All Downloads are FREE. Search and download functionalities are using the official Maven repository.

com.blazegraph.gremlin.structure.BlazeGraph 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
*/
package com.blazegraph.gremlin.structure;

import static com.blazegraph.gremlin.util.Lambdas.toMap;
import static java.util.stream.Collectors.toList;

import java.util.AbstractMap.SimpleEntry;
import java.util.Arrays;
import java.util.Collection;
import java.util.Iterator;
import java.util.LinkedHashMap;
import java.util.List;
import java.util.Map;
import java.util.Optional;
import java.util.UUID;
import java.util.concurrent.atomic.AtomicLong;
import java.util.function.Function;
import java.util.stream.Collectors;
import java.util.stream.Stream;

import org.apache.commons.configuration.Configuration;
import org.apache.tinkerpop.gremlin.process.computer.GraphComputer;
import org.apache.tinkerpop.gremlin.structure.Direction;
import org.apache.tinkerpop.gremlin.structure.Edge;
import org.apache.tinkerpop.gremlin.structure.Element;
import org.apache.tinkerpop.gremlin.structure.Graph;
import org.apache.tinkerpop.gremlin.structure.Graph.Features.VertexFeatures;
import org.apache.tinkerpop.gremlin.structure.Property;
import org.apache.tinkerpop.gremlin.structure.T;
import org.apache.tinkerpop.gremlin.structure.Transaction;
import org.apache.tinkerpop.gremlin.structure.Vertex;
import org.apache.tinkerpop.gremlin.structure.VertexProperty;
import org.apache.tinkerpop.gremlin.structure.VertexProperty.Cardinality;
import org.apache.tinkerpop.gremlin.structure.util.ElementHelper;
import org.apache.tinkerpop.gremlin.structure.util.StringFactory;
import org.openrdf.model.Literal;
import org.openrdf.model.Resource;
import org.openrdf.model.Statement;
import org.openrdf.model.URI;
import org.openrdf.model.Value;
import org.openrdf.query.BindingSet;
import org.openrdf.query.Query;
import org.openrdf.query.impl.MapBindingSet;
import org.openrdf.repository.RepositoryConnection;

import com.bigdata.rdf.internal.XSD;
import com.bigdata.rdf.internal.impl.extensions.DateTimeExtension;
import com.bigdata.rdf.model.BigdataBNode;
import com.bigdata.rdf.model.BigdataStatement;
import com.bigdata.rdf.model.BigdataURI;
import com.bigdata.rdf.model.BigdataValueFactory;
import com.bigdata.rdf.sail.RDRHistory;
import com.bigdata.rdf.sail.model.RunningQuery;
import com.bigdata.rdf.store.AbstractTripleStore;
import com.blazegraph.gremlin.embedded.BasicRepositoryProvider;
import com.blazegraph.gremlin.embedded.BlazeGraphEmbedded;
import com.blazegraph.gremlin.embedded.BlazeGraphEmbedded.BlazeTransaction;
import com.blazegraph.gremlin.listener.BlazeGraphAtom;
import com.blazegraph.gremlin.listener.BlazeGraphEdit;
import com.blazegraph.gremlin.listener.BlazeGraphEdit.Action;
import com.blazegraph.gremlin.util.CloseableIterator;
import com.blazegraph.gremlin.util.Code;
import com.blazegraph.gremlin.util.LambdaLogger;
import com.blazegraph.gremlin.util.Streams;

import info.aduna.iteration.CloseableIteration;

/**
 * Blazegraph/tinkerpop3 integration.  Handles the mapping between the 
 * tinkerpop3 data model and a custom RDF/PG data model using RDF*.
 * 
 * Currently the only concrete implementation of this class is
 * {@link BlazeGraphEmbedded}, which provides an embedded (same JVM)
 * implementation of the Blazegraph/tinkerpop3 API.
 * 
 * See {@link BlazeGraphFeatures} for what tinkerpop3 features this
 * implementation supports. In addition to the tinkerpop3 features, this API
 * also provides the following:
 * 
 * 
    *
  • History API - capture full or partial history of edits to the graph.
  • *
  • Built-in full text index and search API to find graph elements.
  • *
  • Automatic SPARQL to PG translation - run a SPARQL query and get your * results back in property graph form.
  • *
  • Query management API - list and cancel running Sparql queries.
  • *
  • Bulk Load API for fast setup of new graphs.
  • *
* * And an additional two features specific to the embedded implementation: * *
    *
  • Listener API - subscribe to notifications about updates to the graph * (adds and removes of vertices/edges/properties, commits, rollbacks, etc.)
  • *
  • Support for MVCC concurrency model for high-concurrency read access.
  • *
* @author mikepersonick */ @Graph.OptIn("com.blazegraph.gremlin.structure.StructureStandardSuite") @Graph.OptIn(Graph.OptIn.SUITE_STRUCTURE_INTEGRATE) @Graph.OptIn(Graph.OptIn.SUITE_STRUCTURE_PERFORMANCE) @Graph.OptIn(Graph.OptIn.SUITE_GROOVY_ENVIRONMENT) @Graph.OptIn(Graph.OptIn.SUITE_GROOVY_ENVIRONMENT_INTEGRATE) @Graph.OptIn(Graph.OptIn.SUITE_GROOVY_ENVIRONMENT_PERFORMANCE) public abstract class BlazeGraph implements Graph { protected final transient static LambdaLogger log = LambdaLogger.getLogger(BlazeGraph.class); protected final transient static LambdaLogger sparqlLog = LambdaLogger.getLogger(BlazeGraph.class.getName() + ".SparqlLog"); /** * Options that can be specified in the graph configuration. * * @author mikepersonick */ public static interface Options { /** * The {@link BlazeValueFactory} instance this graph should use. * Defaults to {@link BlazeValueFactory#INSTANCE}. */ String VALUE_FACTORY = BlazeGraph.class.getName() + ".valueFactory"; /** * The max query time for Sparql queries before timeout. Defaults to * infinite (0). */ String MAX_QUERY_TIME = BlazeGraph.class.getName() + ".maxQueryTime"; /** * An internal option set by the concrete implementation as a floor * to use when assigning list index values for Cardinality.list * properties. Default is System.currentTimeMillis(), but better is * to use the last commit time of the database (which could be in the * future in cases of clock skew). */ String LIST_INDEX_FLOOR = BlazeGraph.class.getName() + ".listIndexFloor"; /** * Maximum number of chars to print through the SparqlLogger. */ String SPARQL_LOG_MAX = BlazeGraph.class.getName() + ".sparqlLogMax"; /** * Defaults to 10k. */ int DEFAULT_SPARQL_LOG_MAX = 10000; } /** * Enum used by the full text search API. * * @author mikepersonick */ public static enum Match { /** * Match any terms in the search string (OR). */ ANY, /** * Match all terms in the search string (AND). */ ALL, /** * Match the search string exactly using a regex filter. Most expensive * option - use ALL instead if possible. */ EXACT; } /** * Value factory for round-tripping between property graph values * (ids, labels, keys, and values) and RDF values (URIs and Literals). */ private final BlazeValueFactory vf; /** * Configuration of this graph instance. */ protected final Configuration config; /** * Sparql query string generator. */ private final SparqlGenerator sparql; /** * Counter for Cardinality.list vertex property ids, which also serve as * their list index. By starting at System.currentTimeMillis() we are * guaranteed to have monotonically increasing ids/indices * for a given vertex/key, assuming no system clock skew. * * We could use the last commit time on the underlying journal instead, * which would then make us impervious to bad system clock times. */ private final AtomicLong vpIdFactory; /** * Max Query Time used to globally set the query timeout. * * Default is 0 (unlimited) */ private final int maxQueryTime; /** * URI used for labeling elements. */ private final URI TYPE; /** * URI used for list item values. */ private final URI VALUE; /** * Datatype URI for list index for Cardinality.list vertex properties. */ private final URI LI_DATATYPE; /** * Transform functions for converting from RDF query results to property * graph results. */ protected final Transforms transforms; /** * Maximum number of chars to print through the SparqlLogger. Defaults to * 10k. */ protected final int sparqlLogMax; /** * When this is set to true, disables any implicit reads/removes that are * interleaved with the process of adding new data. This means no checking * on vertex and edge id reuse and no cleaning of old property values * (applies to all properties on Edges and VertexProperties and * Cardinality.single properties on Vertices). */ private transient volatile boolean bulkLoad = false; /** * Construct an instance using the supplied configuration. */ protected BlazeGraph(final Configuration config) { this.config = config; this.vf = Optional.ofNullable((BlazeValueFactory) config.getProperty(Options.VALUE_FACTORY)) .orElse(BlazeValueFactory.INSTANCE); final long listIndexFloor = config.getLong( Options.LIST_INDEX_FLOOR, System.currentTimeMillis()); this.vpIdFactory = new AtomicLong(listIndexFloor); this.maxQueryTime = config.getInt(Options.MAX_QUERY_TIME, 0); this.sparqlLogMax = config.getInt(Options.SPARQL_LOG_MAX, Options.DEFAULT_SPARQL_LOG_MAX); this.TYPE = vf.type(); this.VALUE = vf.value(); this.LI_DATATYPE = vf.liDatatype(); this.sparql = new SparqlGenerator(vf); this.transforms = new Transforms(); } /** * Return the factory used to round-trip between Tinkerpop values and * RDF values. */ public BlazeValueFactory valueFactory() { return vf; } /** * RDF value factory for Sesame model objects. */ public abstract BigdataValueFactory rdfValueFactory(); /** * Provide a connection to the SAIL repository for read and write * operations. */ protected abstract RepositoryConnection cxn(); /** * Returns whether the graph is in bulkLoad (true) or incremental update * (false) mode. Incremental update is the default mode * * @see #setBulkLoad(boolean) */ public boolean isBulkLoad() { return bulkLoad; } /** * When this is set to true, disables any implicit reads/removes that are * interleaved with the process of adding new data. This means no checking * on vertex and edge id reuse and no cleaning of old property values * (applies to all properties on Edges and VertexProperties and * Cardinality.single properties on Vertices). This results in considerably * greater write throughput and is suitable for loading a new data set into * an empty graph or loading data that does overlap with any data already in * an existing graph. * * Default is incremental update (bulkLoad = false). * */ public void setBulkLoad(final boolean bulkLoad) { this.bulkLoad = bulkLoad; } /** * Execute the supplied code fragment in bulk load mode and reset to * incremental mode when finished. * * @see #setBulkLoad(boolean) */ public void bulkLoad(final Code code) { if (isBulkLoad()) { Code.wrapThrow(code); } else { setBulkLoad(true); Code.wrapThrow(code, () -> setBulkLoad(false)); } } /** * Bulk load a Graph (TinkerGraph or otherwise). Uses the Graph's * features to determine VertexProperty key cardinality. Vertex and Edge * ids will be toString()-ed, VertexProperty ids will be ignored. */ public void bulkLoad(final Graph g) { final VertexFeatures vf = g.features().vertex(); bulkLoad(g, key -> vf.getCardinality(key)); } /** * Bulk load a Graph (TinkerGraph or otherwise), using the supplied * function to determine VertexProperty key cardinality. Vertex and Edge * ids will be toString()-ed, VertexProperty ids will be ignored. */ public void bulkLoad(final Graph g, final Function getCardinality) { bulkLoad(() -> { g.vertices().forEachRemaining(v -> { final String id = v.id().toString(); final BlazeVertex bv = this.addVertex(T.id, id, T.label, v.label()); v.properties().forEachRemaining(vp -> { final String key = vp.key(); final Object val = vp.value(); final BlazeVertexProperty bvp = bv.property(getCardinality.apply(key), key, val); bv.properties().forEachRemaining(p -> { bvp.property(p.key(), p.value()); }); }); }); g.edges().forEachRemaining(e -> { final String id = e.id().toString(); final String fromId = e.outVertex().id().toString(); final String toId = e.inVertex().id().toString(); final BlazeEdge be = this.addEdge(id, e.label(), fromId, toId); e.properties().forEachRemaining(p -> { be.property(p.key(), p.value()); }); }); }); } /** * Add a vertex. * * @see Graph#addVertex(Object...) */ @Override public BlazeVertex addVertex(final Object... kvs) { ElementHelper.legalPropertyKeyValueArray(kvs); final Optional suppliedId = validateSuppliedId(kvs); final String id = suppliedId.map(String.class::cast) .orElse(nextId()); final String label = ElementHelper.getLabelValue(kvs) .orElse(Vertex.DEFAULT_LABEL); ElementHelper.validateLabel(label); if (!bulkLoad) { final Optional existing = vertex(id); if (existing.isPresent()) { throw Graph.Exceptions.vertexWithIdAlreadyExists(id); } } log.info(() -> "v("+id+", "+label+")"); final BigdataValueFactory rdfvf = rdfValueFactory(); final BigdataURI uri = rdfvf.asValue(vf.elementURI(id)); final BigdataURI rdfLabel = rdfvf.asValue(vf.typeURI(label)); final RepositoryConnection cxn = cxn(); Code.wrapThrow(() -> { cxn.add(uri, TYPE, rdfLabel); }); final BlazeVertex vertex = new BlazeVertex(this, uri, rdfLabel); ElementHelper.attachProperties(vertex, VertexProperty.Cardinality.set, kvs); return vertex; } /** * Add an edge. Helper for {@link BlazeVertex#addEdge(String, Vertex, Object...)} * to consolidate these operations in one place. * * @see Vertex#addEdge(String, Vertex, Object...) */ public BlazeEdge addEdge(final BlazeVertex from, final BlazeVertex to, final String label, final Object... kvs) { ElementHelper.validateLabel(label); ElementHelper.legalPropertyKeyValueArray(kvs); final Optional suppliedId = validateSuppliedId(kvs); final String id = suppliedId.map(String.class::cast) .orElse(nextId()); if (!bulkLoad) { final Optional existing = edge(id); if (existing.isPresent()) { throw Graph.Exceptions.vertexWithIdAlreadyExists(id); } } log.info(() -> "v("+from+")-e("+id+", "+label+")->v("+to+")"); final BigdataValueFactory rdfvf = rdfValueFactory(); final URI uri = rdfvf.asValue(vf.elementURI(id)); final BigdataURI rdfLabel = rdfvf.asValue(vf.typeURI(label)); final BigdataStatement edgeStmt = rdfvf.createStatement(from.rdfId(), uri, to.rdfId()); final RepositoryConnection cxn = cxn(); Code.wrapThrow(() -> { // blaze:person:1 blaze:knows:7 blaze:person:2 . cxn.add(edgeStmt); // <> rdf:type "knows" . cxn.add(rdfvf.createBNode(edgeStmt), TYPE, rdfLabel); }); final BlazeEdge edge = new BlazeEdge(this, edgeStmt, rdfLabel, from, to); ElementHelper.attachProperties(edge, kvs); return edge; } /** * Private helper used exclusively by {@link #bulkLoad(Graph)}. */ private BlazeEdge addEdge(final String id, final String label, final String fromId, final String toId) { ElementHelper.validateLabel(label); if (!bulkLoad) { final Optional existing = edge(id); if (existing.isPresent()) { throw Graph.Exceptions.vertexWithIdAlreadyExists(id); } } final BigdataValueFactory rdfvf = rdfValueFactory(); final BigdataURI uri = rdfvf.asValue(vf.elementURI(id)); final BigdataURI rdfLabel = rdfvf.asValue(vf.typeURI(label)); final BigdataURI fromURI = rdfvf.asValue(vf.elementURI(fromId)); final BigdataURI toURI = rdfvf.asValue(vf.elementURI(toId)); final BigdataStatement edgeStmt = rdfvf.createStatement(fromURI, uri, toURI); final RepositoryConnection cxn = cxn(); Code.wrapThrow(() -> { // blaze:person:1 blaze:knows:7 blaze:person:2 . cxn.add(edgeStmt); // <> rdf:type "knows" . cxn.add(rdfvf.createBNode(edgeStmt), TYPE, rdfLabel); }); final BlazeEdge edge = new BlazeEdge(this, edgeStmt, rdfLabel); return edge; } /** * Generate an element id (vertex or edge) if not supplied via T.id. */ private final String nextId() { final String id = UUID.randomUUID().toString(); return id;//.substring(id.length()-5); } /** * Helper for {@link BlazeEdge#property(String, Object)} and * {@link BlazeVertexProperty#property(String, Object)}. * * @param element the BlazeEdge or BlazeVertexProperty * @param key the property key * @param val the property value * @return a BlazeProperty */ BlazeProperty property(final BlazeReifiedElement element, final String key, final V val) { final BigdataValueFactory rdfvf = rdfValueFactory(); final BigdataBNode s = element.rdfId(); final URI p = rdfvf.asValue(vf.propertyURI(key)); final Literal lit = rdfvf.asValue(vf.toLiteral(val)); final RepositoryConnection cxn = cxn(); Code.wrapThrow(() -> { final BigdataStatement stmt = rdfvf.createStatement(s, p, lit); if (!bulkLoad) { // remove (<> ?) cxn.remove(s, p, null); } // add (<> "val") cxn.add(stmt); }); final BlazeProperty prop = new BlazeProperty(this, element, p, lit); return prop; } /** * Helper for {@link BlazeVertex#property(Cardinality, String, Object, Object...)}. * * @param vertex the BlazeVertex * @param cardinality the property cardinality * @param key the property key * @param val the property value * @param kvs the properties to attach to the BlazeVertexProperty * @return a BlazeVertexProperty */ BlazeVertexProperty vertexProperty(final BlazeVertex vertex, final Cardinality cardinality, final String key, final V val, final Object... kvs) { final BigdataValueFactory rdfvf = rdfValueFactory(); final URI s = vertex.rdfId(); final URI p = rdfvf.asValue(vf.propertyURI(key)); final Literal lit = rdfvf.asValue(vf.toLiteral(val)); final BigdataStatement stmt; if (cardinality == Cardinality.list) { final Literal timestamp = rdfvf.createLiteral( String.valueOf(vpIdFactory.getAndIncrement()), LI_DATATYPE); stmt = rdfvf.createStatement(s, p, timestamp); } else { stmt = rdfvf.createStatement(s, p, lit); } final BigdataBNode sid = rdfvf.createBNode(stmt); final String vpId = vertexPropertyId(stmt); final RepositoryConnection cxn = cxn(); Code.wrapThrow(() -> { if (cardinality == Cardinality.list) { // timestamp . cxn.add(stmt); // << timestamp>> rdf:value "val" . cxn.add(sid, VALUE, lit); } else { if (cardinality == Cardinality.single && !bulkLoad) { final String queryStr = sparql.cleanVertexProps(s, p, lit); update(queryStr); } // << stmt >> "val" . cxn.add(stmt); } }); final BlazeProperty prop = new BlazeProperty<>(this, vertex, p, lit); final BlazeVertexProperty bvp = new BlazeVertexProperty<>(prop, vpId, sid); ElementHelper.attachProperties(bvp, kvs); return bvp; } /** * Template the BlazeVertexProperty ids. */ private final String VERTEX_PROPERTY_ID = "<<%s, %s, %s>>"; /** * Construct a BlazeVertexProperty id. */ private String vertexPropertyId(final URI s, final URI p, final Literal o) { return String.format(VERTEX_PROPERTY_ID, s.toString(), p.toString(), o.toString()); } /** * Construct a BlazeVertexProperty id. */ private String vertexPropertyId(final Statement stmt) { return vertexPropertyId((URI) stmt.getSubject(), (URI) stmt.getPredicate(), (Literal) stmt.getObject()); } /** * Validate user supplied ids for vertices and edges. */ private final Optional validateSuppliedId(final Object... kvs) { final Optional suppliedId = ElementHelper.getIdValue(kvs); if (suppliedId.isPresent() && !(suppliedId.get() instanceof String)) { throw Vertex.Exceptions.userSuppliedIdsOfThisTypeNotSupported(); } return suppliedId; } /** * Remove a vertex, cleaning any attached edges and vertex properties (and * their properties). * * @see Vertex#remove() */ void remove(final BlazeVertex vertex) { final String queryStr = sparql.removeVertex(vertex); update(queryStr); } /** * Remove an edge or a vertex property (and its properties). * * @see Edge#remove() * @see VertexProperty#remove() */ void remove(final BlazeReifiedElement element) { final String queryStr = sparql.removeReifiedElement(element); update(queryStr); } /** * Remove a property. * * @see Property#remove() */ void remove(final BlazeProperty prop) { final RepositoryConnection cxn = cxn(); Code.wrapThrow(() -> { cxn.remove(prop.element().rdfId(), prop.rdfKey(), prop.rdfValue()); }); } /** * Fast vertex count using Sparql aggregation. * * @return vertex count */ public int vertexCount() { final String queryStr = sparql.vertexCount(); return count(queryStr); } /** * Lookup a vertex by id. * * @param vertexId the id * @return the BlazeVertex, if one exists for the supplied id */ public Optional vertex(final String vertexId) { try (final CloseableIterator it = vertices(vertexId)) { final Optional v = it.hasNext() ? Optional.of((BlazeVertex) it.next()) : Optional.empty(); if (it.hasNext()) { throw new IllegalStateException("Multiple vertices found with id: " + vertexId); } return v; } } /** * Lookup vertices by (optional) ids. Return type strengthened from normal * iterator. You MUST close this iterator when finished. */ @Override public CloseableIterator vertices(final Object... vertexIds) { final List uris = validateIds(vertexIds); final String queryStr = sparql.vertices(uris); final Stream stream = _select(queryStr, nextQueryId()) .map(transforms.vertex); return CloseableIterator.of(stream); } /** * Fast edge count using Sparql aggregation. * * @return edge count */ public int edgeCount() { final String queryStr = sparql.edgeCount(); return count(queryStr); } /** * Lookup an edge by id. * * @param edgeId the id * @return the BlazeEdge, if one exists for the supplied id */ public Optional edge(final Object edgeId) { try (CloseableIterator it = edges(edgeId)) { final Optional e = it.hasNext() ? Optional.of((BlazeEdge) it.next()) : Optional.empty(); if (it.hasNext()) { throw new IllegalStateException("Multiple edges found with id: " + edgeId); } return e; } } /** * Lookup edges by (optional) ids. Return type strengthened from normal * iterator. You MUST close this iterator when finished. */ @Override public CloseableIterator edges(final Object... edgeIds) { final List uris = validateIds(edgeIds); final String queryStr = sparql.edges(uris); final Stream stream = _select(queryStr, nextQueryId()) .map(transforms.edge); return CloseableIterator.of(stream); } /** * Lookup edges from a source vertex. * * @see Vertex#edges(Direction, String...) */ CloseableIterator edgesFromVertex(final BlazeVertex src, final Direction dir, final String... edgeLabels) { final List uris = Stream.of(edgeLabels).map(vf::typeURI).collect(toList()); final String queryStr = sparql.edgesFromVertex(src, dir, uris); final Stream stream = _select(queryStr, nextQueryId()) .map(transforms.edge); return CloseableIterator.of(stream); } /** * Iterate properties on a BlazeEdge or BlazeVertexProperty. Return type * strengthened from normal iterator. You MUST close this iterator when * finished. * * @see Edge#properties(String...) * @see VertexProperty#properties(String...) */ CloseableIterator> properties( final BlazeReifiedElement element, final String... keys) { final List uris = Stream.of(keys).map(vf::propertyURI).collect(toList()); final String queryStr = sparql.properties(element, uris); final Stream> stream = _select(queryStr, nextQueryId()) .map(transforms.property(element)); return CloseableIterator.of(stream); } /** * Iterate properties on a BlazeVertex. Return type * strengthened from normal iterator. You MUST close this iterator when * finished. * * @see Vertex#properties(String...) */ CloseableIterator> properties( final BlazeVertex vertex, final String... keys) { final List uris = Stream.of(keys).map(vf::propertyURI).collect(toList()); final String queryStr = sparql.vertexProperties(vertex, uris); final Stream> stream = _select(queryStr, nextQueryId()) .map(transforms.vertexProperty(vertex)); return CloseableIterator.of(stream); } /** * Run a Sparql count aggregation and parse the count from the result set. */ private int count(final String queryStr) { try (Stream stream = _select(queryStr, nextQueryId())) { return stream.map(bs -> (Literal) bs.getValue("count")) .map(Literal::intValue) .findFirst().get(); } } /** * Helper to parse URIs from a list of element ids for lookup of * vertices or edges. * * @see #vertices(Object...) * @see #edges(Object...) */ private List validateIds(final Object... elementIds) { ElementHelper.validateMixedElementIds(Element.class, elementIds); return Stream.of(elementIds) .map(elementId -> { if (elementId instanceof String) { return (String) elementId; } else if (elementId instanceof Element) { final Object id = ((Element) elementId).id(); if (id instanceof String) { return (String) id; } } throw new IllegalArgumentException( "Unknown element id type: " + elementId + " ("+ elementId.getClass()+")"); }) .map(vf::elementURI) .collect(Collectors.toList()); } /** * Return the transaction object for this graph. Implementation-specific. * * @see BlazeTransaction */ @Override public abstract Transaction tx(); /** * Close the graph. Implementation-specific. * * @see BlazeGraphEmbedded#close */ @Override public abstract void close(); /** * Return the graph configuration. */ @Override public Configuration configuration() { return config; } /** * Return the features. * * @see BlazeGraphFeatures */ @Override public BlazeGraphFeatures features() { return BlazeGraphFeatures.INSTANCE; } /** * Standard graph toString() representation using TP3 StringFactory. */ @Override public String toString() { return StringFactory.graphString(this, "vertices:" + vertexCount() + " edges:" + edgeCount()); } /** * Variables not currently supported. * * TODO FIXME */ @Override public Variables variables() { throw Graph.Exceptions.variablesNotSupported(); } /** * GraphComputer not currently supported. * * TODO FIXME Implement GraphComputer over DASL API */ @Override public GraphComputer compute() throws IllegalArgumentException { throw Graph.Exceptions.graphComputerNotSupported(); } /** * GraphComputer not currently supported. * * TODO FIXME Implement GraphComputer over DASL API */ @Override public C compute(Class graphComputerClass) throws IllegalArgumentException { throw Graph.Exceptions.graphComputerNotSupported(); } /** * Generate a query id for internal Sparql queries. */ private String nextQueryId() { return UUID.randomUUID().toString(); } /** * Project a subgraph using a SPARQL query. *

* Warning: You MUST close this iterator when finished. */ public CloseableIterator project(final String queryStr) throws Exception { return this.project(queryStr, nextQueryId()); } /** * Project a subgraph using a SPARQL query. * * This version allows passing an external system ID to allow association * between queries in the query engine when using an Embedded Client. * *

* Warning: You MUST close this iterator when finished. */ public CloseableIterator project(final String queryStr, String externalQueryId) throws Exception { final Stream stream = _project(queryStr, externalQueryId) .map(transforms.graphAtom) .filter(Optional::isPresent) .map(Optional::get); return CloseableIterator.of(stream); } /** * Select results using a SPARQL query. *

* Warning: You MUST close this iterator when finished. */ public CloseableIterator select(final String queryStr) { return this.select(queryStr, nextQueryId()); } /** * Select results using a SPARQL query. *

* Warning: You MUST close this iterator when finished. */ public CloseableIterator select( final String queryStr, final String extQueryId) { final Stream stream = _select(queryStr, extQueryId) .map(transforms.bindingSet); return CloseableIterator.of(stream); } /** * Select results using a SPARQL query. */ public boolean ask(final String queryStr) { return ask(queryStr, nextQueryId()); } /** * Select results using a SPARQL query. */ public boolean ask(final String queryStr, final String extQueryId) { return _ask(queryStr, extQueryId); } /** * Update graph using SPARQL Update. */ public void update(final String queryStr) { update(queryStr, nextQueryId()); } /** * Update graph using SPARQL Update. */ public void update(final String queryStr, final String extQueryId) { _update(queryStr, extQueryId); } /** * If history is enabled, return an iterator of historical graph edits * related to any of the supplied ids. To enable history, make sure * that the RDR History class is enabled (it is by default with * {@link BasicRepositoryProvider}. Only vertex and edges ids currently * supported. *

* Warning: You MUST close this iterator when finished. * * @see AbstractTripleStore.Options#STATEMENT_IDENTIFIERS * @see AbstractTripleStore.Options#RDR_HISTORY_CLASS * @see RDRHistory */ public CloseableIterator history(final String... ids) { return history(Arrays.asList(ids)); } /** * @see #history(String...) */ public CloseableIterator history(final List ids) { validateHistoryIds(ids); final List uris = ids.stream().map(vf::elementURI).collect(toList()); final String queryStr = sparql.history(uris); final Stream stream = _select(queryStr, nextQueryId()) .map(transforms.history) .filter(Optional::isPresent) .map(Optional::get); return CloseableIterator.of(stream); } /** * Make sure ids only contains edge and vertex ids (no vertex properties). */ private void validateHistoryIds(final List ids) { ids.forEach(id -> { if (id.startsWith("<<") && id.endsWith(">>")) { throw new IllegalArgumentException( "History not (yet) supported for VertexProperty elements: " + id); } }); } /** * Search for properties in the graph using the built-in full text index. * * @param search * The search string. Tokens in the search string will be matched * according to the supplied match parameter. * @param match * The match behavior. * @return * Any properties whose value matches the search. */ public CloseableIterator> search(final String search, final Match match) { final String queryStr = sparql.search(search, match); final Stream> stream = _select(queryStr, nextQueryId()) .map(transforms.search()); return CloseableIterator.of(stream); } /** * Convert a Sesame iteration into a Java 8 Stream. Backed by database * resources so these streams must be closed when finished. (Does not * close the connection but will close the Sesame iteration / Blaze * iterators). * * @param * the type of Sesame element (BindingSet or Statement) * * @author mikepersonick */ protected class GraphStreamer { /** * Sesame iteration. */ private final CloseableIteration result; /** * Custom onClose() behavior. */ private final Optional onClose; /** * Iterator wrapping the Sesame iteration. */ private final Iterator it; /** * Construct a streamer using the supplied Sesame iteration and optional * onClose() behavior. The iteration will be closed automatically by * this class, so the onClose code does not need to do it. */ public GraphStreamer(final CloseableIteration result, final Optional onClose) { this.result = result; this.onClose = onClose; /* * Wrap the iteration as an iterator and ultimately wrap the * iterator as a stream with the appropriate onClose() behavior. */ this.it = new Iterator() { @Override public boolean hasNext() { return Code.wrapThrow(() -> { return result.hasNext(); }); } @Override public E next() { return Code.wrapThrow( () -> { /* try */ return result.next(); }, () -> { /* finally */ if (!hasNext()) { close(); } } ); } }; } private boolean streamed = false; /** * Provide a one-time stream against the Sesame iteration with the * appropriate onClose behavior. You MUST close this stream to * release database resources. */ public Stream stream() { if (streamed) throw Exceptions.alreadyStreamed(); final Stream s = Streams.of(it).onClose(() -> close()); streamed = true; return s; } private boolean closed = false; /** * Close the Sesame iteration and run the custom onClose * behavior if supplied. */ public void close() { if (closed) return; Code.wrapThrow( () -> { /* try */ result.close(); onClose.ifPresent(Runnable::run); }, () -> { /* finally */ closed = true; } ); } } /** * Internal Sparql select method. Prepare a Sparql select query, turn * Sesame iteration into a stream of binding sets using {@link GraphStreamer}. * Callers MUST close this stream when finished. */ protected abstract Stream _select( final String queryStr, final String extQueryId); /** * Internal Sparql project method. Prepare a Sparql graph query, turn * Sesame iteration into a stream of statements using {@link GraphStreamer}. * Callers MUST close this stream when finished. */ protected abstract Stream _project( final String queryStr, final String extQueryId); /** * Internal Sparql ask method. Prepare a Sparql ask query and return result. */ protected abstract boolean _ask( final String queryStr, final String extQueryId); /** * Internal Sparql update method. Prepare a Sparql Update and execute. */ protected abstract void _update( final String queryStr, final String extQueryId); /** * Utility function to set the Query timeout to the global * setting if it is configured. */ protected void setMaxQueryTime(final Query query) { if (maxQueryTime > 0) { query.setMaxQueryTime(maxQueryTime); } } /** * Return a Collection of running queries */ public abstract Collection getRunningQueries(); /** * Kill a running query specified by the UUID. Do nothing if the query has * completed. */ public abstract void cancel(UUID queryId); /** * Kill a running query specified by the RunningQuery object. Do nothing if * the query has completed. */ public abstract void cancel(RunningQuery r); protected static class Exceptions { public static IllegalStateException alreadyClosed() { return new IllegalStateException("Graph has already been closed."); } public static IllegalStateException alreadyStreamed() { return new IllegalStateException("Sesame iteration has already been streamed."); } } /** * So that BlazeGraphEmbedded can use it for listeners. */ protected Function> graphAtomTransform() { return transforms.graphAtom; } /** * Transform functions from RDF results (binding sets and statements) to * property graph results. * * @author mikepersonick */ protected class Transforms { /** * Binding set to vertex. */ private final Function vertex = bs -> { final BigdataURI uri = (BigdataURI) bs.getValue("vertex"); // final Literal label = (Literal) bs.getValue("label"); final BigdataURI label = (BigdataURI) bs.getValue("label"); final BlazeVertex vertex = new BlazeVertex(BlazeGraph.this, uri, label); return vertex; }; /** * Binding set to edge. */ private final Function edge = bs -> { log.debug(() -> bs); final BigdataBNode s = (BigdataBNode) bs.getValue("edge"); final BigdataStatement stmt = s.getStatement(); // final Literal label = (Literal) bs.getValue("label"); final BigdataURI label = (BigdataURI) bs.getValue("label"); final BigdataURI fromURI = (BigdataURI) stmt.getSubject(); // final Literal fromLabel = (Literal) bs.getValue("fromLabel"); final BigdataURI fromLabel = (BigdataURI) bs.getValue("fromLabel");; final BlazeVertex from = new BlazeVertex(BlazeGraph.this, fromURI, fromLabel); final BigdataURI toURI = (BigdataURI) stmt.getObject(); // final Literal toLabel = (Literal) bs.getValue("toLabel"); final BigdataURI toLabel = (BigdataURI) bs.getValue("toLabel");; final BlazeVertex to = new BlazeVertex(BlazeGraph.this, toURI, toLabel); return new BlazeEdge(BlazeGraph.this, stmt, label, from, to); }; /** * Binding set to property (for Edge and VertexProperty elements). */ private final Function> property(final BlazeReifiedElement e) { return bs -> { log.debug(() -> bs); final URI key = (URI) bs.getValue("key"); final Literal val = (Literal) bs.getValue("val"); final BlazeProperty prop = new BlazeProperty<>(BlazeGraph.this, e, key, val); return prop; }; } /** * Binding set to vertex property (for Vertex elements). */ private final Function> vertexProperty(final BlazeVertex v) { return bs -> { log.debug(() -> bs); final Literal val = (Literal) bs.getValue("val"); final BigdataBNode sid = (BigdataBNode) bs.getValue("vp"); final BigdataStatement stmt = sid.getStatement(); final URI key = stmt.getPredicate(); final String vpId = vertexPropertyId(stmt); final BlazeProperty prop = new BlazeProperty<>(BlazeGraph.this, v, key, val); final BlazeVertexProperty bvp = new BlazeVertexProperty<>(prop, vpId, sid); return bvp; }; } /** * RDF binding set to PG binding set. */ private final Function bindingSet = bs -> { log.debug(() -> bs); final Map map = bs.getBindingNames().stream() .map(key -> new SimpleEntry(key, bs.getBinding(key).getValue())) .map(e -> { final String key = e.getKey(); final Object val = e.getValue(); final SimpleEntry _e; if (val instanceof Literal) { _e = new SimpleEntry<>(key, vf.fromLiteral((Literal) val)); } else if (val instanceof URI) { _e = new SimpleEntry<>(key, vf.fromURI((URI) val)); } else { final BigdataBNode sid = (BigdataBNode) val; final BigdataStatement stmt = sid.getStatement(); if (stmt.getObject() instanceof URI) { // return edge id _e = new SimpleEntry<>(key, vf.fromURI(stmt.getPredicate())); } else { // return vertex property id _e = new SimpleEntry<>(key, vertexPropertyId(stmt)); } } return _e; }) .collect(toMap(LinkedHashMap::new)); return new BlazeBindingSet(map); }; /** * Atomic unit of RDF data (statement) to atomic unit of PG data * (graph atom). */ private final Function> graphAtom = stmt -> { final Resource s = stmt.getSubject(); final URI p = stmt.getPredicate(); final Value o = stmt.getObject(); if (s instanceof URI) { if (o instanceof URI && !TYPE.equals(p)) { // blaze:a blaze:x blaze:b /* * We actually want to ignore these and wait for * <> rdf:type blaze:label . * to emit a VertexAtom. */ return Optional.empty(); } final BigdataURI uri = (BigdataURI) s; final String vertexId = vf.fromURI(uri); if (TYPE.equals(p)) { // blaze:a rdf:type "label" . final String label = vf.fromURI((URI) o); return Optional.of( new BlazeGraphAtom.VertexAtom(vertexId, label)); } else { final Literal lit = (Literal) o; final URI dt = lit.getDatatype(); if (dt != null && LI_DATATYPE.equals(dt)) { // blaze:a blaze:key "0"^^"bg:listIndex" /* * We actually want to ignore these and wait for * <> rdfs:value "val" . * to emit a VertexPropertyAtom. */ return Optional.empty(); } else { // blaze:a blaze:key "val" . final String key = vf.fromURI(p); final Object val = vf.fromLiteral(lit); final String vpId = vertexPropertyId(uri, p, lit); return Optional.of( new BlazeGraphAtom.VertexPropertyAtom(vertexId, key, val, vpId)); } } } else { // s instanceof BigdataBNode final BigdataBNode sid = (BigdataBNode) s; final BigdataStatement reified = sid.getStatement(); if (TYPE.equals(p)) { // <> rdf:type "label" . final String edgeId = vf.fromURI(reified.getPredicate()); final String fromId = vf.fromURI((URI) reified.getSubject()); final String toId = vf.fromURI((URI) reified.getObject()); final String label = vf.fromURI((URI) o); return Optional.of( new BlazeGraphAtom.EdgeAtom(edgeId, label, fromId, toId)); } else if (VALUE.equals(p)) { // <> rdfs:value "val" . final String vertexId = vf.fromURI((URI) reified.getSubject()); final String key = vf.fromURI(reified.getPredicate()); final Object val = vf.fromLiteral((Literal) o); final String vpId = vertexPropertyId(reified); return Optional.of( new BlazeGraphAtom.VertexPropertyAtom(vertexId, key, val, vpId)); } else { final String key = vf.fromURI(p); final Object val = vf.fromLiteral((Literal) o); if (reified.getObject() instanceof URI) { // <> blaze:key "val" . final String edgeId = vf.fromURI(reified.getPredicate()); return Optional.of( new BlazeGraphAtom.PropertyAtom(edgeId, key, val)); } else { // <> blaze:key "val" . // <> blaze:key "val" . final String vpId = vertexPropertyId(reified); return Optional.of( new BlazeGraphAtom.PropertyAtom(vpId, key, val)); } } } }; /** * History query result to graph edit. */ private final Function> history = bs -> { log.debug(() -> bs); final BigdataBNode sid = (BigdataBNode) bs.getValue("sid"); final BigdataStatement stmt = sid.getStatement(); final URI a = (URI) bs.getValue("action"); final Literal t = (Literal) bs.getValue("time"); if (!t.getDatatype().equals(XSD.DATETIME)) { throw new RuntimeException("Unexpected timestamp in result: " + bs); } final BlazeGraphEdit.Action action; if (a.equals(RDRHistory.Vocab.ADDED)) { action = Action.Add; } else if (a.equals(RDRHistory.Vocab.REMOVED)) { action = Action.Remove; } else { throw new RuntimeException("Unexpected action in result: " + bs); } final long timestamp = DateTimeExtension.getTimestamp(t.getLabel()); return graphAtom.apply(stmt).map(atom -> new BlazeGraphEdit(action, atom, timestamp)); }; /** * Search query result to property. */ private final Function> search() { return bs -> { log.debug(() -> bs); final Property prop; if (bs.hasBinding("edge")) { /* * Edge property */ final BlazeEdge edge = (BlazeEdge) this.edge.apply(bs); prop = this.property(edge).apply(bs); } else { final BlazeVertex vertex = (BlazeVertex) this.vertex.apply(bs); if (bs.hasBinding("vpVal")) { /* * VertexProperty property */ final MapBindingSet remap = new MapBindingSet() {{ addBinding("vp", bs.getValue("vp")); addBinding("val", bs.getValue("vpVal")); }}; final BlazeVertexProperty vp = (BlazeVertexProperty) this.vertexProperty(vertex).apply(remap); prop = this.property(vp).apply(bs); } else { /* * Vertex property */ prop = this.vertexProperty(vertex).apply(bs); } } return prop; }; } } }