net.java.rdf.sommer
Class SesameMapper

java.lang.Object
  net.java.rdf.sommer.Mapping
      net.java.rdf.sommer.SesameMapper

All Implemented Interfaces:

Mapper, RDFFactory, RewriteMapper

public class SesameMapper
extends Mapping

Author:

Henry Story

Nested Class Summary

(package private) class	SesameMapper.FilterFieldsSerialiser filter for fields that have given annotations Note: this is used by findFromCifpConstructor (perhaps the class should be inernal to the method) also: this should be more subtle: if fields are private and in a superclass, then should we find a solution?
(package private) class	SesameMapper.IdFinderSerialiser Tries to find the Identity of an object by looking for identiy or inverse functional properties.
(package private) class	SesameMapper.MappingSerialiser take an unmapped object and add all its relations to the graph
class	SesameMapper.QueryByExampleSerialiser
static class	SesameMapper.SesameInit

Nested classes/interfaces inherited from interface net.java.rdf.sommer.Mapper

Mapper.Init

Field Summary

(package private) org.openrdf.model.Resource[]	graphs
private boolean	inference
(package private) SesameMapper.SesameInit	init
(package private) java.util.HashMap<java.lang.Class,JavaInstanceMapper>	literalMap
(package private) static java.util.logging.Logger	log
private java.util.WeakHashMap<java.lang.Object,org.openrdf.model.Resource>	obj2Resource A mapping from objects to resources is one to one.
(package private) org.openrdf.repository.RepositoryConnection	rep
private java.util.WeakHashMap<org.openrdf.model.Resource,java.util.List<java.lang.Object>>	resource2Obj A mapping from resources to objects is one to many.
private static java.util.TimeZone	TZ
(package private) java.util.Date	updateTime
org.openrdf.model.ValueFactory	vf
(package private) org.openrdf.model.Resource	writeGrphs
(package private) static javax.xml.datatype.DatatypeFactory	xmldf

Fields inherited from class net.java.rdf.sommer.Mapping

classmap

Constructor Summary

SesameMapper(SesameMapper.SesameInit init, java.lang.String... urls)
Constructs a repository tied to a graph

Method Summary

(package private) void	addContext(java.lang.String context) add a context to the list of 'read only contexts'
(package private) void	addJavaInstance(JavaInstanceMapper map)
<T> T	addObject(T object) Insert the object and all its dependent objects into the graph
boolean	addObjects(java.lang.Object... objects) Insert all objects
boolean	addRelation(org.openrdf.model.Resource source, org.openrdf.model.URI relation, boolean inverse, java.lang.Object obj)
private void	addStatement(Statement s)
void	addStatements(java.util.Collection<Statement> statements) Add the following statements to the writeable graph.
void	cifpName(java.lang.Object thiz, java.lang.Class clazz, java.lang.Class[] argTypes, java.lang.Object[] values) Set the name for thiz to an existing one if one exists.
<T> T	cifpObject(java.lang.Class<T> clazz, java.lang.Class[] argTypes, java.lang.Object[] values) Check to see if we don't allready have an object that corresponds to this cifp
boolean	clear() note this clear here only clears what was written to the graph.
java.lang.Object	createLiteral(java.lang.String s)
java.lang.Object	createLiteral(java.lang.String s, java.lang.String lang)
java.lang.Object	createLiteralType(java.lang.String s, java.lang.String uriType)
<T> T	createObjectWithId(java.lang.String uri, java.lang.Class<T> clazz) Create an object of type with a given uri
java.lang.Object	createResource(java.lang.String uri)
<T> RDFCollection<T>	createVirtualCollection(java.lang.Object thiz, rdf relation, java.util.Collection<T> oldFieldValue, java.lang.reflect.Type fieldType) get the value of a virtual collection field.
static java.lang.String	dateToXsdString(java.util.Date date)
void	enableInferencing(boolean on)
boolean	equals(java.lang.Object mapped1, java.lang.Object mapped2) Equality of mapped objects.
private org.openrdf.model.Resource	findFromCifpConstructor(java.lang.reflect.Constructor con, java.lang.Object obj)
(package private) org.openrdf.model.Value	findKnownMappedValueFor(java.lang.Object obj) Just find the Value of the object if it is known or a literal
(package private) org.openrdf.model.Value	findKnownMappedValueFor(java.lang.Object obj, java.lang.String literalType) Find a known mapped value of the object, whith the given type
<T> java.util.Collection<T>	getAllObjectsOfType(java.lang.Class<T> clazz)
private org.openrdf.model.Resource	getCIFPValue(java.lang.Class clazz, java.lang.Class[] argTypes, java.lang.Object[] values) Find the object, if one exists that has the relations given by the rdf annotations on the arguments of the constructor clazz, to the values.
java.util.List<java.lang.Class>	getClassesOf(java.lang.String uri) find the classes of the object with id uri
<E> java.util.ArrayList<E>	getCollection(org.openrdf.model.Resource subj, org.openrdf.model.URI relationUri, boolean inverse, java.lang.Class<E> clazz) Get the collection starting with subj and the given relation, with contents of type clazz
<T> java.util.Collection<T>	getCollectionField(java.lang.Object thiz, rdf relation, java.util.Collection<T> oldFieldValue, java.lang.reflect.Type fieldType) get the value of a collection field.
java.lang.Object	getField(java.lang.Class fieldClass, java.lang.Object sourceObj, rdf relation, java.lang.Object fieldVal) Get the value of a field
<T> T	getObjectById(java.lang.String uri, java.lang.Class<T> clazz) Fetch an object of type with a given uri.
private java.util.ArrayList<java.lang.String>	getTypesOf(org.openrdf.model.Resource res) Get all the known types of resources res.
(package private) org.openrdf.model.Resource	getWriteGraphs()
java.lang.String	graphId()
void	importFrom(java.io.Reader r, java.lang.String baseUri, java.lang.String mimeType)
void	importInto(java.io.Reader r, java.lang.String baseUri, java.lang.String mimeType, java.lang.String context) import graph serialisation into the given context.
private void	initInstanceMapper()
boolean	isDirty(java.util.Date update)
boolean	isInferencingEnabled()
boolean	isMapped(java.lang.Object o) note: We should be able to ask the object directly now that it implements SommerMapable except of course that we are adding this directly
org.openrdf.model.Value	map(java.lang.Object obj) Map the java object to a RDF Value.
org.openrdf.model.Value	map(java.lang.Object obj, java.lang.String literalType) Map the java object to a RDF Value.
private <T> T	map(org.openrdf.model.Value value, java.lang.Class<T> sprClzz) map the resource into an object of the given class
void	name(java.lang.Object obj, org.openrdf.model.Resource id) remember the name of the object
void	output(java.io.OutputStream out) Output the contents of this graph todo: add type of output (N3,rdf/xml)
void	output(java.io.Writer out) Output the contents of this graph todo: add type of output (N3,rdf/xml)
<T> java.util.Collection<T>	queryByExample(java.lang.Object eg, java.lang.Class<T> clazz) Query the map by giving an example object, and return all objects that fit the example.
boolean	remove(java.lang.Object obj) remove all statements concerning the resource this object is mapped to.
void	removeAll(org.openrdf.model.Resource source, org.openrdf.model.URI relation, boolean inverse)
boolean	removeRelation(org.openrdf.model.Resource source, org.openrdf.model.URI relation, boolean inverse, java.lang.Object obj)
private void	removeStatement(Statement s)
void	removeStatements(java.util.Collection<Statement> statements) remove the following statements from the graphs.
org.openrdf.repository.RepositoryConnection	rep() todo: It should be possible to set different repositories for different Maps.
<T> RDFCollection<T>	replaceVirtualCollection(java.lang.Object thiz, rdf relation, java.util.Collection<T> oldColl, java.util.Collection<T> newColl, java.lang.reflect.Type genType) Replace the elements in the collection with the new elements.
(package private) <T> T	resourceGet(org.openrdf.model.Resource name, java.lang.Class<T> clazz) look in our map for an object with rdf name and the most precise subclass
(package private) void	setChanged()
<T> java.util.Collection<T>	setCollectionField(java.lang.Object sourceObj, rdf relation, java.util.Collection<T> newColl, java.lang.reflect.Type type) Set a Collection on a given field.
java.lang.Object	setField(java.lang.Class fieldClass, java.lang.Object obj, rdf relation, java.lang.Object value) Set a given field
long	size() The Size in triples of the graph, mapped to.
void	smush() Smush this graph.
boolean	unmap(java.lang.Object obj) This will remove the object from the map, but the relations will remain in the database.

Methods inherited from class net.java.rdf.sommer.Mapping

cogitate, mostSpecificSubClass, preload

Methods inherited from class java.lang.Object

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

Field Detail

log

static transient java.util.logging.Logger log

vf

public transient org.openrdf.model.ValueFactory vf

rep

transient org.openrdf.repository.RepositoryConnection rep

graphs

org.openrdf.model.Resource[] graphs

xmldf

static javax.xml.datatype.DatatypeFactory xmldf

TZ

private static final java.util.TimeZone TZ

inference

private boolean inference

init

SesameMapper.SesameInit init

writeGrphs

org.openrdf.model.Resource writeGrphs

obj2Resource

private java.util.WeakHashMap<java.lang.Object,org.openrdf.model.Resource> obj2Resource

A mapping from objects to resources is one to one. We only give one name to an object. The other owl:sameAs names can be found in the database.

resource2Obj

private java.util.WeakHashMap<org.openrdf.model.Resource,java.util.List<java.lang.Object>> resource2Obj

A mapping from resources to objects is one to many. This is because many objects can have the same name. For example - 2 objects implement different classes (one way to get multiple inheritance in java) - another object is created in a constructor that is an ifp or cifp and so the new object receives the same name

updateTime

java.util.Date updateTime

literalMap

java.util.HashMap<java.lang.Class,JavaInstanceMapper> literalMap

Constructor Detail

SesameMapper

SesameMapper(SesameMapper.SesameInit init,
             java.lang.String... urls)

Constructs a repository tied to a graph

Parameters:

init - Initialisation behavior

urls - the name of the graph or null for the default graph

Method Detail

getWriteGraphs

org.openrdf.model.Resource getWriteGraphs()

addContext

void addContext(java.lang.String context)

add a context to the list of 'read only contexts'

Parameters:

context -

isDirty

public boolean isDirty(java.util.Date update)

setChanged

void setChanged()

rep

public org.openrdf.repository.RepositoryConnection rep()

todo: It should be possible to set different repositories for different Maps. As such todo: we may not want this to be static... At the same time we may also want different maps for different todo: repositories

Returns:

importFrom

public void importFrom(java.io.Reader r,
                       java.lang.String baseUri,
                       java.lang.String mimeType)
                throws java.io.IOException,
                       GraphUpdateException,
                       ParseException

Parameters:

r - a Reader on the rdf

baseUri - the base to resolve relative uri from

mimeType - of the rdf serialisation: rdfxml, n3, turtle, ntriples, ...

Throws:

java.io.IOException

GraphUpdateException

ParseException

importInto

public void importInto(java.io.Reader r,
                       java.lang.String baseUri,
                       java.lang.String mimeType,
                       java.lang.String context)
                throws java.io.IOException,
                       GraphUpdateException,
                       ParseException

Description copied from interface: Mapper

import graph serialisation into the given context. Add the context to the list of contexts we are working with *

Parameters:

r - the reader containing the serialisation of the graph

baseUri - the base uri to xxx relative uris

mimeType - format of the serialisation ( currently the mime type)

context - the context to drop the information into

Throws:

java.io.IOException

GraphUpdateException

ParseException

clear

public boolean clear()

note this clear here only clears what was written to the graph. it does not remove all the facts from the read only graphs. Is this more useful that what we wanted before?

Returns:

replaceVirtualCollection

public <T> RDFCollection<T> replaceVirtualCollection(java.lang.Object thiz,
                                                     rdf relation,
                                                     java.util.Collection<T> oldColl,
                                                     java.util.Collection<T> newColl,
                                                     java.lang.reflect.Type genType)

Description copied from interface: RewriteMapper

Replace the elements in the collection with the new elements.

Parameters:

thiz - the object on which the field is attached

relation - the annotation on the field

oldColl - the old collection that the field was set to

newColl - the new collection the it should be replaced by

genType - the type of the objects in the collection

Returns:

the collection containing the objects related to thiz by relation. This may be the same object as oldColl

setCollectionField

public <T> java.util.Collection<T> setCollectionField(java.lang.Object sourceObj,
                                                      rdf relation,
                                                      java.util.Collection<T> newColl,
                                                      java.lang.reflect.Type type)

Description copied from interface: RewriteMapper

Set a Collection on a given field.

relation - rdf annotation on the field

Returns:

The object the field is set to

createVirtualCollection

public <T> RDFCollection<T> createVirtualCollection(java.lang.Object thiz,
                                                    rdf relation,
                                                    java.util.Collection<T> oldFieldValue,
                                                    java.lang.reflect.Type fieldType)

Description copied from interface: RewriteMapper

get the value of a virtual collection field. This will only be called once: on mapping of the object into the graph. Later the object present in the field will be returned. should this be also called when the object gets unmapped, so that it can return an unmapped collection?

Parameters:

thiz - the object on which the field is attached

relation - annotation on this field

oldFieldValue - the old value of the field - the one currently referred to

fieldType - Gernics type info of the Collection

Returns:

The object the field is set to

getCollectionField

public <T> java.util.Collection<T> getCollectionField(java.lang.Object thiz,
                                                      rdf relation,
                                                      java.util.Collection<T> oldFieldValue,
                                                      java.lang.reflect.Type fieldType)

Description copied from interface: RewriteMapper

get the value of a collection field. These are collections that have the (at)functional annotation (or something with a better name, when I get round to it) They are collections that really appear as collections in the rdf store.

relation - annotation on this field

oldFieldValue - the old value of the field - the one currently referred to

fieldType - the generic type of the collection

Returns:

The object the field is set to

setField

public java.lang.Object setField(java.lang.Class fieldClass,
                                 java.lang.Object obj,
                                 rdf relation,
                                 java.lang.Object value)

Description copied from interface: RewriteMapper

Set a given field

Parameters:

fieldClass - The class of the field being set

obj - The object on which the field is attached

relation - The rdf annotation on the field

value - the value that the field should be set to

Returns:

The object the field is set to

getField

public java.lang.Object getField(java.lang.Class fieldClass,
                                 java.lang.Object sourceObj,
                                 rdf relation,
                                 java.lang.Object fieldVal)

Description copied from interface: RewriteMapper

Get the value of a field

Parameters:

fieldClass - the class of the field

sourceObj - the object on which the field is located

relation - the rdf annotation on the field

fieldVal - the current value of the field

Returns:

the value of the field

name

public void name(java.lang.Object obj,
                 org.openrdf.model.Resource id)

remember the name of the object

Parameters:

obj -

id -

cifpName

public void cifpName(java.lang.Object thiz,
                     java.lang.Class clazz,
                     java.lang.Class[] argTypes,
                     java.lang.Object[] values)

Description copied from interface: RewriteMapper

Set the name for thiz to an existing one if one exists. The relations together form a CIFP (see http://esw.w3.org/topic/CIFP) The array size for relations and values should be the same

Parameters:

thiz - the object to be named

clazz - the class of the constructor

argTypes - the argument types identifying the particular constructor

values - the values

cifpObject

public <T> T cifpObject(java.lang.Class<T> clazz,
                        java.lang.Class[] argTypes,
                        java.lang.Object[] values)

Description copied from interface: RewriteMapper

Check to see if we don't allready have an object that corresponds to this cifp

Parameters:

clazz - the class on which the constructor is being called

argTypes - the types of the arguments of the constructor, to help find the constructor

Returns:

the object that allready corresponded to that cifp, or null if none

getCIFPValue

private org.openrdf.model.Resource getCIFPValue(java.lang.Class clazz,
                                                java.lang.Class[] argTypes,
                                                java.lang.Object[] values)

Find the object, if one exists that has the relations given by the rdf annotations on the arguments of the constructor clazz, to the values.

Parameters:

clazz - the class of the constructor

argTypes - the argument types identifying the particular constructor

values - the values that may have allready have a mapping in the rdf graph

Returns:

an rdf resource that has the required realtions to the mapped values, or null

addJavaInstance

void addJavaInstance(JavaInstanceMapper map)

initInstanceMapper

private void initInstanceMapper()

findKnownMappedValueFor

org.openrdf.model.Value findKnownMappedValueFor(java.lang.Object obj)

Just find the Value of the object if it is known or a literal

Parameters:

obj -

Returns:

findKnownMappedValueFor

org.openrdf.model.Value findKnownMappedValueFor(java.lang.Object obj,
                                                java.lang.String literalType)

Find a known mapped value of the object, whith the given type

Parameters:

obj - the object we are looking a mapping for

literalType - the type of the literal

Returns:

dateToXsdString

public static java.lang.String dateToXsdString(java.util.Date date)

map

public org.openrdf.model.Value map(java.lang.Object obj,
                                   java.lang.String literalType)

Map the java object to a RDF Value.

Parameters:

obj -

literalType - the type the object should be mapped to, or "" if not a literal

Returns:

map

public org.openrdf.model.Value map(java.lang.Object obj)

Map the java object to a RDF Value.

Parameters:

obj -

Returns:

findFromCifpConstructor

private org.openrdf.model.Resource findFromCifpConstructor(java.lang.reflect.Constructor con,
                                                           java.lang.Object obj)

smush

public void smush()

Smush this graph. Need this because don't have an inferencing engine

map

private <T> T map(org.openrdf.model.Value value,
                  java.lang.Class<T> sprClzz)

map the resource into an object of the given class

Parameters:

value -

sprClzz - the class or super class of any object returned

Returns:

the object, or null if no object could be constructed

getTypesOf

private java.util.ArrayList<java.lang.String> getTypesOf(org.openrdf.model.Resource res)

Get all the known types of resources res.

Parameters:

res -

Returns:

the known types as a list of URI types

resourceGet

<T> T resourceGet(org.openrdf.model.Resource name,
                  java.lang.Class<T> clazz)

look in our map for an object with rdf name and the most precise subclass

Parameters:

name - the rdf

clazz - the supertype of the object to be returned

Returns:

the object in our map, if one exists

removeRelation

public boolean removeRelation(org.openrdf.model.Resource source,
                              org.openrdf.model.URI relation,
                              boolean inverse,
                              java.lang.Object obj)

addRelation

public boolean addRelation(org.openrdf.model.Resource source,
                           org.openrdf.model.URI relation,
                           boolean inverse,
                           java.lang.Object obj)

Parameters:

source -

relation -

obj -

removeAll

public void removeAll(org.openrdf.model.Resource source,
                      org.openrdf.model.URI relation,
                      boolean inverse)

getCollection

public <E> java.util.ArrayList<E> getCollection(org.openrdf.model.Resource subj,
                                                org.openrdf.model.URI relationUri,
                                                boolean inverse,
                                                java.lang.Class<E> clazz)

Get the collection starting with subj and the given relation, with contents of type clazz

Parameters:

subj - the name of the object from which to go

relationUri - the relation from subj

inverse - if the inverse of the relationURi is desired

clazz - the class to return (could be a subclass if this is warranted by the database)

Returns:

a collection of java objects that are related by the relationUri relation to the collection

isMapped

public boolean isMapped(java.lang.Object o)

Description copied from interface: Mapper

note: We should be able to ask the object directly now that it implements SommerMapable except of course that we are adding this directly

Returns:

true if the object is mapped, false otherwise

getObjectById

public <T> T getObjectById(java.lang.String uri,
                           java.lang.Class<T> clazz)

Description copied from interface: Mapper

Fetch an object of type with a given uri. Note. There may be a good case for having a similar method that can also take a literal.

Parameters:

uri - the id of the object

clazz - the type of the object looked for

Returns:

null if none available in the database or an object mapped (if there is already an object mapped to that id, it should return that one)

createObjectWithId

public <T> T createObjectWithId(java.lang.String uri,
                                java.lang.Class<T> clazz)

Description copied from interface: Mapper

Create an object of type with a given uri

Parameters:

uri - the id of the object

clazz - the type of the object looked for

Returns:

an object (null only if the class has no @rdf annotation?)

getAllObjectsOfType

public <T> java.util.Collection<T> getAllObjectsOfType(java.lang.Class<T> clazz)

addObject

public <T> T addObject(T object)

Description copied from interface: Mapper

Insert the object and all its dependent objects into the graph

Parameters:

object - the object whose relations should be added to the graph

Returns:

the same object, for method chaining

addObjects

public boolean addObjects(java.lang.Object... objects)

Description copied from interface: Mapper

Insert all objects

Returns:

addStatements

public void addStatements(java.util.Collection<Statement> statements)

Description copied from interface: Mapper

Add the following statements to the writeable graph. The Subject and Object of the Statement can be mapped objects, in which case their id will be used. If they are not mapped objects, what should one do? Map them first and then add the statement?

addStatement

private void addStatement(Statement s)

removeStatements

public void removeStatements(java.util.Collection<Statement> statements)

Description copied from interface: Mapper

remove the following statements from the graphs. (this may require making diffs of non writeable graphs)

removeStatement

private void removeStatement(Statement s)

remove

public boolean remove(java.lang.Object obj)

Description copied from interface: Mapper

remove all statements concerning the resource this object is mapped to. The object will no longer appear in the rdf graph after this. note: On inferencing graphs this may not lead to the resource being removed if the resource is implied by other facts in the database. todo: this needs to be looked at a lot more carefully todo: is it a good thing to be able to remove an object? It could be quite a problem, as it could invalidate a number of other objects...

Returns:

false if it never was part of the graph

unmap

public boolean unmap(java.lang.Object obj)

Description copied from interface: Mapper

This will remove the object from the map, but the relations will remain in the database. todo: understand what this would mean for all the objects in points to or all objects pointed to it return true if unmapped, false if object was never mapped

output

public void output(java.io.OutputStream out)

Description copied from interface: Mapper

Output the contents of this graph todo: add type of output (N3,rdf/xml)

output

public void output(java.io.Writer out)

Description copied from interface: Mapper

Output the contents of this graph todo: add type of output (N3,rdf/xml)

graphId

public java.lang.String graphId()

equals

public boolean equals(java.lang.Object mapped1,
                      java.lang.Object mapped2)

Description copied from interface: Mapper

Equality of mapped objects. As far as the mapper knows these two objects are equals. Inferencing graphs will do a better job of finding object equality than non inferencing graphs. We can't use Object.equals(o) because that is too closely tied to the hash signature. In the graph equality between objects can be discovered over time, as new relations are added to the database. Equality between object may possibly also be broken, as relations are altered. if either object is not mapped, then this returns false if both objects are mapped and == then this is true if there is a owl:sameAs relation between them then it is true else it is false.

Parameters:

mapped1 - a mapped object

mapped2 - another mapped objects

Returns:

as far as this mapper knows mapped1 and mapped2 are the same

queryByExample

public <T> java.util.Collection<T> queryByExample(java.lang.Object eg,
                                                  java.lang.Class<T> clazz)

Description copied from interface: Mapper

Query the map by giving an example object, and return all objects that fit the example. Here the example is an object with @rdf annotated fields. Only looks at fields with values. Note: objects with simple literal values such as char, int, etc... are always set. So beware.

Parameters:

eg - example object. Should not be a mapped object or an object with a id specified by @rdf URI id since otherwise the only object returneable would be the one given. The object can have fields that are set either to mapped or unmapped objects.

clazz - of objects to return (should this not just be the same as the class of o?

Returns:

a Collection of objects of the given that fit the example.

createLiteral

public java.lang.Object createLiteral(java.lang.String s)

createLiteral

public java.lang.Object createLiteral(java.lang.String s,
                                      java.lang.String lang)

createLiteralType

public java.lang.Object createLiteralType(java.lang.String s,
                                          java.lang.String uriType)

createResource

public java.lang.Object createResource(java.lang.String uri)

size

public long size()

Description copied from interface: Mapper

The Size in triples of the graph, mapped to.

Returns:

the size of the graph, or -1 if there was a problem finding out

getClassesOf

public java.util.List<java.lang.Class> getClassesOf(java.lang.String uri)

Description copied from interface: Mapper

find the classes of the object with id uri

Parameters:

uri - the id of the object looked for

Returns:

a list of classes

enableInferencing

public void enableInferencing(boolean on)

isInferencingEnabled

public boolean isInferencingEnabled()