Page MenuHomePhorge
Files F16570998

No OneTemporary
Actions

View Options

diff --git a/nepomuk/query/queryparser.cpp b/nepomuk/query/queryparser.cpp
index 3bdcb257c8..fe215d3094 100644
--- a/nepomuk/query/queryparser.cpp
+++ b/nepomuk/query/queryparser.cpp
@@ -1,755 +1,782 @@
/*
This file is part of the Nepomuk KDE project.
Copyright (C) 2007-2011 Sebastian Trueg <trueg@kde.org>
This library is free software; you can redistribute it and/or
modify it under the terms of the GNU Library General Public
License version 2 as published by the Free Software Foundation.
This library is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
Library General Public License for more details.
You should have received a copy of the GNU Library General Public License
along with this library; see the file COPYING.LIB. If not, write to
the Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor,
Boston, MA 02110-1301, USA.
*/
#include "queryparser.h"
#include "query.h"
#include "query_p.h"
#include "literalterm.h"
#include "resourceterm.h"
#include "andterm.h"
#include "orterm.h"
#include "negationterm.h"
#include "comparisonterm.h"
#include "dateparser_p.h"
#include "nfo.h"
#include <QtCore/QRegExp>
#include <QtCore/QSet>
#include <QtCore/QMutex>
#include <QtCore/QMutexLocker>
#include <kdebug.h>
#include <klocale.h>
#include "resourcemanager.h"
#include "property.h"
#include "literal.h"
#include <Soprano/Node>
#include <Soprano/Model>
#include <Soprano/QueryResultIterator>
#include <Soprano/Vocabulary/RDFS>
#include <Soprano/Vocabulary/RDF>
using namespace Nepomuk::Query;
namespace {
Nepomuk::Query::ComparisonTerm::Comparator fieldTypeRelationFromString( const QString& s ) {
if ( s == "=" ) {
return Nepomuk::Query::ComparisonTerm::Equal;
}
else if ( s == ":" ) {
return Nepomuk::Query::ComparisonTerm::Contains;
}
else if ( s == ">" ) {
return Nepomuk::Query::ComparisonTerm::Greater;
}
else if ( s == "<" ) {
return Nepomuk::Query::ComparisonTerm::Smaller;
}
else if ( s == ">=" ) {
return Nepomuk::Query::ComparisonTerm::GreaterOrEqual;
}
else if ( s == "<=" ) {
return Nepomuk::Query::ComparisonTerm::SmallerOrEqual;
}
else {
kDebug() << "FIXME: Unsupported relation:" << s;
return Nepomuk::Query::ComparisonTerm::Equal;
}
}
QString stripQuotes( const QString& s, bool* hadQuotes = 0 ) {
if ( s[0] == '\'' ||
s[0] == '\"' ) {
if( hadQuotes )
*hadQuotes = true;
return s.mid( 1 ).left( s.length()-2 );
}
else {
if( hadQuotes )
*hadQuotes = false;
return s;
}
}
QUrl tryToBeIntelligentAboutParsingUrl( const QString& s ) {
if ( s.contains( '%' ) && !s.contains( '/' ) ) {
return QUrl::fromEncoded( s.toAscii() );
}
else {
return QUrl( s );
}
}
Soprano::LiteralValue createLiteral( const QString& s, bool globbing ) {
// 1. check if it is a number
QString clearString(s);
clearString.remove(QLatin1Char('\''));
clearString.remove(QLatin1Char('"'));
bool b = false;
int i = clearString.toInt( &b );
if ( b )
return Soprano::LiteralValue( i );
double d = clearString.toDouble( &b );
if ( b )
return Soprano::LiteralValue( d );
// 2. no number - continue with the original string
// no globbing if we have quotes or if there already is a wildcard
if ( s[0] == QLatin1Char('\'') ||
s[0] == QLatin1Char('\"') ) {
return s;
}
//
// we can only do query term globbing for strings longer than 3 chars
//
if( globbing && s.length() > 3 && !s.endsWith('*') && !s.endsWith('?') )
return QString(s + '*');
else
return s;
}
bool positiveTerm( const QString& s ) {
if(s.isEmpty())
return true;
else if(s == "+")
return true;
else if(s == "-")
return false;
else //unrecognized capture
return true;
}
- // TODO do this dorectly while parsing (but only once we have a real parser based on bison or whatever
- Nepomuk::Query::Term resolveFields( const Nepomuk::Query::Term& term, const Nepomuk::Query::QueryParser* parser )
- {
- switch( term.type() ) {
- case Nepomuk::Query::Term::And:
- case Nepomuk::Query::Term::Or: {
- QList<Nepomuk::Query::Term> newSubTerms;
- foreach( const Nepomuk::Query::Term& t, static_cast<const Nepomuk::Query::GroupTerm&>( term ).subTerms() ) {
- Nepomuk::Query::Term resolvedTerm = resolveFields(t, parser);
- if ( resolvedTerm.isValid() )
- newSubTerms << resolvedTerm;
- else
- return Nepomuk::Query::Term();
- }
- if ( term.isAndTerm() )
- return Nepomuk::Query::AndTerm( newSubTerms );
- else
- return Nepomuk::Query::OrTerm( newSubTerms );
- }
-
-
- case Nepomuk::Query::Term::Negation: {
- return Nepomuk::Query::NegationTerm::negateTerm( resolveFields( term.toNegationTerm().subTerm(), parser ) );
- }
-
-
- case Nepomuk::Query::Term::Comparison: {
- Nepomuk::Query::ComparisonTerm newTerm;
- newTerm.setComparator( term.toComparisonTerm().comparator() );
- newTerm.setProperty( term.toComparisonTerm().property() );
- newTerm.setSubTerm( resolveFields( term.toComparisonTerm().subTerm(), parser ) );
-
- // needLiteralRange states if we need a literal range. A value of false does not mean that we need a resource though.
- bool needLiteralRange = newTerm.subTerm().isLiteralTerm() && !newTerm.subTerm().toLiteralTerm().value().isString();
-
- // A very dumb test to see if the property is set or not: does the URI have a scheme.
- // With a proper parser and in-place property matching there will be no need for this anymore
- if ( newTerm.property().uri().scheme().isEmpty() ) {
- QList<Nepomuk::Types::Property> properties = parser->matchProperty( term.toComparisonTerm().property().uri().toString() );
- if ( properties.count() > 0 ) {
- if ( properties.count() == 1 ) {
- newTerm.setProperty( properties.first() );
- if(needLiteralRange) {
- QVariant v = newTerm.subTerm().toLiteralTerm().value().variant();
- v.convert(properties[0].literalRangeType().dataType());
- newTerm.setSubTerm(Nepomuk::Query::LiteralTerm(v));
- if(newTerm.comparator() == Nepomuk::Query::ComparisonTerm::Contains && v.type() != QVariant::String) {
- newTerm.setComparator(Nepomuk::Query::ComparisonTerm::Equal);
- }
- }
- return newTerm;
- }
- else {
- // we only use a max of 4 properties, otherwise the queries get too big
- // in addition we try to exclude properties which do not make sense:
- // - numerical values can never match a resource
- Nepomuk::Query::OrTerm orTerm;
- for( int i = 0; i < properties.count() && orTerm.subTerms().count() < 4; ++i ) {
- const Nepomuk::Types::Property& property = properties[i];
- if(!property.range().isValid() || !needLiteralRange) {
- QVariant v = newTerm.subTerm().toLiteralTerm().value().variant();
- if(property.range().isValid() ||
- v.convert(property.literalRangeType().dataType())) {
- Nepomuk::Query::ComparisonTerm t( newTerm );
- t.setProperty( property );
- t.setSubTerm(Nepomuk::Query::LiteralTerm(v));
- // only strings can be matched via bif:contains
- if(t.comparator() == Nepomuk::Query::ComparisonTerm::Contains && v.type() != QVariant::String) {
- t.setComparator(Nepomuk::Query::ComparisonTerm::Equal);
- }
- orTerm.addSubTerm( t );
- }
- }
- }
- if(!orTerm.isValid()) {
- // FIXME: actually the whole query might be invalid at this point.
- return Nepomuk::Query::Term();
- }
- return orTerm;
- }
- }
- else {
- return Nepomuk::Query::Term();
- }
- }
- }
-
- default:
- return term;
- }
- }
-
// A filename pattern needs to contain one dot and at least one '*' or '?':
// *.mp3
// hello?.txt
// hello?.*
// test*.???
bool isFilenamePattern( const QString& s )
{
return( !s.contains(' ') &&
s.count('.') == 1 &&
s.count('*') + s.count('?') > 0 );
}
Nepomuk::Query::ComparisonTerm createFilenamePatternTerm( const QString& s )
{
QString regex = QRegExp::escape(s);
regex.replace( "\\*", QLatin1String( ".*" ) );
regex.replace( "\\?", QLatin1String( "." ) );
regex.replace("\\", "\\\\");
regex.prepend('^');
regex.append('$');
return Nepomuk::Query::ComparisonTerm( Nepomuk::Vocabulary::NFO::fileName(),
Nepomuk::Query::LiteralTerm( regex ),
Nepomuk::Query::ComparisonTerm::Regexp );
}
/**
* Merging literal terms is an optimization which is based on the assumption that most
* users want to search for the full text terms they enter in the value of the same
* property.
* Since merging two literals "foo" and "bar" into one term "foo AND bar" effectively
* changes the result set (the former allows that "foo" occurs in a property value
* different from "bar" while the latter forces them to occur in the same.)
* But the resulting query is much faster.
*/
Nepomuk::Query::Term mergeLiteralTerms( const Nepomuk::Query::Term& term )
{
if( term.isAndTerm() ) {
AndTerm mergedTerm;
QStringList fullTextTerms;
Q_FOREACH( const Term& st, term.toAndTerm().subTerms() ) {
if( st.isLiteralTerm() ) {
fullTextTerms << st.toLiteralTerm().value().toString();
}
else {
mergedTerm.addSubTerm( st );
}
}
mergedTerm.addSubTerm( LiteralTerm( fullTextTerms.join( QString::fromLatin1(" AND ") ) ) );
return mergedTerm.optimized();
}
else {
return term;
}
}
+ /**
+ * Tries to setup the sub-term of the given comparisonterm according to its property.
+ * This can fail if the sub-term cannot be converted to the property's range.
+ *
+ * This is used in QueryParser::Private::resolveFields
+ */
+ bool setupComparisonTermSubTerm(Nepomuk::Query::ComparisonTerm& ct)
+ {
+ // property with resource range
+ if(ct.property().range().isValid()) {
+ if(ct.subTerm().isLiteralTerm()) {
+ // here we need a string. everything else does not work since we match labels
+ return ct.subTerm().toLiteralTerm().value().isString();
+ }
+ else {
+ return true;
+ }
+ }
+
+ // property with literal range
+ else if(ct.subTerm().isLiteralTerm()) {
+ QVariant v = ct.subTerm().toLiteralTerm().value().variant();
+ if(v.convert(ct.property().literalRangeType().dataType())) {
+ ct.setSubTerm(Nepomuk::Query::LiteralTerm(v));
+ // only strings can be matched via bif:contains
+ if(ct.comparator() == Nepomuk::Query::ComparisonTerm::Contains && v.type() != QVariant::String) {
+ ct.setComparator(Nepomuk::Query::ComparisonTerm::Equal);
+ }
+ return true;
+ }
+ }
+
+ // fallback
+ return false;
+ }
+
#ifndef Q_CC_MSVC
#warning Make the parser handle different data, time, and datetime encodings as well as suffixes like MB or GB
#endif
#if 0
QDateTime parseDateTime( const Soprano::LiteralValue& literal )
{
//TODO: change to DateTime parser once complete
Nepomuk::Search::DateParser date( literal.toString() );
if( date.hasDate() ) {
return QDateTime( date.getDate() );
}
else {
Nepomuk::Search::TimeParser time( literal.toString() );
if(time.hasTime() )
return QDateTime(QDate::currentDate(), time.next() );
else
return QDateTime(); //return invalid datetime
}
}
Soprano::LiteralValue parseSizeType( const Soprano::LiteralValue& literal )
{
const double KiB = 1024.0;
const double MiB = KiB * 1024.0;
const double GiB = MiB * 1024.0;
const double TiB = GiB * 1024.0;
const double KB = 1000.0;
const double MB = KB * 1000.0;
const double GB = MB * 1000.0;
const double TB = GB * 1000.0;
QHash<QString, double> sizes;
sizes.insert( "KiB", KiB );
sizes.insert( "MiB", MiB );
sizes.insert( "GiB", GiB );
sizes.insert( "TiB", TiB );
sizes.insert( "KB", KB );
sizes.insert( "MB", MB );
sizes.insert( "GB", GB );
sizes.insert ("TB", TB );
for ( QHash<QString, double>::const_iterator i = sizes.constBegin();
i != sizes.constEnd(); ++i ) {
QRegExp cur( QString("^([\\d]+.?[\\d]*)[\\s]*%1$").arg( i.key() ) );
if( cur.indexIn( literal.toString() ) != -1 ) {
double value = cur.cap( 1 ).toDouble();
double newValue = value * i.value();
kDebug() << "Found value" << value << i.key() << "->" << newValue;
return Soprano::LiteralValue( newValue );
}
}
return literal;
}
Nepomuk::Query::Term resolveLiteralValues( const Nepomuk::Query::Term& term )
{
switch( term.type() ) {
case Nepomuk::Query::Term::Comparison: {
Nepomuk::Query::ComparisonTerm cterm = term.toComparisonTerm();
Nepomuk::Types::Property p( cterm.property() );
if ( p.literalRangeType().isValid() ) {
Q_ASSERT( cterm.subTerm().isLiteralTerm() );
Nepomuk::Query::ComparisonTerm newTerm;
newTerm.setComparator( cterm.comparator() );
newTerm.setProperty( QUrl(cterm.property()) );
// now try to resolve the literal
const Nepomuk::Query::LiteralTerm subTerm = cterm.subTerm().toLiteralTerm();
const Nepomuk::Types::Literal lt = p.literalRangeType();
if ( lt.dataType() == QVariant::DateTime &&
!subTerm.value().isDateTime() ) {
QDateTime dateTime = parseDateTime( subTerm.value() );
if ( dateTime.isValid() ) {
newTerm.setSubTerm( Nepomuk::Query::LiteralTerm( dateTime ) );
return newTerm;
}
}
else if ( lt.dataType() == QVariant::Int &&
!subTerm.value().isInt() ) {
newTerm.setSubTerm( Nepomuk::Query::LiteralTerm( parseSizeType( subTerm.value() ) ) );
return newTerm;
}
}
return term;
}
case Nepomuk::Query::Term::And:
case Nepomuk::Query::Term::Or: {
QList<Nepomuk::Query::Term> newSubTerms;
foreach( const Nepomuk::Query::Term& t, static_cast<const Nepomuk::Query::GroupTerm&>( term ).subTerms() ) {
newSubTerms << resolveLiteralValues(t);
}
if ( term.isAndTerm() )
return Nepomuk::Query::AndTerm( newSubTerms );
else
return Nepomuk::Query::OrTerm( newSubTerms );
}
default:
return term;
}
}
#endif
// a field differs from a plain term in that it does never allow comparators
const char* s_fieldNamePattern = "([^\\s\"':=<>]+|(?:([\"'])[^\"':=<>]+\\%1))";
const char* s_plainTermPattern = "([^\\s\"':=<>]+|(?:([\"'])[^\"']+\\%1))";
const char* s_inExclusionPattern = "([\\+\\-]?)";
const char* s_uriPattern = "<([^<>]+)>";
const char* s_comparatorPattern = "(:|\\<=|\\>=|=|\\<|\\>)";
/**
* Creating QRegExp is expensive, copying them is cheap. Thus, we keep static
* instances of the regexps around which we only have to create once.
*/
class QueryParserRegExpPool
{
public:
QueryParserRegExpPool()
: plainTermRx( QLatin1String(s_inExclusionPattern)
+ QString::fromLatin1(s_plainTermPattern).arg( 3 ) ),
fieldRx( QLatin1String(s_inExclusionPattern)
+ QString::fromLatin1(s_fieldNamePattern).arg( 3 )
+ QLatin1String(s_comparatorPattern)
+ QString::fromLatin1(s_plainTermPattern).arg( 6 ) ),
propertyRx( QLatin1String(s_inExclusionPattern)
+ QLatin1String(s_uriPattern)
+ QLatin1String(s_comparatorPattern)
+ QString::fromLatin1(s_plainTermPattern).arg( 5 ) ),
resourceRx( QLatin1String(s_inExclusionPattern)
+ QLatin1String(s_uriPattern)
+ QLatin1String("(?::|=)")
+ QLatin1String(s_uriPattern) ),
fieldFieldRx( QLatin1String(s_inExclusionPattern)
+ QString::fromLatin1(s_fieldNamePattern).arg( 3 )
+ QLatin1String(s_comparatorPattern)
+ QLatin1String("\\(")
+ QString::fromLatin1(s_fieldNamePattern).arg( 6 )
+ QLatin1String(s_comparatorPattern)
+ QString::fromLatin1(s_plainTermPattern).arg( 9 )
+ QLatin1String("\\)") )
{
}
// match a simple search text
// captures: 1 - The optional + or - sign (may be empty)
// 2 - the search text (including optional paranthesis)
QRegExp plainTermRx;
// match a field search term: fieldname + relation (:, =, etc) + search text with optional paranthesis
// captures: 1 - The optional + or - sign (may be empty)
// 2 - fieldname
// 3 - relation
// 4 - search text (including optional paranthesis)
QRegExp fieldRx;
// match a property URI search term: property URI + relation (:, =, etc) + search text with optional paranthesis
// captures: 1 - The optional + or - sign (may be empty)
// 2 - property URI
// 3 - relation
// 4 - search text (including optional paranthesis)
QRegExp propertyRx;
// match a property URI search term: property URI + relation (:, =, etc) + resource URI
// captures: 1 - The optional + or - sign (may be empty)
// 2 - property URI
// 3 - resource URI
QRegExp resourceRx;
QRegExp fieldFieldRx;
};
// the one global instance used for the statis QueryParser methods
K_GLOBAL_STATIC( QueryParserRegExpPool, s_regExpPool )
}
class Nepomuk::Query::QueryParser::Private
{
public:
+ QueryParser* q;
+
QSet<QString> andKeywords;
QSet<QString> orKeywords;
mutable QHash<QString, QList<Types::Property> > fieldMatchCache;
QMutex fieldMatchCacheMutex;
+
+ /// set by resolveFields if the query is in fact invalid
+ bool m_invalidQuery;
+
+ /**
+ * Resolve the fields in all ComparisonTerms by looking up possible matching
+ * properties.
+ * This method will set m_invalidQuery in case there is one term for which
+ * no property can be matched.
+ */
+ Nepomuk::Query::Term resolveFields( const Nepomuk::Query::Term& term );
};
+
+Term QueryParser::Private::resolveFields(const Term &term)
+{
+ switch( term.type() ) {
+ case Nepomuk::Query::Term::And:
+ case Nepomuk::Query::Term::Or: {
+ QList<Nepomuk::Query::Term> newSubTerms;
+ foreach( const Nepomuk::Query::Term& t, static_cast<const Nepomuk::Query::GroupTerm&>( term ).subTerms() ) {
+ Nepomuk::Query::Term resolvedTerm = resolveFields(t);
+ if ( resolvedTerm.isValid() )
+ newSubTerms << resolvedTerm;
+ else
+ return Nepomuk::Query::Term();
+ }
+ if ( term.isAndTerm() )
+ return Nepomuk::Query::AndTerm( newSubTerms );
+ else
+ return Nepomuk::Query::OrTerm( newSubTerms );
+ }
+
+
+ case Nepomuk::Query::Term::Negation: {
+ return Nepomuk::Query::NegationTerm::negateTerm( resolveFields( term.toNegationTerm().subTerm() ) );
+ }
+
+
+ case Nepomuk::Query::Term::Comparison: {
+ Nepomuk::Query::ComparisonTerm newTerm;
+ newTerm.setComparator( term.toComparisonTerm().comparator() );
+ newTerm.setProperty( term.toComparisonTerm().property() );
+ newTerm.setSubTerm( resolveFields( term.toComparisonTerm().subTerm() ) );
+
+ // A very dumb test to see if the property is set or not: does the URI have a scheme.
+ // With a proper parser and in-place property matching there will be no need for this anymore
+ if ( newTerm.property().uri().scheme().isEmpty() ) {
+ QList<Nepomuk::Types::Property> properties = q->matchProperty( term.toComparisonTerm().property().uri().toString() );
+ // we only use a max of 4 properties, otherwise the queries get too big
+ // in addition we try to exclude properties which do not make sense:
+ // - numerical values can never match a resource
+ Nepomuk::Query::OrTerm orTerm;
+ for( int i = 0; i < properties.count() && orTerm.subTerms().count() < 4; ++i ) {
+ const Nepomuk::Types::Property& property = properties[i];
+ Nepomuk::Query::ComparisonTerm t( newTerm );
+ t.setProperty( property );
+ if(setupComparisonTermSubTerm(t)) {
+ orTerm.addSubTerm( t );
+ }
+ }
+ if(!orTerm.isValid()) {
+ m_invalidQuery = true;
+ return Nepomuk::Query::Term();
+ }
+ return orTerm.optimized();
+ }
+ }
+
+ default:
+ return term;
+ }
+}
+
+
Nepomuk::Query::QueryParser::QueryParser()
: d( new Private() )
{
+ d->q = this;
+
QString andListStr = i18nc( "Boolean AND keyword in desktop search strings. "
"You can add several variants separated by spaces, "
"e.g. retain the English one alongside the translation; "
"keywords are not case sensitive. Make sure there is "
"no conflict with the OR keyword.",
"and" );
foreach ( const QString &andKeyword, andListStr.split( ' ', QString::SkipEmptyParts ) ) {
d->andKeywords.insert( andKeyword.toLower() );
}
QString orListStr = i18nc( "Boolean OR keyword in desktop search strings. "
"You can add several variants separated by spaces, "
"e.g. retain the English one alongside the translation; "
"keywords are not case sensitive. Make sure there is "
"no conflict with the AND keyword.",
"or" );
foreach ( const QString &orKeyword, orListStr.split( ' ', QString::SkipEmptyParts ) ) {
d->orKeywords.insert( orKeyword.toLower() );
}
}
Nepomuk::Query::QueryParser::~QueryParser()
{
delete d;
}
QList<Nepomuk::Types::Property> Nepomuk::Query::QueryParser::matchProperty( const QString& fieldName ) const
{
kDebug() << fieldName;
QMutexLocker lock( &d->fieldMatchCacheMutex );
QHash<QString, QList<Types::Property> >::ConstIterator it = d->fieldMatchCache.constFind( fieldName );
if( it != d->fieldMatchCache.constEnd() ) {
return it.value();
}
else {
lock.unlock();
QList<Nepomuk::Types::Property> results;
//
// Due to the limited number of properties in the database a REGEX filter
// is actually faster than a fulltext query via bif:contains (this is what
// experiments showed).
//
QString query = QString( "select distinct ?p where { "
"graph ?g { "
"?p a %1 . "
"?p %2 ?l . "
"FILTER(REGEX(STR(?l),'%3*','i') || REGEX(STR(?p),'%3*','i')) . "
"} "
"}" )
.arg( Soprano::Node::resourceToN3( Soprano::Vocabulary::RDF::Property() ) )
.arg( Soprano::Node::resourceToN3( Soprano::Vocabulary::RDFS::label() ) )
.arg( fieldName );
kDebug() << "Match query:" << query;
Soprano::QueryResultIterator labelHits
= Nepomuk::ResourceManager::instance()->mainModel()->executeQuery( query, Soprano::Query::QueryLanguageSparql );
while ( labelHits.next() ) {
QUrl property = labelHits.binding( "p" ).uri();
results << property;
kDebug() << "Found property match" << property;
}
lock.relock();
d->fieldMatchCache.insert( fieldName, results );
return results;
}
}
Nepomuk::Query::Query Nepomuk::Query::QueryParser::parse( const QString& query ) const
{
return parse( query, NoParserFlags );
}
Nepomuk::Query::Query Nepomuk::Query::QueryParser::parse( const QString& query, ParserFlags flags ) const
{
// TODO: a "real" parser which can handle all of the Xesam user language
// This one for example does not handle nesting at all.
Nepomuk::Query::Query final;
QList<Term> terms;
bool inOrBlock = false;
bool inAndBlock = false;
int pos = 0;
// create local copies of the regexps for thread safety purposes
const QRegExp resourceRx = s_regExpPool->resourceRx;
const QRegExp propertyRx = s_regExpPool->propertyRx;
const QRegExp fieldFieldRx = s_regExpPool->fieldFieldRx;
const QRegExp fieldRx = s_regExpPool->fieldRx;
const QRegExp plainTermRx = s_regExpPool->plainTermRx;
while ( pos < query.length() ) {
// skip whitespace
while ( pos < query.length() && query[pos].isSpace() ) {
kDebug() << "Skipping space at" << pos;
++pos;
}
Term term;
if ( pos < query.length() ) {
if ( resourceRx.indexIn( query, pos ) == pos ) {
kDebug() << "matched resource term at" << pos << resourceRx.cap( 0 );
term = ComparisonTerm( tryToBeIntelligentAboutParsingUrl( resourceRx.cap( 2 ) ),
ResourceTerm( tryToBeIntelligentAboutParsingUrl( resourceRx.cap( 3 ) ) ),
ComparisonTerm::Equal );
if ( !positiveTerm( resourceRx.cap( 1 ) ) ) {
term = NegationTerm::negateTerm( term );
}
pos += resourceRx.matchedLength();
}
else if ( propertyRx.indexIn( query, pos ) == pos ) {
kDebug() << "matched property term at" << pos << propertyRx.cap( 0 );
ComparisonTerm ct;
ct.setProperty( tryToBeIntelligentAboutParsingUrl( propertyRx.cap( 2 ) ) );
ct.setSubTerm( LiteralTerm( createLiteral( propertyRx.cap( 4 ), flags&QueryTermGlobbing ) ) );
QString comparator = propertyRx.cap( 3 );
ct.setComparator( fieldTypeRelationFromString( comparator ) );
pos += propertyRx.matchedLength();
if ( !positiveTerm(propertyRx.cap( 1 ) ) ) {
term = NegationTerm::negateTerm( ct );
}
else {
term = ct;
}
}
else if ( fieldFieldRx.indexIn( query, pos ) == pos ) {
kDebug() << "matched field field term at" << pos
<< fieldFieldRx.cap( 0 )
<< fieldFieldRx.cap( 2 )
<< fieldFieldRx.cap( 4 )
<< fieldFieldRx.cap( 5 )
<< fieldFieldRx.cap( 7 )
<< fieldFieldRx.cap( 8 );
ComparisonTerm ct;
ct.setProperty( QUrl(stripQuotes( fieldFieldRx.cap( 2 ) )) );
QString comparator = fieldFieldRx.cap( 4 );
ct.setComparator( fieldTypeRelationFromString( comparator ) );
ct.setSubTerm( ComparisonTerm( QUrl(stripQuotes( fieldFieldRx.cap( 5 ) )),
LiteralTerm( createLiteral( fieldFieldRx.cap( 8 ), flags&QueryTermGlobbing ) ),
fieldTypeRelationFromString( fieldFieldRx.cap( 7 ) ) ) );
pos += fieldFieldRx.matchedLength();
if ( !positiveTerm( fieldFieldRx.cap( 1 ) ) ) {
term = NegationTerm::negateTerm( ct );
}
else {
term = ct;
}
}
else if ( fieldRx.indexIn( query, pos ) == pos ) {
kDebug() << "matched field term at" << pos << fieldRx.cap( 0 ) << fieldRx.cap( 2 ) << fieldRx.cap( 4 ) << fieldRx.cap( 5 );
if( stripQuotes ( fieldRx.cap( 2 ) ).compare( QString( "inFolder" ), Qt::CaseInsensitive ) == 0 ) {
KUrl url( fieldRx.cap( 5 ) );
kDebug() << "found include path" << url;
FileQuery fileQuery(final);
if ( positiveTerm( fieldRx.cap( 1 ) ) )
fileQuery.addIncludeFolder(url);
else
fileQuery.addExcludeFolder(url);
final = fileQuery;
pos += fieldRx.matchedLength();
}
else {
ComparisonTerm ct;
ct.setProperty( QUrl( stripQuotes( fieldRx.cap( 2 ) ) ) );
ct.setSubTerm( LiteralTerm( createLiteral( fieldRx.cap( 5 ), flags&QueryTermGlobbing ) ) );
QString comparator = fieldRx.cap( 4 );
ct.setComparator( fieldTypeRelationFromString( comparator ) );
pos += fieldRx.matchedLength();
if ( !positiveTerm(fieldRx.cap( 1 ) ) ) {
NegationTerm nt;
nt.setSubTerm( ct );
term = nt;
}
else {
term = ct;
}
}
}
else if ( plainTermRx.indexIn( query, pos ) == pos ) {
QString value = plainTermRx.cap( 2 );
if ( d->orKeywords.contains( value.toLower() ) ) {
inOrBlock = true;
}
else if ( d->andKeywords.contains( value.toLower() ) ) {
inAndBlock = true;
}
else {
kDebug() << "matched literal at" << pos << value;
if( flags&DetectFilenamePattern && isFilenamePattern(value) ) {
term = createFilenamePatternTerm( value );
}
else {
term = LiteralTerm( createLiteral( value, flags&QueryTermGlobbing ) );
}
if ( !positiveTerm(plainTermRx.cap( 1 ) ) ) {
term = NegationTerm::negateTerm( term );
}
}
pos += plainTermRx.matchedLength();
}
else {
kDebug() << "Invalid query at" << pos << query;
return Query();
}
if ( term.isValid() ) {
if ( inOrBlock && !terms.isEmpty() ) {
OrTerm orTerm;
orTerm.addSubTerm( terms.takeLast() );
orTerm.addSubTerm( term );
terms.append( orTerm );
}
else if ( inAndBlock && !terms.isEmpty() ) {
AndTerm andTerm;
andTerm.addSubTerm( terms.takeLast() );
andTerm.addSubTerm( term );
terms.append( andTerm );
}
else {
terms.append( term );
}
}
}
}
if ( terms.count() == 1 ) {
final.setTerm( terms[0] );
}
else if ( terms.count() > 0 ) {
AndTerm t;
t.setSubTerms( terms );
final.setTerm( t );
}
- final.setTerm( mergeLiteralTerms( resolveFields( final.term(), this ) ) );
- return final;
+ d->m_invalidQuery = false;
+ final.setTerm( mergeLiteralTerms( d->resolveFields( final.term() ) ) );
+ if(d->m_invalidQuery) {
+ return Query();
+ }
+ else {
+ return final;
+ }
}
// static
Nepomuk::Query::Query Nepomuk::Query::QueryParser::parseQuery( const QString& query )
{
QueryParser parser;
return parser.parse( query );
}
// static
Nepomuk::Query::Query Nepomuk::Query::QueryParser::parseQuery( const QString& query, ParserFlags flags )
{
QueryParser parser;
return parser.parse( query, flags );
}

File Metadata

Mime Type
text/x-diff
Expires
Fri, Nov 1, 9:49 AM (1 d, 19 h)
Storage Engine
blob
Storage Format
Raw Data
Storage Handle
10076160
Default Alt Text
(34 KB)

Event Timeline