lucene/queryparser/src/java/org/apache/lucene/queryparser/flexible/standard/parser/StandardSyntaxParser.jj

options {
  STATIC=false;
  JAVA_UNICODE_ESCAPE=true;
  USER_CHAR_STREAM=true;
  IGNORE_CASE=false;
  JDK_VERSION="1.5";
}

PARSER_BEGIN(StandardSyntaxParser)
package org.apache.lucene.queryparser.flexible.standard.parser;

/*
 * Licensed to the Apache Software Foundation (ASF) under one or more
 * contributor license agreements.  See the NOTICE file distributed with
 * this work for additional information regarding copyright ownership.
 * The ASF licenses this file to You under the Apache License, Version 2.0
 * (the "License"); you may not use this file except in compliance with
 * the License.  You may obtain a copy of the License at
 *
 *     http://www.apache.org/licenses/LICENSE-2.0
 *
 * Unless required by applicable law or agreed to in writing, software
 * distributed under the License is distributed on an "AS IS" BASIS,
 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 * See the License for the specific language governing permissions and
 * limitations under the License.
 */

import java.io.StringReader;
import java.util.Vector;
import java.util.Arrays;

import org.apache.lucene.queryparser.flexible.messages.Message;
import org.apache.lucene.queryparser.flexible.messages.MessageImpl;
import org.apache.lucene.queryparser.flexible.core.QueryNodeParseException;
import org.apache.lucene.queryparser.flexible.core.messages.QueryParserMessages;
import org.apache.lucene.queryparser.flexible.core.nodes.AndQueryNode;
import org.apache.lucene.queryparser.flexible.core.nodes.BooleanQueryNode;
import org.apache.lucene.queryparser.flexible.core.nodes.BoostQueryNode;
import org.apache.lucene.queryparser.flexible.core.nodes.FieldQueryNode;
import org.apache.lucene.queryparser.flexible.core.nodes.FuzzyQueryNode;
import org.apache.lucene.queryparser.flexible.core.nodes.ModifierQueryNode;
import org.apache.lucene.queryparser.flexible.core.nodes.GroupQueryNode;
import org.apache.lucene.queryparser.flexible.core.nodes.OrQueryNode;
import org.apache.lucene.queryparser.flexible.standard.nodes.RegexpQueryNode;
import org.apache.lucene.queryparser.flexible.core.nodes.SlopQueryNode;
import org.apache.lucene.queryparser.flexible.core.nodes.QueryNode;
import org.apache.lucene.queryparser.flexible.core.nodes.QuotedFieldQueryNode;
import org.apache.lucene.queryparser.flexible.core.parser.SyntaxParser;
import org.apache.lucene.queryparser.flexible.standard.nodes.TermRangeQueryNode;

/**
 * Parser for the standard Lucene syntax
 */
public class StandardSyntaxParser implements SyntaxParser {

 
   // syntax parser constructor
   public StandardSyntaxParser() {
     this(new FastCharStream(new StringReader("")));
  }
     /** Parses a query string, returning a {@link org.apache.lucene.queryparser.flexible.core.nodes.QueryNode}.
     *  @param query  the query string to be parsed.
     *  @throws ParseException if the parsing fails
     */
    public QueryNode parse(CharSequence query, CharSequence field) throws QueryNodeParseException {
      ReInit(new FastCharStream(new StringReader(query.toString())));
      try {
        // TopLevelQuery is a Query followed by the end-of-input (EOF)
        QueryNode querynode = TopLevelQuery(field);
        return querynode;
      }
      catch (ParseException tme) {
            tme.setQuery(query);
            throw tme;
      }
      catch (Error tme) {
          Message message = new MessageImpl(QueryParserMessages.INVALID_SYNTAX_CANNOT_PARSE, query, tme.getMessage());
          QueryNodeParseException e = new QueryNodeParseException(tme);
            e.setQuery(query);
            e.setNonLocalizedMessage(message);
            throw e;
      }
    }
   
}

PARSER_END(StandardSyntaxParser)

/* ***************** */
/* Token Definitions */
/* ***************** */

<*> TOKEN : {
  <#_NUM_CHAR:   ["0"-"9"] >
// every character that follows a backslash is considered as an escaped character
| <#_ESCAPED_CHAR: "\\" ~[] >
| <#_TERM_START_CHAR: ( ~[ " ", "\t", "\n", "\r", "\u3000", "+", "-", "!", "(", ")", ":", "^",
                           "<", ">", "=", "[", "]", "\"", "{", "}", "~", "\\", "/" ]
                       | <_ESCAPED_CHAR> ) >
| <#_TERM_CHAR: ( <_TERM_START_CHAR> | <_ESCAPED_CHAR> | "-" | "+" ) >
| <#_WHITESPACE: ( " " | "\t" | "\n" | "\r" | "\u3000") >
| <#_QUOTED_CHAR: ( ~[ "\"", "\\" ] | <_ESCAPED_CHAR> ) >
}

<DEFAULT, Range> SKIP : {
  < <_WHITESPACE>>
}

<DEFAULT> TOKEN : {
  <AND:       ("AND" | "&&") >
| <OR:        ("OR" | "||") >
| <NOT:       ("NOT" | "!") >
| <PLUS:      "+" >
| <MINUS:     "-" >
| <LPAREN:    "(" >
| <RPAREN:    ")" >
| <OP_COLON:  ":" >
| <OP_EQUAL:  "=" >
| <OP_LESSTHAN:   "<"  >
| <OP_LESSTHANEQ: "<=" >
| <OP_MORETHAN:   ">"  >
| <OP_MORETHANEQ: ">=" >
| <CARAT:     "^" > : Boost
| <QUOTED:     "\"" (<_QUOTED_CHAR>)* "\"">
| <TERM:      <_TERM_START_CHAR> (<_TERM_CHAR>)*  >
| <FUZZY_SLOP:     "~" ( (<_NUM_CHAR>)+ ( "." (<_NUM_CHAR>)+ )? )? >
| <REGEXPTERM: "/" (~[ "/" ] | "\\/" )* "/" >
| <RANGEIN_START: "[" > : Range
| <RANGEEX_START: "{" > : Range
}

<Boost> TOKEN : {
<NUMBER:    (<_NUM_CHAR>)+ ( "." (<_NUM_CHAR>)+ )? > : DEFAULT
}

<Range> TOKEN : {
<RANGE_TO: "TO">
| <RANGEIN_END: "]"> : DEFAULT
| <RANGEEX_END: "}"> : DEFAULT
| <RANGE_QUOTED: "\"" (~["\""] | "\\\"")+ "\"">
| <RANGE_GOOP: (~[ " ", "]", "}" ])+ >
}

ModifierQueryNode.Modifier Modifiers() : {
  ModifierQueryNode.Modifier ret = ModifierQueryNode.Modifier.MOD_NONE;
}
{
  [
     <PLUS> { ret = ModifierQueryNode.Modifier.MOD_REQ; }
     | <MINUS> { ret = ModifierQueryNode.Modifier.MOD_NOT; }
     | <NOT> { ret = ModifierQueryNode.Modifier.MOD_NOT; }
  ]
  { return ret; }
}

// This makes sure that there is no garbage after the query string
QueryNode TopLevelQuery(CharSequence field) : 
{
  QueryNode q;
}
{
  q=Query(field) <EOF>
  {
     return q;
  }
}

// These changes were made to introduce operator precedence:
// - Clause() now returns a QueryNode. 
// - The modifiers are consumed by Clause() and returned as part of the QueryNode Object
// - Query does not consume conjunctions (AND, OR) anymore. 
// - This is now done by two new non-terminals: ConjClause and DisjClause
// The parse tree looks similar to this:
//       Query ::= DisjQuery ( DisjQuery )*
//   DisjQuery ::= ConjQuery ( OR ConjQuery )* 
//   ConjQuery ::= Clause ( AND Clause )*
//      Clause ::= [ Modifier ] ... 


QueryNode Query(CharSequence field) :
{
  Vector<QueryNode> clauses = null;
  QueryNode c, first=null;
}
{
  first=DisjQuery(field)
  (
    c=DisjQuery(field)
    { 
       if (clauses == null) {
           clauses = new Vector<QueryNode>();
           clauses.addElement(first);
        }
        clauses.addElement(c);
    }
    )*
    {
        if (clauses != null) { 
        return new BooleanQueryNode(clauses);
      } else {
          // Handle the case of a "pure" negation query which
          // needs to be wrapped as a boolean query, otherwise
          // the returned result drops the negation.
          if (first instanceof ModifierQueryNode) {
            ModifierQueryNode m = (ModifierQueryNode) first;
            if (m.getModifier() == ModifierQueryNode.Modifier.MOD_NOT) {
              return new BooleanQueryNode(Arrays.<QueryNode> asList(m));
            }
          }
          return first;
      }
    }
}

QueryNode DisjQuery(CharSequence field) : {
  QueryNode first, c;
  Vector<QueryNode> clauses = null;
}
{
  first = ConjQuery(field)
  (
   <OR> c=ConjQuery(field)
   { 
     if (clauses == null) {
         clauses = new Vector<QueryNode>();
         clauses.addElement(first); 
     } 
     clauses.addElement(c);
   }
  )*
  {
    if (clauses != null) { 
      return new OrQueryNode(clauses);
    } else {
        return first;
    }
  }
}

QueryNode ConjQuery(CharSequence field) : {
  QueryNode first, c;
  Vector<QueryNode> clauses = null;
}
{
  first = ModClause(field)
  (
   <AND> c=ModClause(field)
   { 
     if (clauses == null) {
         clauses = new Vector<QueryNode>();
         clauses.addElement(first);
     } 
     clauses.addElement(c); 
   }
  )*
  {
    if (clauses != null) {     
      return new AndQueryNode(clauses);
    } else {
        return first;
    }
  }
}

// QueryNode Query(CharSequence field) :
// {
// List clauses = new ArrayList();
//   List modifiers = new ArrayList();
//   QueryNode q, firstQuery=null;
//   ModifierQueryNode.Modifier mods;
//   int conj;
// }
// {
//   mods=Modifiers() q=Clause(field)
//   {
//     if (mods == ModifierQueryNode.Modifier.MOD_NONE) firstQuery=q;
//     
//     // do not create modifier nodes with MOD_NONE
//      if (mods != ModifierQueryNode.Modifier.MOD_NONE) {
//          q = new ModifierQueryNode(q, mods);
//         }
//      clauses.add(q);
//   }
//   (
//     conj=Conjunction() mods=Modifiers() q=Clause(field)
//     { 
//       // do not create modifier nodes with MOD_NONE
//         if (mods != ModifierQueryNode.Modifier.MOD_NONE) {
//          q = new ModifierQueryNode(q, mods);
//         }
//          clauses.add(q);
//        //TODO: figure out what to do with AND and ORs
//   }
//   )*
//     {
//      if (clauses.size() == 1 && firstQuery != null)
//         return firstQuery;
//       else {
//       return new BooleanQueryNode(clauses);
//       }
//     }
// }

QueryNode ModClause(CharSequence field) : {
  QueryNode q; 
  ModifierQueryNode.Modifier mods;
}
{
   mods=Modifiers() q= Clause(field) {
        if (mods != ModifierQueryNode.Modifier.MOD_NONE) {
           q = new ModifierQueryNode(q, mods);
        }
        return q;
   }
}

QueryNode Clause(CharSequence field) : {
  QueryNode q;
  Token fieldToken=null, boost=null, operator=null, term=null;
  FieldQueryNode qLower, qUpper;
  boolean lowerInclusive, upperInclusive;
  
  boolean group = false;
}
{
(
  LOOKAHEAD(3)
  fieldToken=<TERM> (
    ( <OP_COLON> | <OP_EQUAL> ) {field=EscapeQuerySyntaxImpl.discardEscapeChar(fieldToken.image);} q=Term(field)
  | ( operator=<OP_LESSTHAN> | operator=<OP_LESSTHANEQ> | operator=<OP_MORETHAN> | operator=<OP_MORETHANEQ> ) {field=EscapeQuerySyntaxImpl.discardEscapeChar(fieldToken.image);}( term=<TERM> | term=<QUOTED> | term=<NUMBER> )
    {
        if (term.kind == QUOTED) {
            term.image = term.image.substring(1, term.image.length()-1);
        }
        switch (operator.kind) {
            case OP_LESSTHAN:
              lowerInclusive = true;
              upperInclusive = false;

               qLower = new FieldQueryNode(field,
                                         "*", term.beginColumn, term.endColumn);
            qUpper = new FieldQueryNode(field,
                                 EscapeQuerySyntaxImpl.discardEscapeChar(term.image), term.beginColumn, term.endColumn);
                
                break;
            case OP_LESSTHANEQ:
              lowerInclusive = true;
              upperInclusive = true;
            
                qLower = new FieldQueryNode(field, 
                                         "*", term.beginColumn, term.endColumn);
                qUpper = new FieldQueryNode(field, 
                                         EscapeQuerySyntaxImpl.discardEscapeChar(term.image), term.beginColumn, term.endColumn);
                break;
            case OP_MORETHAN:
              lowerInclusive = false;
              upperInclusive = true;
            
                qLower = new FieldQueryNode(field, 
                                         EscapeQuerySyntaxImpl.discardEscapeChar(term.image), term.beginColumn, term.endColumn);
                qUpper = new FieldQueryNode(field, 
                                         "*", term.beginColumn, term.endColumn);
                break;
            case OP_MORETHANEQ:
              lowerInclusive = true;
              upperInclusive = true;

                qLower = new FieldQueryNode(field, 
                                         EscapeQuerySyntaxImpl.discardEscapeChar(term.image), term.beginColumn, term.endColumn);
                qUpper = new FieldQueryNode(field, 
                                         "*", term.beginColumn, term.endColumn);
                break;
            default:
                throw new Error("Unhandled case: operator="+operator.toString());
        }
        q = new TermRangeQueryNode(qLower, qUpper, lowerInclusive, upperInclusive);
    }
  )
| [
    LOOKAHEAD(2)
    fieldToken=<TERM>
    ( <OP_COLON> | <OP_EQUAL> ) {field=EscapeQuerySyntaxImpl.discardEscapeChar(fieldToken.image);}
  ]
  (
   q=Term(field)
   | <LPAREN> q=Query(field) <RPAREN> (<CARAT> boost=<NUMBER>)? {group=true;}
  )
)
    {
      if (boost != null) {
      float f = (float)1.0;
      try {
        f = Float.parseFloat(boost.image);
        // avoid boosting null queries, such as those caused by stop words
          if (q != null) {
            q = new BoostQueryNode(q, f);
          }
      } catch (Exception ignored) {
        /* Should this be handled somehow? (defaults to "no boost", if
             * boost number is invalid)
             */
      }
      }
      if (group) { q = new GroupQueryNode(q);}
      return q;
    }
}


QueryNode Term(CharSequence field) : {
  Token term, boost=null, fuzzySlop=null, goop1, goop2;
  boolean fuzzy = false;
  boolean regexp = false;
  boolean startInc=false;
  boolean endInc=false;
  QueryNode q =null; 
  FieldQueryNode qLower, qUpper;
  float defaultMinSimilarity = org.apache.lucene.search.FuzzyQuery.defaultMinSimilarity;
}
{
  (
     (
      term=<TERM> { q = new FieldQueryNode(field, EscapeQuerySyntaxImpl.discardEscapeChar(term.image), term.beginColumn, term.endColumn); }
       | term=<REGEXPTERM> { regexp=true; }
       | term=<NUMBER>
     )
     [ fuzzySlop=<FUZZY_SLOP> { fuzzy=true; } ]
     [ <CARAT> boost=<NUMBER> [ fuzzySlop=<FUZZY_SLOP> { fuzzy=true; } ] ]
     {
       if (fuzzy) {
           float fms = defaultMinSimilarity;
           try {
            fms = Float.parseFloat(fuzzySlop.image.substring(1));
           } catch (Exception ignored) { }
           if(fms < 0.0f){
                throw new ParseException(new MessageImpl(QueryParserMessages.INVALID_SYNTAX_FUZZY_LIMITS));
          } else if (fms >= 1.0f && fms != (int) fms) {
            throw new ParseException(new MessageImpl(QueryParserMessages.INVALID_SYNTAX_FUZZY_EDITS));
          }
          q = new FuzzyQueryNode(field, EscapeQuerySyntaxImpl.discardEscapeChar(term.image), fms, term.beginColumn, term.endColumn);
       } else if (regexp) {
         String re = term.image.substring(1, term.image.length()-1);
         q = new RegexpQueryNode(field, re, 0, re.length());
       }
     }
     | ( ( <RANGEIN_START> {startInc=true;} | <RANGEEX_START> )
         ( goop1=<RANGE_GOOP>|goop1=<RANGE_QUOTED>|goop1=<RANGE_TO> )
         ( <RANGE_TO> )
         ( goop2=<RANGE_GOOP>|goop2=<RANGE_QUOTED>|goop2=<RANGE_TO> )
         ( <RANGEIN_END> {endInc=true;} | <RANGEEX_END>))
       [ <CARAT> boost=<NUMBER> ]
        {
          if (goop1.kind == RANGE_QUOTED) {
            goop1.image = goop1.image.substring(1, goop1.image.length()-1);
          }
          if (goop2.kind == RANGE_QUOTED) {
            goop2.image = goop2.image.substring(1, goop2.image.length()-1);
          }
          
          qLower = new FieldQueryNode(field, 
                                   EscapeQuerySyntaxImpl.discardEscapeChar(goop1.image), goop1.beginColumn, goop1.endColumn);
      qUpper = new FieldQueryNode(field,
                                   EscapeQuerySyntaxImpl.discardEscapeChar(goop2.image), goop2.beginColumn, goop2.endColumn);
          q = new TermRangeQueryNode(qLower, qUpper, startInc ? true : false, endInc ? true : false);
        }
     | term=<QUOTED> {q = new QuotedFieldQueryNode(field, EscapeQuerySyntaxImpl.discardEscapeChar(term.image.substring(1, term.image.length()-1)), term.beginColumn + 1, term.endColumn - 1);}
       [ fuzzySlop=<FUZZY_SLOP> ]
       [ <CARAT> boost=<NUMBER> ]
       {       
         int phraseSlop = 0;

         if (fuzzySlop != null) {
           try {
             phraseSlop = (int)Float.parseFloat(fuzzySlop.image.substring(1));
             q = new SlopQueryNode(q, phraseSlop);    
           }
           catch (Exception ignored) {
            /* Should this be handled somehow? (defaults to "no PhraseSlop", if
           * slop number is invalid)
           */
           }
         }
              
       }
  )
  {
    if (boost != null) {
      float f = (float)1.0;
      try {
        f = Float.parseFloat(boost.image);
        // avoid boosting null queries, such as those caused by stop words
          if (q != null) {
            q = new BoostQueryNode(q, f);
          }
      } catch (Exception ignored) {
        /* Should this be handled somehow? (defaults to "no boost", if
           * boost number is invalid)
           */
      }
    }
      return q;
  }
}