net.sourceforge.nite.search.rewriter
Class DefaultQueryRewriter

java.lang.Object
  net.sourceforge.nite.search.rewriter.DefaultQueryRewriter

All Implemented Interfaces:

ParserTreeConstants, QueryRewriter, RewriterConstants

public class DefaultQueryRewriter
extends java.lang.Object
implements QueryRewriter, ParserTreeConstants, RewriterConstants

The default implementation of the QueryRewriter, implementing our first guess at a set of rewriting rules.

We implement the ParserTreeConstants so we can use them in creating Nodes, and RewriterConstants so we can use the sample queries in testing via the main method.

Author:

nmayo@inf.ed.ac.uk

Field Summary

Fields inherited from interface net.sourceforge.nite.search.ParserTreeConstants

JJTCONDITION, JJTLOGICAL, jjtNodeName, JJTQUERY, JJTSIMPLEQUERY

Fields inherited from interface net.sourceforge.nite.search.rewriter.RewriterConstants

BREAK_LINE, TEST_QUERIES

Constructor Summary

DefaultQueryRewriter()
Create a DefaultQueryRewriter with (optionally) a SearchableCorpus attached.

DefaultQueryRewriter(SearchableCorpus sc)

Method Summary

static void	main(java.lang.String[] args) Main test method.
NodeQuery	rewrite(NodeQuery query) Rewrite the query plan according to the rules of our algorithm.
java.util.List	rewriteResult(java.util.List result) Take the result tree and rewrite it to the form expected given the original input query.

Methods inherited from class java.lang.Object

equals, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

DefaultQueryRewriter

public DefaultQueryRewriter()

Create a DefaultQueryRewriter with (optionally) a SearchableCorpus attached.

DefaultQueryRewriter

public DefaultQueryRewriter(SearchableCorpus sc)

Method Detail

main

public static void main(java.lang.String[] args)

Main test method. Instantiates the class without a SearchableCorpus, just running the rewrite() method on a parsed query string. The NodeQuery query plan is rewritten according to the algorithm.

rewrite

public NodeQuery rewrite(NodeQuery query)

Rewrite the query plan according to the rules of our algorithm. Ignore exists quantifier, and split query into complex composition of more concise simple queries with, where possible:

• one variable binding per sub query
• tests evaluated on earliset subquery possible
• component queries sequenced by increasing order of (estimated) match set size

In terms of the dnf vector containing the query plan: Redistribute sets of ANDed NodeConditions to single variables where possible, joining them with NodeLogicals. Keep conditions which are joined by implication or OR together. Make all exists booleans false.

TO DO: Meanwhile record a mapping for the results output. TO DO: We may need to reorder forall booleans?

Specified by:

rewrite in interface QueryRewriter

Parameters:

query - the NodeQuery which is to be rewritten

Returns:

the rewritten query

rewriteResult

public java.util.List rewriteResult(java.util.List result)

Take the result tree and rewrite it to the form expected given the original input query.

Specified by:

rewriteResult in interface QueryRewriter

Parameters:

result - the results of the query