/*
* 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.
*/
package org.exoplatform.services.jcr.impl.core.query.lucene;
import org.apache.lucene.index.IndexReader;
import org.apache.lucene.index.Term;
import org.apache.lucene.index.TermDocs;
import org.apache.lucene.index.TermEnum;
import org.apache.lucene.index.TermPositions;
import java.io.IOException;
import java.util.HashMap;
import java.util.Map;
import java.util.WeakHashMap;
import javax.jcr.PropertyType;
/**
* Implements a variant of the lucene class <code>org.apache.lucene.search.FieldCacheImpl</code>.
* The lucene FieldCache class has some sort of support for custom comparators
* but it only works on the basis of a field name. There is no further control
* over the terms to iterate, that's why we use our own implementation.
*/
public class SharedFieldCache
{
/**
* Expert: Stores term text values and document ordering data.
*/
public static class ValueIndex
{
/**
* Some heuristic factor that determines whether the array is sparse. Note that if less then
* 1% is set, we already count the array as sparse. This is because it will become memory consuming
* quickly by keeping the (sparse) arrays
*/
private static final int SPARSE_FACTOR = 100;
/**
* Values indexed by document id.
*/
private final Comparable<?>[] values;
/**
* Values (Comparable) map indexed by document id.
*/
public final Map<Integer, Comparable<?>> valuesMap;
/**
* Boolean indicating whether the {@link #valuesMap} impl has to be used
*/
public final boolean sparse;
/**
* Creates one of these objects
*/
public ValueIndex(Comparable<?>[] values, int setValues)
{
if (isSparse(values, setValues))
{
this.sparse = true;
this.values = null;
if (setValues == 0)
{
this.valuesMap = null;
}
else
{
this.valuesMap = getValuesMap(values, setValues);
}
}
else
{
this.sparse = false;
this.values = values;
this.valuesMap = null;
}
}
public Comparable<?> getValue(int i)
{
if (sparse)
{
return valuesMap == null ? null : valuesMap.get(i);
}
else
{
return values[i];
}
}
private static Map<Integer, Comparable<?>> getValuesMap(Comparable<?>[] values, int setValues)
{
Map<Integer, Comparable<?>> map = new HashMap<Integer, Comparable<?>>(setValues);
for (int i = 0; i < values.length && setValues > 0; i++)
{
if (values[i] != null)
{
map.put(i, values[i]);
setValues--;
}
}
return map;
}
private static boolean isSparse(Comparable<?>[] values, int setValues)
{
// some really simple test to test whether the array is sparse.
// Currently, when less then 1% is set, the array is already sparse
// for this typical cache to avoid memory issues
if (setValues * SPARSE_FACTOR < values.length)
{
return true;
}
return false;
}
}
/**
* Reference to the single instance of <code>SharedFieldCache</code>.
*/
public static final SharedFieldCache INSTANCE = new SharedFieldCache();
/**
* The internal cache. Maps Entry to array of interpreted term values.
*/
private final Map<IndexReader, Map<Key, ValueIndex>> cache = new WeakHashMap<IndexReader, Map<Key, ValueIndex>>();
/**
* Private constructor.
*/
private SharedFieldCache()
{
}
/**
* Creates a <code>ValueIndex</code> for a <code>field</code> and a term
* <code>prefix</code>. The term prefix acts as the property name for the
* shared <code>field</code>.
* <br>
* This method is an adapted version of: <code>FieldCacheImpl.getStringIndex()</code>
*
* @param reader the <code>IndexReader</code>.
* @param field name of the shared field.
* @param prefix the property name, will be used as term prefix.
* @return a ValueIndex that contains the field values and order
* information.
* @throws IOException if an error occurs while reading from the index.
*/
public ValueIndex getValueIndex(IndexReader reader, String field, String prefix) throws IOException
{
if (reader instanceof ReadOnlyIndexReader)
{
reader = ((ReadOnlyIndexReader)reader).getBase();
}
field = field.intern();
ValueIndex ret = lookup(reader, field, prefix);
if (ret == null)
{
Comparable<?>[] retArray = new Comparable[reader.maxDoc()];
int setValues = 0;
if (retArray.length > 0)
{
IndexFormatVersion version = IndexFormatVersion.getVersion(reader);
boolean hasPayloads = version.isAtLeast(IndexFormatVersion.V3);
TermDocs termDocs;
byte[] payload = null;
int type;
if (hasPayloads)
{
termDocs = reader.termPositions();
payload = new byte[1];
}
else
{
termDocs = reader.termDocs();
}
TermEnum termEnum = reader.terms(new Term(field, prefix));
int prefixLength = prefix.length();
try
{
if (termEnum.term() == null)
{
throw new RuntimeException("no terms in field " + field);
}
do
{
Term term = termEnum.term();
String text;
if (term.field() != field || !(text = term.text()).startsWith(prefix))
{
break;
}
String value = text.substring(prefixLength);
termDocs.seek(termEnum);
while (termDocs.next())
{
type = PropertyType.UNDEFINED;
if (hasPayloads)
{
TermPositions termPos = (TermPositions)termDocs;
termPos.nextPosition();
if (termPos.isPayloadAvailable())
{
payload = termPos.getPayload(payload, 0);
type = payload[0];
}
}
setValues++;
retArray[termDocs.doc()] = getValue(value, type);
}
}
while (termEnum.next());
}
finally
{
termDocs.close();
termEnum.close();
}
}
ValueIndex value = new ValueIndex(retArray, setValues);
store(reader, field, prefix, value);
return value;
}
return ret;
}
/**
* See if a <code>ValueIndex</code> object is in the cache.
*/
ValueIndex lookup(IndexReader reader, String field, String prefix)
{
synchronized (cache)
{
Map<Key, ValueIndex> readerCache = cache.get(reader);
if (readerCache == null)
{
return null;
}
return readerCache.get(new Key(field, prefix));
}
}
/**
* Put a <code>ValueIndex</code> <code>value</code> to cache.
*/
void store(IndexReader reader, String field, String prefix, ValueIndex value)
{
synchronized (cache)
{
Map<Key, ValueIndex> readerCache = cache.get(reader);
if (readerCache == null)
{
readerCache = new HashMap<Key, ValueIndex>();
cache.put(reader, readerCache);
}
readerCache.put(new Key(field, prefix), value);
}
}
/**
* Returns a comparable for the given <code>value</code> that is read from
* the index.
*
* @param value the value as read from the index.
* @param type the property type.
* @return a comparable for the <code>value</code>.
*/
private Comparable<?> getValue(String value, int type)
{
switch (type)
{
case PropertyType.BOOLEAN :
return ComparableBoolean.valueOf(Boolean.valueOf(value).booleanValue());
case PropertyType.DATE :
return new Long(DateField.stringToTime(value));
case PropertyType.LONG :
return new Long(LongField.stringToLong(value));
case PropertyType.DOUBLE :
return new Double(DoubleField.stringToDouble(value));
default :
return value;
}
}
/**
* A compound <code>Key</code> that consist of <code>field</code>
* and <code>prefix</code> .
*/
static class Key
{
private final String field;
private final String prefix;
/**
* Creates <code>Key</code> for ValueIndex lookup.
*/
Key(String field, String prefix)
{
this.field = field.intern();
this.prefix = prefix.intern();
}
/**
* Returns <code>true</code> if <code>o</code> is a <code>Key</code>
* instance and refers to the same field and prefix.
*/
public boolean equals(Object o)
{
if (o instanceof Key)
{
Key other = (Key)o;
return other.field == field && other.prefix == prefix ;
}
return false;
}
/**
* Composes a hashcode based on the field and prefix.
*/
public int hashCode()
{
return field.hashCode() ^ prefix.hashCode();
}
}
}