add positionLength

NLPIR-team · Mar 13, 2018 · 4167a52 · 4167a52
1 parent b8a6c0d
commit 4167a52
Show file tree

Hide file tree

Showing 14 changed files with 15 additions and 12 deletions.
diff --git a/src/main/java/org/nlpir/lucene/cn/ictclas/finersegmet/FinerTokenizer.java b/src/main/java/org/nlpir/lucene/cn/ictclas/finersegmet/FinerTokenizer.java
@@ -179,6 +179,7 @@ public boolean incrementToken() throws IOException {
 		end = start + length;
 		termAtt.copyBuffer(buffer[current].toCharArray(), 0, length);
 		offsetAtt.setOffset(correctOffset(start), correctOffset(end));
+		posArr.setPositionLength(current+1);
 		typeAtt.setType("word");
 		start = end;
 		current += 1;

diff --git a/src/test/java/org/nlpir/lucene/cn/ictclas/NLPIRTokenizerTest.java b/src/test/java/org/nlpir/lucene/cn/ictclas/NLPIRTokenizerTest.java
@@ -20,28 +20,30 @@
 public class NLPIRTokenizerTest {
 
 	public static void main(String[] args) throws Exception {
-		//NLPIR
+		// NLPIR
 		NLPIRTokenizerAnalyzer nta = new NLPIRTokenizerAnalyzer("", 1, "", "", false);
-		//Index
-		IndexWriterConfig inconf=new IndexWriterConfig(nta);
+		// Index
+		IndexWriterConfig inconf = new IndexWriterConfig(nta);
 		inconf.setOpenMode(OpenMode.CREATE_OR_APPEND);
-		IndexWriter index=new IndexWriter(FSDirectory.open(Paths.get("index/")),inconf);
+		IndexWriter index = new IndexWriter(FSDirectory.open(Paths.get("index/")), inconf);
 		Document doc = new Document();
-		doc.add(new TextField("contents", "特朗普表示，很高兴汉堡会晤后再次同习近平主席通话。我同习主席就重大问题保持沟通和协调、两国加强各层级和各领域交往十分重要。当前，美中关系发展态势良好，我相信可以发展得更好。我期待着对中国进行国事访问。",Field.Store.YES));
+		doc.add(new TextField("contents",
+				"特朗普表示，很高兴汉堡会晤后再次同习近平主席通话。我同习主席就重大问题保持沟通和协调、两国加强各层级和各领域交往十分重要。当前，美中关系发展态势良好，我相信可以发展得更好。我期待着对中国进行国事访问。",
+				Field.Store.YES));
 		index.addDocument(doc);
 		index.flush();
 		index.close();
-		//Search
+		// Search
 		String field = "contents";
 		IndexReader reader = DirectoryReader.open(FSDirectory.open(Paths.get("index/")));
 		IndexSearcher searcher = new IndexSearcher(reader);
 		QueryParser parser = new QueryParser(field, nta);
 		Query query = parser.parse("特朗普习近平");
-		TopDocs top=searcher.search(query, 100);
-		System.out.println("总条数："+top.totalHits);
+		TopDocs top = searcher.search(query, 100);
+		System.out.println("总条数：" + top.totalHits);
 		ScoreDoc[] hits = top.scoreDocs;
-		for(int i=0;i<hits.length;i++) {
-			System.out.println("doc="+hits[i].doc+" score="+hits[i].score);
+		for (int i = 0; i < hits.length; i++) {
+			System.out.println("doc=" + hits[i].doc + " score=" + hits[i].score);
 			Document d = searcher.doc(hits[i].doc);
 			System.out.println(d.get("contents"));
 		}

diff --git a/...TA-INF/maven/lucene-analyzers-nlpir-ictclas/lucene-analyzers-nlpir-ictclas/pom.properties b/...TA-INF/maven/lucene-analyzers-nlpir-ictclas/lucene-analyzers-nlpir-ictclas/pom.properties
@@ -1,5 +1,5 @@
 #Generated by Maven Integration for Eclipse
-#Mon Mar 12 11:49:08 CST 2018
+#Tue Mar 13 13:25:11 CST 2018
 version=6.6.0
 groupId=lucene-analyzers-nlpir-ictclas
 m2e.projectName=nlpir-analysis-cn-ictclas

diff --git a/target/classes/org/nlpir/lucene/cn/ictclas/NLPIRTokenizer.class b/target/classes/org/nlpir/lucene/cn/ictclas/NLPIRTokenizer.class
diff --git a/target/classes/org/nlpir/lucene/cn/ictclas/NLPIRTokenizerAnalyzer.class b/target/classes/org/nlpir/lucene/cn/ictclas/NLPIRTokenizerAnalyzer.class
diff --git a/target/classes/org/nlpir/lucene/cn/ictclas/NLPIRTokenizerFactory.class b/target/classes/org/nlpir/lucene/cn/ictclas/NLPIRTokenizerFactory.class
diff --git a/target/classes/org/nlpir/lucene/cn/ictclas/finersegmet/FinerTokenizer.class b/target/classes/org/nlpir/lucene/cn/ictclas/finersegmet/FinerTokenizer.class
diff --git a/target/classes/org/nlpir/lucene/cn/ictclas/finersegmet/FinerTokenizerAnalyzer.class b/target/classes/org/nlpir/lucene/cn/ictclas/finersegmet/FinerTokenizerAnalyzer.class
diff --git a/target/classes/org/nlpir/lucene/cn/ictclas/finersegmet/FinerTokenizerFactory.class b/target/classes/org/nlpir/lucene/cn/ictclas/finersegmet/FinerTokenizerFactory.class
diff --git a/target/classes/org/nlpir/segment/CNLPIRLibrary.class b/target/classes/org/nlpir/segment/CNLPIRLibrary.class
diff --git a/target/classes/org/nlpir/segment/exception/NLPIRException.class b/target/classes/org/nlpir/segment/exception/NLPIRException.class
diff --git a/target/lucene-analyzers-nlpir-ictclas-6.6.0.jar b/target/lucene-analyzers-nlpir-ictclas-6.6.0.jar
diff --git a/target/maven-archiver/pom.properties b/target/maven-archiver/pom.properties
@@ -1,5 +1,5 @@
 #Generated by Maven
-#Mon Mar 12 11:35:22 CST 2018
+#Tue Mar 13 12:09:58 CST 2018
 version=6.6.0
 groupId=lucene-analyzers-nlpir-ictclas
 artifactId=lucene-analyzers-nlpir-ictclas
diff --git a/target/test-classes/org/nlpir/lucene/cn/ictclas/NLPIRTokenizerTest.class b/target/test-classes/org/nlpir/lucene/cn/ictclas/NLPIRTokenizerTest.class