diff lucene/src/luan/modules/lucene/queryparser/SaneQueryParser.java @ 730:01e68da6983b

add sane-lucene-queryparser source to luan
author Franklin Schmidt <fschmidt@gmail.com>
date Fri, 10 Jun 2016 15:41:15 -0600
parents
children
line wrap: on
line diff
--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/lucene/src/luan/modules/lucene/queryparser/SaneQueryParser.java	Fri Jun 10 15:41:15 2016 -0600
@@ -0,0 +1,247 @@
+package luan.modules.lucene.queryparser;
+
+import java.util.List;
+import java.util.ArrayList;
+import java.util.regex.Pattern;
+import org.apache.lucene.search.Query;
+import org.apache.lucene.search.MatchAllDocsQuery;
+import org.apache.lucene.search.BooleanClause;
+import org.apache.lucene.search.BooleanQuery;
+import org.apache.lucene.search.Sort;
+import org.apache.lucene.search.SortField;
+
+
+public class SaneQueryParser {
+
+	public static Query parseQuery(FieldParser fieldParser,String query) throws ParseException {
+		return new SaneQueryParser(fieldParser,query).parseQuery();
+	}
+
+	private static Pattern specialChar = Pattern.compile("[ \\t\\r\\n\":\\[\\]{}^+\\-(),?*\\\\]");
+
+	public static String literal(String s) {
+		return specialChar.matcher(s).replaceAll("\\\\$0");
+	}
+
+	public static Sort parseSort(FieldParser fieldParser,String sort) throws ParseException {
+		return new SaneQueryParser(fieldParser,sort).parseSort();
+	}
+
+
+	private static final String NOT_IN_TERM = " \t\r\n\":[]{}^+-()";
+	private static final String NOT_IN_FIELD = NOT_IN_TERM + ",";
+	private final FieldParser fieldParser;
+	final Parser parser;
+
+	private SaneQueryParser(FieldParser fieldParser,String query) {
+		this.fieldParser = fieldParser;
+		this.parser = new Parser(query);
+	}
+
+	private Query parseQuery() throws ParseException {
+		Spaces();
+		BooleanQuery bq = new BooleanQuery();
+		while( !parser.endOfInput() ) {
+			bq.add( Term(null) );
+		}
+		BooleanClause[] clauses = bq.getClauses();
+		switch( clauses.length ) {
+		case 0:
+			return new MatchAllDocsQuery();
+		case 1:
+			{
+				BooleanClause bc = clauses[0];
+				if( bc.getOccur() != BooleanClause.Occur.MUST_NOT )
+					return bc.getQuery();
+			}
+		default:
+			return bq;
+		}
+	}
+
+	private BooleanClause Term(String defaultField) throws ParseException {
+		BooleanClause.Occur occur;
+		if( parser.match('+') ) {
+			occur = BooleanClause.Occur.MUST;
+			Spaces();
+		} else if( parser.match('-') ) {
+			occur = BooleanClause.Occur.MUST_NOT;
+			Spaces();
+		} else {
+			occur = BooleanClause.Occur.SHOULD;
+		}
+		String field = QueryField();
+		if( field == null )
+			field = defaultField;
+		Query query = NestedTerm(field);
+		if( query == null )
+			query = RangeTerm(field);
+		if( query == null ) {
+			parser.begin();
+			String match = SimpleTerm();
+			query = fieldParser.getQuery(this,field,match);
+			parser.success();
+		}
+		if( parser.match('^') ) {
+			Spaces();
+			int start = parser.begin();
+			try {
+				while( parser.anyOf("0123456789.") );
+				String match = parser.textFrom(start);
+				float boost = Float.parseFloat(match);
+				query.setBoost(boost);
+			} catch(NumberFormatException e) {
+				throw new ParseException(this,e);
+			}
+			parser.success();
+			Spaces();
+		}
+		BooleanClause bc = new BooleanClause(query,occur);
+		return bc;
+	}
+
+	private Query NestedTerm(String field) throws ParseException {
+		parser.begin();
+		if( !parser.match('(') )
+			return parser.failure(null);
+		BooleanQuery bq = new BooleanQuery();
+		while( !parser.match(')') ) {
+			if( parser.endOfInput() )
+				throw new ParseException(this,"unclosed parentheses");
+			bq.add( Term(field) );
+		}
+		Spaces();
+		BooleanClause[] clauses = bq.getClauses();
+		switch( clauses.length ) {
+		case 0:
+			throw new ParseException(this,"empty parentheses");
+		case 1:
+			{
+				BooleanClause bc = clauses[0];
+				if( bc.getOccur() != BooleanClause.Occur.MUST_NOT )
+					return parser.success(bc.getQuery());
+			}
+		default:
+			return parser.success(bq);
+		}
+	}
+
+	private Query RangeTerm(String field) throws ParseException {
+		parser.begin();
+		if( !parser.anyOf("[{") )
+			return parser.failure(null);
+		boolean includeMin = parser.lastChar() == '[';
+		Spaces();
+		String minQuery = SimpleTerm();
+		TO();
+		String maxQuery = SimpleTerm();
+		if( !parser.anyOf("]}") )
+			throw new ParseException(this,"unclosed range");
+		boolean includeMax = parser.lastChar() == ']';
+		Spaces();
+		Query query = fieldParser.getRangeQuery(this,field,minQuery,maxQuery,includeMin,includeMax);
+		return parser.success(query);
+	}
+
+	private void TO() throws ParseException {
+		parser.begin();
+		if( !(parser.match("TO") && Space()) )
+			throw new ParseException(this,"'TO' expected");
+		Spaces();
+		parser.success();
+	}
+
+	private String SimpleTerm() throws ParseException {
+		parser.begin();
+		String match;
+		if( parser.match('"') ) {
+			int start = parser.currentIndex() - 1;
+			while( !parser.match('"') ) {
+				if( parser.endOfInput() )
+					throw new ParseException(this,"unclosed quotes");
+				parser.anyChar();
+				checkEscape();
+			}
+			match = parser.textFrom(start);
+			Spaces();
+		} else {
+			match = Unquoted(NOT_IN_TERM);
+		}
+		if( match.length() == 0 )
+			throw new ParseException(this);
+		return parser.success(match);
+	}
+
+	private String QueryField() throws ParseException {
+		parser.begin();
+		String match = Field();
+		if( match==null || !parser.match(':') )
+			return parser.failure((String)null);
+		Spaces();
+		return parser.success(match);
+	}
+
+	private String Field() throws ParseException {
+		parser.begin();
+		String match = Unquoted(NOT_IN_FIELD);
+		if( match.length()==0 )
+			return parser.failure((String)null);
+		match = StringFieldParser.escape(this,match);
+		return parser.success(match);
+	}
+
+	private String Unquoted(String exclude) throws ParseException {
+		int start = parser.begin();
+		while( parser.noneOf(exclude) ) {
+			checkEscape();
+		}
+		String match = parser.textFrom(start);
+		Spaces();
+		return parser.success(match);
+	}
+
+	private void checkEscape() {
+		if( parser.lastChar() == '\\' )
+			parser.anyChar();
+	}
+
+	private void Spaces() {
+		while( Space() );
+	}
+
+	private boolean Space() {
+		return parser.anyOf(" \t\r\n");
+	}
+
+
+	// sort
+
+	private Sort parseSort() throws ParseException {
+		Spaces();
+		if( parser.endOfInput() )
+			return null;
+		List<SortField> list = new ArrayList<SortField>();
+		list.add( SortField() );
+		while( !parser.endOfInput() ) {
+			parser.begin();
+			if( !parser.match(',') )
+				throw new ParseException(this,"',' expected");
+			Spaces();
+			parser.success();
+			list.add( SortField() );
+		}
+		return new Sort(list.toArray(new SortField[0]));
+	}
+
+	private SortField SortField() throws ParseException {
+		parser.begin();
+		String field = Field();
+		if( field==null )
+			throw new ParseException(this);
+		boolean reverse = !parser.matchIgnoreCase("asc") && parser.matchIgnoreCase("desc");
+		Spaces();
+		SortField sf = fieldParser.getSortField(this,field,reverse);
+		return parser.success(sf);
+	}
+
+}