/*
* 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 com.hp.hpl.jena.reasoner.rulesys;
import java.util.Iterator;
import java.util.List;
import com.hp.hpl.jena.graph.Factory;
import com.hp.hpl.jena.graph.Graph;
import com.hp.hpl.jena.graph.Node;
import com.hp.hpl.jena.graph.Triple;
import com.hp.hpl.jena.reasoner.BaseInfGraph;
import com.hp.hpl.jena.reasoner.Derivation;
import com.hp.hpl.jena.reasoner.FGraph;
import com.hp.hpl.jena.reasoner.Finder;
import com.hp.hpl.jena.reasoner.FinderUtil;
import com.hp.hpl.jena.reasoner.Reasoner;
import com.hp.hpl.jena.reasoner.ReasonerException;
import com.hp.hpl.jena.reasoner.TriplePattern;
import com.hp.hpl.jena.reasoner.rulesys.impl.FRuleEngineI;
import com.hp.hpl.jena.reasoner.rulesys.impl.FRuleEngineIFactory;
import com.hp.hpl.jena.reasoner.rulesys.impl.SafeGraph;
import com.hp.hpl.jena.util.OneToManyMap;
import com.hp.hpl.jena.util.iterator.ExtendedIterator;
import com.hp.hpl.jena.util.iterator.NullIterator;
/**
* An inference graph interface that runs a set of forward chaining
* rules to conclusion on each added triple and stores the entire
* result set.
* <p>
* This implementation has a horribly inefficient rule chainer built in.
* Once we have this working generalize this to an interface than
* can call out to a rule engine and build a real rule engine (e.g. Rete style). </p>
*/
public class BasicForwardRuleInfGraph extends BaseInfGraph implements ForwardRuleInfGraphI {
/** Table of derivation records, maps from triple to RuleDerivation */
protected OneToManyMap<Triple, Derivation> derivations;
/** The set of deduced triples, this is in addition to base triples in the fdata graph */
protected FGraph fdeductions;
/** A safe wrapped version of the deductions graph used for reporting getDeductions */
protected Graph safeDeductions;
/** Reference to any schema graph data bound into the parent reasoner */
protected Graph schemaGraph;
/** The forward rule engine being used */
protected FRuleEngineI engine;
/** The original rule set as supplied */
private List<Rule> rules;
/** Flag, if true then find results will be filtered to remove functors and illegal RDF */
public boolean filterFunctors = true;
/** Flag which, if true, enables tracing of rule actions to logger.info */
protected boolean traceOn = false;
// private static Logger logger = LoggerFactory.getLogger(BasicForwardRuleInfGraph.class);
//=======================================================================
// Core methods
/**
* Constructor. Creates a new inference graph to which a (compiled) rule set
* and a data graph can be attached. This separation of binding is useful to allow
* any configuration parameters (such as logging) to be set before the data is added.
* Note that until the data is added using {@link #rebind rebind} then any operations
* like add, remove, find will result in errors.
*
* @param reasoner the parent reasoner
* @param schema the (optional) schema data which is being processed
*/
public BasicForwardRuleInfGraph(Reasoner reasoner, Graph schema) {
super(null, reasoner);
instantiateRuleEngine(null);
this.schemaGraph = schema;
}
/**
* Constructor. Creates a new inference graph based on the given rule set.
* No data graph is attached at this stage. This is to allow
* any configuration parameters (such as logging) to be set before the data is added.
* Note that until the data is added using {@link #rebind rebind} then any operations
* like add, remove, find will result in errors.
*
* @param reasoner the parent reasoner
* @param rules the list of rules to use this time
* @param schema the (optional) schema or preload data which is being processed
*/
public BasicForwardRuleInfGraph(Reasoner reasoner, List<Rule> rules, Graph schema)
{
super( null, reasoner );
instantiateRuleEngine( rules );
this.rules = rules;
this.schemaGraph = schema;
}
/**
* Constructor. Creates a new inference graph based on the given rule set
* then processes the initial data graph. No precomputed deductions are loaded.
*
* @param reasoner the parent reasoner
* @param rules the list of rules to use this time
* @param schema the (optional) schema or preload data which is being processed
* @param data the data graph to be processed
*/
public BasicForwardRuleInfGraph(Reasoner reasoner, List<Rule> rules, Graph schema, Graph data) {
this(reasoner, rules, schema);
rebind(data);
}
/**
* Instantiate the forward rule engine to use.
* Subclasses can override this to switch to, say, a RETE imlementation.
* @param rules the rule set or null if there are not rules bound in yet.
*/
protected void instantiateRuleEngine(List<Rule> rules) {
engine = FRuleEngineIFactory.getInstance().createFRuleEngineI(this, rules, false);
}
/**
* Attach a compiled rule set to this inference graph.
* @param ruleStore a compiled set of rules (i.e. the result of an FRuleEngine.compile).
*/
public void setRuleStore(Object ruleStore) {
engine.setRuleStore(ruleStore);
}
/**
* Replace the underlying data graph for this inference graph and start any
* inferences over again. This is primarily using in setting up ontology imports
* processing to allow an imports multiunion graph to be inserted between the
* inference graph and the raw data, before processing.
* @param data the new raw data graph
*/
@Override
public void rebind(Graph data) {
fdata = new FGraph( data );
rebind();
}
/**
* Cause the inference graph to reconsult the underlying graph to take
* into account changes. Normally changes are made through the InfGraph's add and
* remove calls are will be handled appropriately. However, in some cases changes
* are made "behind the InfGraph's back" and this forces a full reconsult of
* the changed data.
*/
@Override
public void rebind() {
version++;
this.setPreparedState(false);
}
/**
* Return the schema graph, if any, bound into this inference graph.
*/
@Override
public Graph getSchemaGraph() {
return schemaGraph;
}
/**
* Perform any initial processing and caching. This call is optional. Most
* engines either have negligable set up work or will perform an implicit
* "prepare" if necessary. The call is provided for those occasions where
* substantial preparation work is possible (e.g. running a forward chaining
* rule system) and where an application might wish greater control over when
* this prepration is done.
*/
@Override
public synchronized void prepare() {
if (this.isPrepared()) return;
this.setPreparedState(true);
// initilize the deductions graph
fdeductions = new FGraph( createDeductionsGraph() );
boolean rulesLoaded = false;
if (schemaGraph != null) {
rulesLoaded = preloadDeductions(schemaGraph);
}
if (rulesLoaded) {
engine.fastInit(fdata);
} else {
engine.init(true, fdata);
}
}
/**
* Adds a set of precomputed triples to the deductions store. These do not, themselves,
* fire any rules but provide additional axioms that might enable future rule
* firing when real data is added. Used to implement bindSchema processing
* in the parent Reasoner.
* @return return true if the rule set has also been loaded
*/
protected boolean preloadDeductions(Graph preloadIn) {
Graph d = fdeductions.getGraph();
BasicForwardRuleInfGraph preload = (BasicForwardRuleInfGraph) preloadIn;
// If the rule set is the same we can reuse those as well
if (preload.rules == rules) {
// Load raw deductions
for (Iterator<Triple> i = preload.find(null, null, null); i.hasNext(); ) {
d.add(i.next());
}
engine.setRuleStore(preload.engine.getRuleStore());
return true;
} else {
return false;
}
}
/**
* Add a new deduction to the deductions graph.
*/
@Override
public void addDeduction(Triple t) {
getDeductionsGraph().add(t);
}
/**
* Set to true to cause functor-valued literals to be dropped from rule output.
* Default is true.
*/
@Override
public void setFunctorFiltering(boolean param) {
filterFunctors = param;
}
/**
* Extended find interface used in situations where the implementator
* may or may not be able to answer the complete query. It will
* attempt to answer the pattern but if its answers are not known
* to be complete then it will also pass the request on to the nested
* Finder to append more results.
* @param pattern a TriplePattern to be matched against the data
* @param continuation either a Finder or a normal Graph which
* will be asked for additional match results if the implementor
* may not have completely satisfied the query.
*/
@Override
public ExtendedIterator<Triple> findWithContinuation(TriplePattern pattern, Finder continuation) {
return findWithContinuation(pattern, continuation, true);
}
/**
* Internals of findWithContinuation implementation which allows control
* over functor filtering.
*/
private ExtendedIterator<Triple> findWithContinuation(TriplePattern pattern, Finder continuation, boolean filter) {
checkOpen();
this.requirePrepared();
ExtendedIterator<Triple> result = null;
if (fdata == null) {
result = fdeductions.findWithContinuation(pattern, continuation);
} else {
if (continuation == null) {
result = fdata.findWithContinuation(pattern, fdeductions);
} else {
result = fdata.findWithContinuation(pattern, FinderUtil.cascade(fdeductions, continuation) );
}
}
if (filter && filterFunctors) {
return result.filterDrop(Functor.acceptFilter);
} else {
return result;
}
}
/**
* Returns an iterator over Triples.
* This implementation assumes that the underlying findWithContinuation
* will have also consulted the raw data.
*/
@Override
public ExtendedIterator<Triple> graphBaseFind(Node subject, Node property, Node object) {
return findWithContinuation(new TriplePattern(subject, property, object), null);
}
/**
* Basic pattern lookup interface.
* This implementation assumes that the underlying findWithContinuation
* will have also consulted the raw data.
* @param pattern a TriplePattern to be matched against the data
* @return a ExtendedIterator over all Triples in the data set
* that match the pattern
*/
@Override
public ExtendedIterator<Triple> find(TriplePattern pattern) {
return findWithContinuation(pattern, null);
}
/**
* Add one triple to the data graph, run any rules triggered by
* the new data item, recursively adding any generated triples.
*/
@Override
public synchronized void performAdd(Triple t) {
version++;
fdata.getGraph().add(t);
if (this.isPrepared()) {
engine.add(t);
}
}
/**
* Return the number of triples in the inferred graph
*/
@Override
public int graphBaseSize() {
checkOpen();
this.requirePrepared();
int baseSize = fdata.getGraph().size();
int dedSize = fdeductions.getGraph().size();
// System.err.println( ">> BasicForwardRuleInfGraph::size = " + baseSize + "(base) + " + dedSize + "(deductions)" );
return baseSize + dedSize;
}
/**
* Removes the triple t (if possible) from the set belonging to this graph.
*/
@Override
public void performDelete(Triple t) {
version++;
if (fdata != null) {
Graph data = fdata.getGraph();
if (data != null) {
data.delete(t);
}
}
if (!this.isPrepared()) {
fdeductions.getGraph().delete(t);
}
}
/**
* Free all resources, any further use of this Graph is an error.
*/
@Override
public void close() {
if (!closed) {
engine = null;
fdeductions = null;
rules = null;
schemaGraph = null;
super.close();
}
}
// =======================================================================
// Implementation of ForwardRuleInfGraphI interface which is used by
// the forward rule engine to invoke functions in this InfGraph
/**
* Adds a new Backward rule as a rules of a forward rule process. Only some
* infgraphs support this.
*/
@Override
public void addBRule(Rule brule) {
throw new ReasonerException("Forward reasoner does not support hybrid rules - " + brule.toShortString());
}
/**
* Deletes a new Backward rule as a rules of a forward rule process. Only some
* infgraphs support this.
*/
@Override
public void deleteBRule(Rule brule) {
throw new ReasonerException("Forward reasoner does not support hybrid rules - " + brule.toShortString());
}
/**
* Return the Graph containing all the static deductions available so far.
* Will force a prepare.
*/
@Override
public Graph getDeductionsGraph() {
prepare();
return safeDeductions;
}
/**
* Create the graph used to hold the deductions. Can be overridden
* by subclasses that need special purpose graph implementations here.
* Assumes the graph underlying fdeductions and associated SafeGraph
* wrapper can be reused if present thus enabling preservation of
* listeners.
*/
protected Graph createDeductionsGraph() {
if (fdeductions != null) {
Graph dg = fdeductions.getGraph();
if (dg != null) {
// Reuse the old graph in order to preserve any listeners
safeDeductions.clear();
return dg;
}
}
Graph dg = Factory.createGraphMem( );
safeDeductions = new SafeGraph( dg );
return dg;
}
/**
* Return the Graph containing all the static deductions available so far.
* Does not trigger a prepare action. Returns a SafeWrapper and so
* can be used for update (thus triggering listeners) but not
* for access to generalized triples
*/
@Override
public Graph getCurrentDeductionsGraph() {
return safeDeductions;
// return fdeductions.getGraph();
}
/**
* Search the combination of data and deductions graphs for the given triple pattern.
* This may different from the normal find operation in the base of hybrid reasoners
* where we are side-stepping the backward deduction step.
*/
@Override
public ExtendedIterator<Triple> findDataMatches(Node subject, Node predicate, Node object) {
return findWithContinuation(new TriplePattern(subject, predicate, object), null, false);
}
/**
* Log a dervivation record against the given triple.
*/
@Override
public void logDerivation(Triple t, Derivation derivation) {
derivations.put(t, derivation);
}
/**
* Assert a new triple in the deduction graph, bypassing any processing machinery.
*/
@Override
public void silentAdd(Triple t) {
fdeductions.getGraph().add(t);
}
//=======================================================================
// support for proof traces
/**
* Set to true to enable derivation caching
*/
@Override
public void setDerivationLogging(boolean recordDerivations) {
this.recordDerivations = recordDerivations;
engine.setDerivationLogging(recordDerivations);
if (recordDerivations) {
derivations = new OneToManyMap<>();
} else {
derivations = null;
}
}
/**
* Return true if derivation logging is enabled.
*/
@Override
public boolean shouldLogDerivations() {
return recordDerivations;
}
/**
* Return the derivation of at triple.
* The derivation is a List of DerivationRecords
*/
@Override
public Iterator<Derivation> getDerivation(Triple t) {
if (derivations == null) {
return new NullIterator<>();
} else {
return derivations.getAll(t);
}
}
/**
* Set the state of the trace flag. If set to true then rule firings
* are logged out to the Log at "INFO" level.
*/
public void setTraceOn(boolean state) {
traceOn = state;
}
/**
* Return true if tracing should be acted on - i.e. if traceOn is true
* and we are past the bootstrap phase.
*/
@Override
public boolean shouldTrace() {
return traceOn && engine.shouldTrace();
}
/**
* Return the number of rules fired since this rule engine instance
* was created and initialized
*/
public long getNRulesFired() {
return engine.getNRulesFired();
}
}