net.sf.openschema

Class OpenSchemaPlanner

java.lang.Object

net.sf.openschema.OpenSchemaPlanner

public class OpenSchemaPlanner
extends Object

Main OpenSchema class. The schema is created from an XML file. At construction, the strategy for selecting local choices is also specified. Once created, the schema is instantiated with a frameset and an ontology. The result of an instantiated schema is a document plan.

Author:

Pablo Ariel Duboue

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
protected static class	OpenSchemaPlanner.Cache Values cache.
protected static class	OpenSchemaPlanner.DecoratedNode A decorated node inner class.
protected class	OpenSchemaPlanner.Node Node inner class.
protected class	OpenSchemaPlanner.Predicate Predicate inner class

Field Summary

Fields

Modifier and Type	Field and Description
protected static int	AGGRBOUNDARY_NODE Aggregation boundary node.
protected LocalChooser	chooser LocalChooser, decides which node to continue the instantiation of the schema.
protected static Frame	EMPTY_FOCUS
protected static int	EMPTY_NODE Empty node, only used for structural purposes.
protected static int	PARBOUNDARY_NODE Paragraph boundary node.
protected static int	PREDICATE_NODE Predicate node, contains a predicate reference plus global-to-local mapping.
protected Map<String,OpenSchemaPlanner.Predicate>	predicates Definition of the rhetorical predicates.
protected OpenSchemaPlanner.Node	top Top-level node of the schema, here is where the instantiation starts.
static boolean	verbose Verbosity flag, defaults to off.

Constructor Summary

Constructors

Constructor and Description
OpenSchemaPlanner(InputSource source, LocalChooser chooser) Constructor from an SAX InputSource.

Method Summary

Modifier and Type	Method and Description
protected Map<String,Object>	assembleFD(List<AttributeValuePair> vs, Set<String> vars) Create a functional description from the JAXB code.
protected void	assembleSchema(OpenSchema schema) Construct the nodes network from a JAXB generated schema class.
protected void	build(OpenSchemaPlanner.Node inNode, OpenSchemaPlanner.Node outNode, List<SchemaNode.Node> schemaNodes, boolean isChoice) Construct the network for a given construction in the schema language.
protected boolean	canBeInstantiated(OpenSchemaPlanner.Node node, Map<String,Frame> varMapping, OpenSchemaPlanner.Cache cache, Ontology ontology, FrameSet frames) Check whether there is valid data to instantiate a node or not, given an initial variable mapping.
protected List<OpenSchemaPlanner.DecoratedNode>	computeConfusionSet(OpenSchemaPlanner.Node node, Map<String,Frame> varMapping, OpenSchemaPlanner.Cache cache, Ontology ontology, FrameSet frames) Compute all the nodes that are reachable from current node without passing through another Predicate node.
String	dump() Show the network (for debugging purposes)
String	dump(boolean asDot) Show the network (for debugging purposes)
protected void	dump(StringBuilder dump, OpenSchemaPlanner.Node current, Set<OpenSchemaPlanner.Node> seen, boolean asDot) Recursive method for dump.
protected static List<Frame>	framesUnderType(Object type, Ontology ontology, FrameSet frames) Retrieve all the frames in a FrameSet which type is a subtype of a given type.
DocumentPlan	instantiate(FrameSet frames, Map<String,Frame> initialMapping, Ontology ontology) Main method, takes data to instantiate the schema, an initial variable mapping and an ontology and returns a DocumentPlan.
protected Map<String,Object>	instantiateClause(Map<String,Object> fd, Map<Object,Frame> valueMapping) Instantiate a given FD by changing all the variable references via a provided variable mapping.
protected Map<String,Object>	instantiatePredicate(OpenSchemaPlanner.Node node, Map<String,Frame> varMapping, OpenSchemaPlanner.Cache cache, boolean commit) Instantiate a predicate by fetching the values from the cache and resolving variables in the output FD of the predicate.
protected List<Map<Object,Frame>>	searchValues(OpenSchemaPlanner.Node node, Map<String,Frame> varMapping, Ontology ontology, FrameSet frames) Search for values that satisfy the node properties and the given mapping of variables.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Detail

verbose

public static boolean verbose

Verbosity flag, defaults to off.

top

protected OpenSchemaPlanner.Node top

Top-level node of the schema, here is where the instantiation starts. The whole network is constructed from the XML in the build method.

predicates

protected Map<String,OpenSchemaPlanner.Predicate> predicates

Definition of the rhetorical predicates. The keys are names ( String), the values are instances of the inner class Predicate.

chooser

protected LocalChooser chooser

LocalChooser, decides which node to continue the instantiation of the schema.

EMPTY_FOCUS

protected static Frame EMPTY_FOCUS

EMPTY_NODE

protected static final int EMPTY_NODE

Empty node, only used for structural purposes.

See Also:

Constant Field Values

PREDICATE_NODE

protected static final int PREDICATE_NODE

Predicate node, contains a predicate reference plus global-to-local mapping.

See Also:

Constant Field Values

AGGRBOUNDARY_NODE

protected static final int AGGRBOUNDARY_NODE

Aggregation boundary node.

See Also:

Constant Field Values

PARBOUNDARY_NODE

protected static final int PARBOUNDARY_NODE

Paragraph boundary node.

See Also:

Constant Field Values

Constructor Detail

OpenSchemaPlanner

public OpenSchemaPlanner(InputSource source,
                         LocalChooser chooser)
                  throws JAXBException

Constructor from an SAX InputSource.

Parameters:

source - A SAX InputSource, can read from a Reader or an InputStream.

chooser - An instance of the abstract class LocalChooser, it should select among the different continuation nodes during instantiation.

Throws:

JAXBException

Method Detail

instantiate

public DocumentPlan instantiate(FrameSet frames,
                                Map<String,Frame> initialMapping,
                                Ontology ontology)

Main method, takes data to instantiate the schema, an initial variable mapping and an ontology and returns a DocumentPlan.

Parameters:

frames - a class implementing the FrameSet interface, for example RDFFrameSet containing the data to fill the schema.

initialMapping - a mapping from variable names to values (Frames).

ontology - a class implementing the Ontology interface, for example RDFOntology.

computeConfusionSet

protected List<OpenSchemaPlanner.DecoratedNode> computeConfusionSet(OpenSchemaPlanner.Node node,
                                                                    Map<String,Frame> varMapping,
                                                                    OpenSchemaPlanner.Cache cache,
                                                                    Ontology ontology,
                                                                    FrameSet frames)

Compute all the nodes that are reachable from current node without passing through another Predicate node. The possible nodes should be able to be instantiated, that is to say, there must be sets of values satisfying the node's predicate properties in the cache. This method invokes the canBeInstantiated method, that will fill the cache the first time it is executed.

Parameters:

node - the current node to start the spanning tree.

varMapping - the current variable-values mapping.

cache - the current values cache.

ontology - used to update the cache.

frames - used to update the cache.

Returns:

a list of DecoratedNode that can be reached and instantiated from the current node.

canBeInstantiated

protected boolean canBeInstantiated(OpenSchemaPlanner.Node node,
                                    Map<String,Frame> varMapping,
                                    OpenSchemaPlanner.Cache cache,
                                    Ontology ontology,
                                    FrameSet frames)

Check whether there is valid data to instantiate a node or not, given an initial variable mapping. This function relies heavily in the cache, and will populate it in case of cache miss.

Parameters:

node - the current node, to analyze the predicate and global-to-local variable mapping.

varMapping - the global variables.

cache - the cache to be queried/populated.

ontology - used to populate the cache.

frames - used to populate the cache.

Returns:

whether or not there's data to instantiate this node.

searchValues

protected List<Map<Object,Frame>> searchValues(OpenSchemaPlanner.Node node,
                                               Map<String,Frame> varMapping,
                                               Ontology ontology,
                                               FrameSet frames)

Search for values that satisfy the node properties and the given mapping of variables.

Parameters:

node - the schema node, containing the properties that the values should satisfy and the mapping between local and global variables.

varMapping - global variables, mapping from name to value.

ontology - the ontology, employed to restrict the search over values of a certain type.

frames - the set of frames over where to perform the search.

Returns:

List of associations of local variables to values.

See Also:

{edu.columbia.openschema.OpenSchema.Node.Property}

framesUnderType

protected static List<Frame> framesUnderType(Object type,
                                             Ontology ontology,
                                             FrameSet frames)

Retrieve all the frames in a FrameSet which type is a subtype of a given type.

Parameters:

type - the parent type.

ontology - the ontology where to check for the relations.

frames - the frameset to look for the data.

Returns:

a list of frames.

instantiatePredicate

protected Map<String,Object> instantiatePredicate(OpenSchemaPlanner.Node node,
                                                  Map<String,Frame> varMapping,
                                                  OpenSchemaPlanner.Cache cache,
                                                  boolean commit)

Instantiate a predicate by fetching the values from the cache and resolving variables in the output FD of the predicate.

Parameters:

node - the node being instantiated.

varMapping - the global variables.

commit - whether or not to remove the values from the cache.

Returns:

the instantiated FD.

instantiateClause

protected Map<String,Object> instantiateClause(Map<String,Object> fd,
                                               Map<Object,Frame> valueMapping)

Instantiate a given FD by changing all the variable references via a provided variable mapping. Recursive function.

Parameters:

fd - the functional description to operate on.

valueMapping - variable-to-value mapping.

Returns:

a new functional description, with the variables replaced.

assembleSchema

protected void assembleSchema(OpenSchema schema)

Construct the nodes network from a JAXB generated schema class.

Parameters:

schema - the JAXB class, used to deal with the XML.

assembleFD

protected Map<String,Object> assembleFD(List<AttributeValuePair> vs,
                                        Set<String> vars)

Create a functional description from the JAXB code. All defined variables are changed for VarRef or PathRef, accordingly.

Parameters:

vs - the JAXB instances being processed.

vars - predicate variables, to replace the variable references.

Returns:

a newly created FD.

build

protected void build(OpenSchemaPlanner.Node inNode,
                     OpenSchemaPlanner.Node outNode,
                     List<SchemaNode.Node> schemaNodes,
                     boolean isChoice)

Construct the network for a given construction in the schema language. Most constructions can be taken as a sequence with an input node and an output node, but the choices. A flag signals whether to construct a choice or otherwise.

Parameters:

inNode - the input to this subnetwork.

outNode - the output of the subnetwork.

schemaNodes - the JAXB instances for this construction.

isChoice - whether to link each node one after the other or all from inNode to outNode.

dump

public String dump()

Show the network (for debugging purposes)

dump

public String dump(boolean asDot)

Show the network (for debugging purposes)

dump

protected void dump(StringBuilder dump,
                    OpenSchemaPlanner.Node current,
                    Set<OpenSchemaPlanner.Node> seen,
                    boolean asDot)

Recursive method for dump.