/**
* 将分支数据转成Query逻辑
*
* @return
*/
List<Query> toQueries(String fieldName) {
List<Query> queries = new ArrayList<Query>(1);
// 生成当前branch 的query
if (lexeme != null) {
queries.add(new TermQuery(new Term(fieldName, lexeme.getLexemeText())));
}
// 生成child branch 的query
if (acceptedBranchs != null && acceptedBranchs.size() > 0) {
if (acceptedBranchs.size() == 1) {
Query onlyOneQuery = optimizeQueries(acceptedBranchs.get(0).toQueries(fieldName));
if (onlyOneQuery != null) {
queries.add(onlyOneQuery);
}
} else {
BooleanQuery orQuery = new BooleanQuery();
for (TokenBranch childBranch : acceptedBranchs) {
Query childQuery = optimizeQueries(childBranch.toQueries(fieldName));
if (childQuery != null) {
orQuery.add(childQuery, Occur.SHOULD);
}
}
if (orQuery.getClauses().length > 0) {
queries.add(orQuery);
}
}
}
// 生成nextBranch的query
if (nextBranch != null) {
queries.addAll(nextBranch.toQueries(fieldName));
}
return queries;
}
/**
* 单连续字窜（不带空格符）单Field查询分析
*
* @param field
* @param query
* @return
*/
public static Query parse(String field, String query) {
if (field == null) {
throw new IllegalArgumentException("parameter \"field\" is null");
}
if (query == null || "".equals(query.trim())) {
return new TermQuery(new Term(field));
}
// 从缓存中取出已经解析的query生产的TokenBranch
TokenBranch root = getCachedTokenBranch(query);
if (root != null) {
return optimizeQueries(root.toQueries(field));
} else {
// System.out.println(System.currentTimeMillis());
root = new TokenBranch(null);
// 对查询条件q进行分词
StringReader input = new StringReader(query.trim());
IKSegmentation ikSeg = new IKSegmentation(input, isMaxWordLength);
try {
for (Lexeme lexeme = ikSeg.next(); lexeme != null; lexeme = ikSeg.next()) {
// 处理词元分支
root.accept(lexeme);
}
} catch (IOException e) {
e.printStackTrace();
}
// 缓存解析结果的博弈树
cachedTokenBranch(query, root);
return optimizeQueries(root.toQueries(field));
}
}
/**
* 组合词元分支
*
* @param _lexeme
* @return 返回当前branch能否接收词元对象
*/
boolean accept(Lexeme _lexeme) {
/*
* 检查新的lexeme 对当前的branch 的可接受类型
* acceptType : REFUSED 不能接受
* acceptType : ACCEPTED 接受
* acceptType : TONEXT 由相邻分支接受
*/
int acceptType = checkAccept(_lexeme);
switch (acceptType) {
case REFUSED:
// REFUSE 情况
return false;
case ACCEPTED:
if (acceptedBranchs == null) {
// 当前branch没有子branch，则添加到当前branch下
acceptedBranchs = new ArrayList<TokenBranch>(2);
acceptedBranchs.add(new TokenBranch(_lexeme));
} else {
boolean acceptedByChild = false;
// 当前branch拥有子branch，则优先由子branch接纳
for (TokenBranch childBranch : acceptedBranchs) {
acceptedByChild = childBranch.accept(_lexeme) || acceptedByChild;
}
// 如果所有的子branch不能接纳，则由当前branch接纳
if (!acceptedByChild) {
acceptedBranchs.add(new TokenBranch(_lexeme));
}
}
// 设置branch的最大右边界
if (_lexeme.getEndPosition() > this.rightBorder) {
this.rightBorder = _lexeme.getEndPosition();
}
break;
case TONEXT:
// 把lexeme放入当前branch的相邻分支
if (this.nextBranch == null) {
// 如果还没有相邻分支，则建立一个不交叠的分支
this.nextBranch = new TokenBranch(null);
}
this.nextBranch.accept(_lexeme);
break;
}
return true;
}
public boolean equals(Object o) {
if (o == null) {
return false;
}
if (this == o) {
return true;
}
if (o instanceof TokenBranch) {
TokenBranch other = (TokenBranch) o;
if (this.lexeme == null || other.getLexeme() == null) {
return false;
} else {
return this.lexeme.equals(other.getLexeme());
}
} else {
return false;
}
}