curType != preType)
diff --git a/src/main/java/com/hankcs/hanlp/seg/WordBasedSegment.java b/src/main/java/com/hankcs/hanlp/seg/WordBasedSegment.java
@@ -414,7 +414,7 @@ public void hit(int begin, int end, CoreDictionary.Attribute value)
                 int j = i + 1;
                 for (; j < vertexes.length - 1; ++j)
                 {
-                    if (!vertexes[j].isEmpty()) break;
+                    if (!vertexes[j].isEmpty() && CharType.get(charArray[j - 1]) != CharType.CT_CNUM) break;
                 }
                 wordNetStorage.add(i, quickAtomSegment(charArray, i - 1, j - 1));
                 i = j;
diff --git a/src/test/java/com/hankcs/hanlp/seg/SegmentTest.java b/src/test/java/com/hankcs/hanlp/seg/SegmentTest.java
@@ -539,6 +539,7 @@ public void testIssue1172()
 
     public void testIssue1421()
     {
+        System.out.println(HanLP.segment("android十一中国版本"));
         System.out.println(HanLP.segment("十一中国放假吗"));
         System.out.println(HanLP.segment("十一发展计划"));
         System.out.println(HanLP.segment("十一通过山海关吗"));

Original file line number	Diff line number	Diff line change
`@@ -161,7 +161,7 @@ protected static List<AtomNode> quickAtomSegment(char[] charArray, int start, in`
`161`	`161`	`int offsetAtom = start;`
`162`	`162`	`int preType = CharType.get(charArray[offsetAtom]);`
`163`	`163`	`int curType;`
`164`		`- while (++offsetAtom < end \|\| (atomNodeList.isEmpty() && offsetAtom < charArray.length))`
	`164`	`+ while (++offsetAtom < end)`
`165`	`165`	`{`
`166`	`166`	`curType = CharType.get(charArray[offsetAtom]);`
`167`	`167`	`if (curType != preType)`
Original file line number	Diff line number	Diff line change
`@@ -414,7 +414,7 @@ public void hit(int begin, int end, CoreDictionary.Attribute value)`
`414`	`414`	`int j = i + 1;`
`415`	`415`	`for (; j < vertexes.length - 1; ++j)`
`416`	`416`	`{`
`417`		`- if (!vertexes[j].isEmpty()) break;`
	`417`	`+ if (!vertexes[j].isEmpty() && CharType.get(charArray[j - 1]) != CharType.CT_CNUM) break;`
`418`	`418`	`}`
`419`	`419`	`wordNetStorage.add(i, quickAtomSegment(charArray, i - 1, j - 1));`
`420`	`420`	`i = j;`
Original file line number	Diff line number	Diff line change
`@@ -539,6 +539,7 @@ public void testIssue1172()`
`539`	`539`
`540`	`540`	`public void testIssue1421()`
`541`	`541`	`{`
	`542`	`+ System.out.println(HanLP.segment("android十一中国版本"));`
`542`	`543`	`System.out.println(HanLP.segment("十一中国放假吗"));`
`543`	`544`	`System.out.println(HanLP.segment("十一发展计划"));`
`544`	`545`	`System.out.println(HanLP.segment("十一通过山海关吗"));`