com.bigdata.rdf.lexicon.BlobsWriteProc Maven / Gradle / Ivy
/**
Copyright (C) SYSTAP, LLC DBA Blazegraph 2006-2016. All rights reserved.
Contact:
SYSTAP, LLC DBA Blazegraph
2501 Calvert ST NW #106
Washington, DC 20008
[email protected]
This program is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation; version 2 of the License.
This program 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 General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program; if not, write to the Free Software
Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
*/
/*
* Created on May 21, 2007
*/
package com.bigdata.rdf.lexicon;
import java.io.Externalizable;
import java.io.IOException;
import java.io.ObjectInput;
import java.io.ObjectOutput;
import org.apache.log4j.Logger;
import org.openrdf.model.Value;
import com.bigdata.btree.IIndex;
import com.bigdata.btree.keys.IKeyBuilder;
import com.bigdata.btree.keys.KeyBuilder;
import com.bigdata.btree.proc.AbstractKeyArrayIndexProcedure;
import com.bigdata.btree.proc.AbstractKeyArrayIndexProcedureConstructor;
import com.bigdata.btree.proc.AbstractLocalSplitResultAggregator;
import com.bigdata.btree.proc.IParallelizableIndexProcedure;
import com.bigdata.btree.proc.IResultHandler;
import com.bigdata.btree.proc.SplitValuePair;
import com.bigdata.btree.raba.IRaba;
import com.bigdata.btree.raba.codec.IRabaCoder;
import com.bigdata.io.DataOutputBuffer;
import com.bigdata.io.LongPacker;
import com.bigdata.io.ShortPacker;
import com.bigdata.rdf.internal.IV;
import com.bigdata.rdf.internal.VTE;
import com.bigdata.rdf.internal.impl.AbstractIV;
import com.bigdata.rdf.internal.impl.BlobIV;
import com.bigdata.rdf.lexicon.BlobsWriteProc.Result;
import com.bigdata.relation.IMutableRelationIndexWriteProcedure;
import com.bigdata.service.Split;
/**
* This unisolated operation inserts {@link Value}s into the
* {termCode,hash(Value),counter}:Value index, assigning {@link IV}s to
* {@link Value}s as a side-effect.
*
* Note that it is perfectly possible that a concurrent client will overlap in
* the terms being inserted. The results will always be fully consistent if the
* rules of the road are observed since (a) unisolated operations are
* single-threaded; and (b) {@link IV}s are assigned in an unisolated atomic
* operation by {@link BlobsWriteProc}.
*
* @author Bryan Thompson
*/
public class BlobsWriteProc extends AbstractKeyArrayIndexProcedure implements
IParallelizableIndexProcedure, IMutableRelationIndexWriteProcedure {
private static final Logger log = Logger.getLogger(BlobsWriteProc.class);
/**
*
*/
private static final long serialVersionUID = 1L;
/**
* Serialized as extended metadata. When true unknown terms
* are NOT added to the database.
*/
private boolean readOnly;
@Override
public final boolean isReadOnly() {
return readOnly;
}
/**
* Serialized as extended metadata. When true blank nodes
* are stored in the lexicon's forward index.
*/
private boolean toldBNodes;
public final boolean isToldBNodes() {
return toldBNodes;
}
/**
* De-serialization constructor.
*/
public BlobsWriteProc() {
}
protected BlobsWriteProc(final IRabaCoder keySer, final IRabaCoder valSer,
final int fromIndex, final int toIndex, final byte[][] keys,
final byte[][] vals, final boolean readOnly,
final boolean storeBlankNodes) {
super(keySer, valSer, fromIndex, toIndex, keys, vals);
this.readOnly = readOnly;
this.toldBNodes = storeBlankNodes;
}
public static class BlobsWriteProcConstructor extends
AbstractKeyArrayIndexProcedureConstructor {
private final boolean readOnly;
private final boolean toldBNodes;
@Override
public final boolean sendValues() {
return true;
}
public BlobsWriteProcConstructor(final boolean readOnly,
final boolean toldBNodes) {
this.readOnly = readOnly;
this.toldBNodes = toldBNodes;
}
@Override
public BlobsWriteProc newInstance(final IRabaCoder keySer,
final IRabaCoder valSer, final int fromIndex,
final int toIndex, final byte[][] keys, final byte[][] vals) {
if(log.isInfoEnabled())
log.info("ntuples=" + (toIndex - fromIndex));
return new BlobsWriteProc(keySer, valSer, fromIndex, toIndex, keys,
vals, readOnly, toldBNodes);
}
} // class BlobsWriteProcConstructor
/**
* For each term whose serialized key is mapped to the current index
* partition, lookup the term in the terms index. If it is there
* then wrap its key as its {@link IV}. Otherwise, note the of terms in the
* collision bucket, and insert {hash(term),counter},term} entry into the
* terms index.
*
* @param ndx
* The terms index.
*
* @return The {@link Result}, which contains the discovered / assigned term
* identifiers.
*
* TODO There is no point sending bnodes when readOnly and NOT in
* told bnodes mode because the caller is unable to unify a blank
* node with an entry in the index.
*/
@Override
public Result applyOnce(final IIndex ndx, final IRaba keys, final IRaba vals) {
final int numTerms = keys.size();
assert numTerms > 0 : "numTerms="+numTerms;
// Helper class for index operations.
final BlobsIndexHelper helper = new BlobsIndexHelper();
// used to format the keys for the TERMS index.
final IKeyBuilder keyBuilder = helper.newKeyBuilder();
// used to store the discovered / assigned hash collision counters.
final int[] counters = new int[numTerms];
/*
* Note: The baseKey is shorter than the full key (it does not include
* the hash collision counter).
*/
final byte[] baseKey = new byte[keyBuilder.capacity()
- BlobsIndexHelper.SIZEOF_COUNTER];
// Temporary buffer used to format the [toKey] for the bucket scan.
final byte[] tmp = new byte[BlobsIndexHelper.SIZEOF_PREFIX_KEY];
// // Used to report the size of each collision bucket.
// final AtomicInteger bucketSize = new AtomicInteger(0);
// Incremented by the size of each collision bucket probed.
long totalBucketSize = 0L;
// The size of the largest collision bucket.
int maxBucketSize = 0;
final DataOutputBuffer kbuf = new DataOutputBuffer(
0/* existingDataLength */, baseKey);
for (int i = 0; i < numTerms; i++) {
// Copy key into reused buffer to reduce allocation.
// final byte[] baseKey = getKey(i);
keys.copy(i, kbuf.reset());
// decode the VTE from the flags.
final VTE vte = AbstractIV
.getVTE(KeyBuilder.decodeByte(baseKey[0]));
final int counter;
if (!toldBNodes && vte == VTE.BNODE) {
/*
* Do not enter blank nodes into the TERMS index.
*
* For this case, we just assign a term identifier and leave it
* at that. If two different documents by some chance happen to
* specify the same blank node ID they will still be assigned
* distinct term identifiers. The only way that you can get the
* same term identifier for a blank node is to have the blank
* node ID matched in a canonicalizing map of blank nodes by the
* client. That map, of course, should be scoped to the document
* in which the blank node IDs appear.
*/
if (readOnly) {
// blank nodes can not be resolved by the index.
counter = BlobsIndexHelper.NOT_FOUND;
/*
* FIXME Use this to track down people who pass in a blank
* node on a read-only request when we are not using told
* bnodes. Under these conditions we can not unify the blank
* node with the TERMS index so the node should not have
* been passed in at all (for efficiency reasons).
*/
//throw new UnsupportedOperationException();
} else {
/*
* We are not in a told bnode mode and this is not a
* read-only request. The TERMS index will be used to assign
* a unique counter to complete the blank node's key. That
* counter is just the current size of the collision bucket
* at the time that we check the index. The collision bucket
* is increased by ONE since we insert the blank node into
* the index.
*/
// The size of the collision bucket (aka the assigned ctr).
counter = helper.addBNode(ndx, keyBuilder, baseKey,
vals.get(i), tmp);
}
} else {
/*
* The serialized BigdataValue object.
*
* TODO Avoid materialization of this, preferring to operate on
* streams in the source IRaba. We will need to compare it with
* the value[] on any other tuples in the collision bucket. We
* also need to decode the byte which represents the termCode
* can can be used to derive the VTE of the Value.
*/
final byte[] val = vals.get(i);
counter = helper.resolveOrAddValue(ndx, readOnly,
keyBuilder, baseKey, val, tmp, null/* bucketSize */);
}
if (!readOnly && counter < 0)
throw new AssertionError("counter=" + counter);
counters[i] = counter;
if (counter != BlobsIndexHelper.NOT_FOUND) {
/*
* TODO This does not update the bucketSize when the Value was
* not found in the index. We could do this by changing the
* return value of resolveOrAddValue() to -rangeCount and
* casting to an (int). The (-rangeCount) could then be
* normalized to a marker as we pass the information back to the
* client. [Or just enable the bucketSize argument above.]
*/
if (maxBucketSize < counter) {
maxBucketSize += counter;
}
totalBucketSize += counter;
}
}
return new Result(totalBucketSize, maxBucketSize, counters);
} // apply(ndx)
@Override
protected void readMetadata(final ObjectInput in) throws IOException,
ClassNotFoundException {
super.readMetadata(in);
readOnly = in.readBoolean();
}
/**
* Writes metadata (not the keys or values, but just other metadata used by
* the procedure).
*
* The default implementation writes toIndex - fromIndex,
* which is the #of keys.
*
* @param out
*
* @throws IOException
*/
@Override
protected void writeMetadata(final ObjectOutput out) throws IOException {
super.writeMetadata(out);
out.writeBoolean(readOnly);
}
/**
* Object encapsulates the discovered / assigned {@link IV}s and provides
* efficient serialization for communication of those data to the client.
*
* @author Bryan
* Thompson
*/
public static class Result implements Externalizable {
/**
* The total size of the hash collision buckets examined across all
* {@link Value}s in the request. Each time a {@link Value} is resolved
* to a hash collision bucket, the size of that bucket is incremented
* against this field. Thus it will double count a bucket if the same
* bucket is visited more than once for the request.
*/
public long totalBucketSize;
/**
* The size of the largest hash collision bucket examined across all the
* {@link Value}s in the request.
*/
public int maxBucketSize;
/**
* The counters assigned to each {@link Value} in the request. The
* indices of this array are correlated with the indices of the array
* provided to the request.
*
* Note: The actual counters values are SHORTs, not INTs. However,
* {@link BlobsIndexHelper#NOT_FOUND} is an INT value used to indicate
* that the desired {@link BlobIV} was not discovered in the index. That
* means that we need to interchange and represent the counters as an
* int[].
*/
public int[] counters;
private static final long serialVersionUID = 1L;
/**
* De-serialization constructor.
*/
public Result() {
}
/**
*
* @param totalBucketSize
* The total bucket size across all buckets examined.
* @param maxBucketSize
* The size of the largest collision bucket examined.
* @param ivs
* The assigned/resolved collision counters.
*/
public Result(final long totalBucketSize, final int maxBucketSize,
final int[] counters) {
if(counters == null)
throw new IllegalArgumentException();
this.totalBucketSize = totalBucketSize;
this.maxBucketSize = maxBucketSize;
this.counters = counters;
}
private final static transient short VERSION0 = 0x0;
@Override
public void readExternal(final ObjectInput in) throws IOException,
ClassNotFoundException {
final short version = ShortPacker.unpackShort(in);
if (version != VERSION0) {
throw new IOException("Unknown version: " + version);
}
final int n = (int) LongPacker.unpackLong(in);
totalBucketSize = LongPacker.unpackLong(in);
maxBucketSize = LongPacker.unpackInt(in);
counters = new int[n];
for (int i = 0; i < n; i++) {
final short tmp = ShortPacker.unpackShort(in);
counters[i] = tmp == Short.MAX_VALUE ? BlobsIndexHelper.NOT_FOUND
: tmp;
}
}
@Override
public void writeExternal(final ObjectOutput out) throws IOException {
final int n = counters.length;
ShortPacker.packShort(out, VERSION0);
// The #of results.
LongPacker.packLong(out, n);
// The total bucket size across all buckets examined.
LongPacker.packLong(out, totalBucketSize);
// The size of the largest collision bucket examined.
LongPacker.packLong(out, maxBucketSize);
/*
* Write out the assigned/resolved collision counters.
*
* Note: This uses a packed short encoding for the collision
* counters. If we see the marker for an unresolved collision
* counter (NOT_FOUND) then it is replaced with [Short.MAX_VALUE].
* This is fine as long as the collision counter is a byte. Since
* the [short] value is in [0:Short.MAX_VALUE] we can then pack it
* into the output stream.
*/
for (int i = 0; i < n; i++) {
final int c = counters[i];
final short tmp = c == BlobsIndexHelper.NOT_FOUND ? Short.MAX_VALUE
: (short) c;
ShortPacker.packShort(out, tmp);
}
}
} // class Result
/**
* {@link Split}-wise aggregation followed by combining the results across
* those splits in order to return an aggregated result whose counters[] is
* 1:1 with the original keys[][].
*/
@Override
protected IResultHandler newAggregator() {
return new BlobResultAggregator(getKeys().size());
}
/**
* Aggregator collects the individual results in an internal ordered map and
* assembles the final result when it is requested from the individual
* results. With this approach there is no overhead or contention when the
* results are being produced in parallel and they can be combined
* efficiently within a single thread in {@link #getResult()}.
*
* @author bryan
*/
private class BlobResultAggregator extends AbstractLocalSplitResultAggregator {
/**
*
* @param size
* The #of elements in the request (which is the same as the
* cardinality of the aggregated result).
*/
public BlobResultAggregator(final int size) {
super(size);
}
@Override
protected Result newResult(final int size, SplitValuePair[] a) {
long totalBucketSize = 0;
int maxBucketSize = 0;
final int[] counters = new int[size];
for (int i = 0; i < a.length; i++) {
final Split split = a[i].key;
final Result tmp = a[i].val;
totalBucketSize += tmp.totalBucketSize;
maxBucketSize = Math.max(maxBucketSize, tmp.maxBucketSize);
System.arraycopy(tmp.counters/* src */, 0/* srcPos */, counters/* dest */, split.fromIndex/* destPos */,
split.ntuples/* length */);
}
/*
* Return the aggregated result.
*/
final Result r = new Result(totalBucketSize, maxBucketSize, counters);
return r;
}
} // BlobResultHandler
}