|
Home · All Classes · Main Classes · Grouped Classes · Modules · Functions | |
The QAbstractXmlNodeModel class provides an abstract interface for representing arbitrary data as XML. More...
#include <QAbstractXmlNodeModel>
Inherits QSharedData.
Inherited by QPatternist::AccelTree, QPatternist::QObjectNodeModel, and QSimpleXmlNodeModel.
Note: All the functions in this class are thread-safe.
This class was introduced in Qt 4.4.
The QAbstractXmlNodeModel class provides an abstract interface for representing arbitrary data as XML.
The QAbstractXmlNodeModel specifies the interface that a node model must implement for that node model be accessible to XQuery queries. A node model represents data as a structure that can be queried as if the data were XML.
QAbstractXmlNodeModel bridges the gap between the arbitrary structure of the non-XML data to be queried and the well-defined structure of XML data understood by QXmlQuery.
Consider a chemistry application that reads a chemistryData file containing non-XML data representing a chemical structure of molecules and atoms. The application will query this chemistry data with an XQuery it reads from a queryFile. We write a custom subclass of QAbstractXmlNodeModel (ChemistryNodeModel) that reads the chemistryData file and builds a data structure, perhaps composed of objects of our own classes molecule and atom. Clearly, this data structure is not XML. Our custom subclass will know how to traverse this non-XML structure and present it through the XDM interface.
QFile queryFile(argv[1]);
QFile chemistryData(argv[2]);
QString moleculeName = argv[3];
QXmlQuery query;
query.setQuery(&queryFile, QUrl::fromLocalFile(queryFile.fileName()));
ChemistryNodeModel myNodeModel(query.namePool(), chemistryData);
QXmlNodeModelIndex startNode = myNodeModel.nodeFor(moleculeName);
query.bindVariable("queryRoot", startNode);
QFile out;
out.open(stdout, QIODevice::WriteOnly);
QXmlSerializer serializer(query, &out);
query.evaluateTo(&serializer);
The application first creates an instance of QXmlQuery and calls setQuery() to read the queryFile containing the XQuery we want to run. Then it creates an instance of our custom node model class, ChemistryNodeModel, which is a subclass of QAbstractXmlNodeModel. Its constructor is called with the name pool obtained from our QXmlQuery, and with the chemistryFile containing the structure of molecules and atoms to be queried. The name pool is required because our custom node model has the member function name(), which returns the name of any node in the model. The query and the custom node model must use the same name pool for constructing these names. The constructor would then read the chemistryFile and build the custom node model structure.
To connect the query to the custom node model, we must bind a variable name used in the query to a node in the model. The variable can then be used in the query as a starting node. First, an index for the desired starting node is retrieved by calling QAbstractXmlNodeModel::createIndex(). Then the index is bound to a variable name, in this case queryRoot, by passing the name and the index to QXmlQuery::bindVariable(). The query can then use a variable reference $queryRoot to refer to the starting node. Note that if the query uses multiple $variable references, a call to QXmlQuery::bindVariable() is required to bind each different variable name to a node in the model.
The query is executed when the application calls one of the QXmlQuery evaluation functions, evaluateTo() in this case, which passes myNodeModel to the query engine as a delegate the engine can use for traversing the non-XML chemistry data as if it were XML. The application uses QXmlQuery::evaluateTo(), because it then uses a serializer to out the query result as XML to stdout. We could have used QXmlQuery::evaluateTo() to get a list of result items, or QXmlQuery::evaluateTo() if the query evaluated to a sequence of xs:string values.
During query execution, the engine iterates over the node model using nextFromSimpleAxis() to get the index of the next node to be visited. The engine can get the name of a node by calling name() with the node's index. stringValue(), baseUri(), documentUri() and kind() are also called as needed with a node index.
The example demonstrates the standard pattern for using a subclass of QAbstractXmlNodeModel in combination with QXmlQuery to perform an XQuery.
Because the XDM interface presented by QAbstractXmlNodeModel allows QXmlQuery to operate on non-XML data as if it were XML, implementing subclasses of QAbstractXmlNodeModel can involve a significant amount of work. The QSimpleXmlNodeModel class is provided to simplify the implementation for many common use cases.
Because the node model can be accessed concurrently by threads in the QtXmlPatterns module, subclasses of QAbstractXmlNodeModel must be written to be thread-safe. Classes that simplify implementing thread-safety include QReadLocker and QWriteLocker.
See the example File System Example for a demonstration.
A smart pointer to a QAbstractXmlNodeModel instance.
Four axes that only contain one node each.
| Constant | Value | Description |
|---|---|---|
| QAbstractXmlNodeModel::Parent | 0 | The parent of the context node |
| QAbstractXmlNodeModel::FirstChild | 1 | The first child of the context node |
| QAbstractXmlNodeModel::PreviousSibling | 2 | The previous sibling of the context node |
| QAbstractXmlNodeModel::NextSibling | 3 | The next sibling of the context node |
Default constructor.
Destructor.
Returns the attributes that element element has.
The caller guarantees that element is always an element.
Returns the base URI for n.
The caller guarantees that n is not null and that it belongs to this QAbstractXmlNodeModel instance.
The base URI of nodes can be extracted using the fn:base-uri() function, and is typically used for resolving relative URIs appearing directly, or as child of the node. It is conformant to just return the document URI, although that might not properly reflect the underlying data.
This function maps to the dm:base-uri accessor, which the specification completely specifies. Here's a summary:
The implementation guarantees to return a valid QUrl instance, or a default constructed QUrl instance. If a node has no base URI, such as in the case of that a comment has no parent, a default constructed QUrl is returned.
See also XQuery 1.0 and XPath 2.0 Data Model (XDM), 5.2 base-uri Accessor.
Returns the relative document order between ni1 and ni2.
This is used for the is operator, and to sort nodes in document order.
The caller guarantees that ni1 and ni2 are not null and that they belong to this QAbstractXmlNodeModel instance.
If ni1 is identical to ni2, QXmlNodeModelIndex::Is is returned. If ni1 precedes ni2 in document order, QXmlNodeModelIndex::Precedes is returned. If ni1 follows ni2 in document order, QXmlNodeModelIndex::Follows is returned.
Creates a node index with data as its internal data.
What data is, is not constrained.
This is an overloaded member function, provided for convenience.
Creates a node index with pointer and additionalData as its internal data.
What pointer and additionalData is, is not constrained.
This is an overloaded member function, provided for convenience.
Creates a QXmlNodeModelIndex that houses data and additionalData.
Returns the document URI of n.
The document URI identifies the resource which is the document. For instance, if the document would be a regular file, it would perhaps be the file:/ or http:// URL of the location. The document URI is used for resolving URIs and to simply know where the document is.
If the data model maps to a URI in a natural way -- use that. Otherwise return the company or product's URI. It can be any URI as long as its valid and absolute.
The caller guarantees that n is not null and that it belongs to this QAbstractXmlNodeModel instance.
This function maps to the dm:document-uri accessor, which the specification completely specifies. Here's a summary:
- If n is a document node, return an absolute, valid QUrl containing the document URI, or a default constructed QUrl. The latter signals that no document URI is available for the document node. - For all other kinds of nodes, return a default constructed QUrl.
See also XQuery 1.0 and XPath 2.0 Data Model (XDM), 5.4 document-uri Accessor, QUrl::isValid(), and QUrl::isRelative().
Returns the element that has id id. XQuery's id() function ends up calling this.
Returns the element node whose typed value is of type ID and equals id, or the element that has an attribute whose typed value is of type ID and equals id. If there is no such element, a default constructed QXmlNodeModelIndex instance is returned. The implementor guarantees the returned node, if not null, is an element.
In effect it's not sufficient for an attribute or element to merely be called id in order to be of type ID. However, the reserved name xml:id is always recognized as so.
In id the namespace URI and the prefix is undefined, while the local name is the ID that should be looked up.
See also XQuery 1.0 and XPath 2.0 Functions and Operators, 15.5.2 fn:id.
Determines what node kind ni is. Simply, whether ni is an element or comment for instance.
The caller guarantees that ni is not null and that it belongs to this QAbstractXmlNodeModel instance.
This function maps to the dm:node-kind() accessor.
See also XQuery 1.0 and XPath 2.0 Data Model (XDM), 5.10 node-kind Accessor.
Returns the name of ni.
The caller guarantees that ni is not null and that it belong to this QAbstractXmlNodeModel instance.
If a node does not have a name, such as a comment code, a null QXmlName is returned. QXmlName instances must be created with the same QXmlQuery instance that is used for evaluating queries using this QAbstractXmlNodeModel instance.
This function maps to the dm:node-name() accessor.
As specified, if ni is a processing instruction, a QXmlName is returned where the local name is the target name and the namespace URI and prefix is empty.
See also XQuery 1.0 and XPath 2.0 Data Model (XDM), 5.11 node-name Accessor and QXmlName.
Returns the in-scope namespaces of n.
The caller guarantees that n is not null and that it belong to this QAbstractXmlNodeModel instance.
This corresponds to the dm:namespace-nodes accessor.
Note: This is not only the namespace declarations that appear on this element, but takes also into account namespace bindings of the ancestors.
The caller guarantees that n is an Element and belongs to this QAbstractXmlNodeModel instance.
When QtXmlPatterns calls iterate(), QSimpleXmlNodeModel create iterators that calls nextFromSimpleAxis() and "emulates" real XPath axes using QSimpleXmlNodeModel::SimpleAxis. Therefore, the implementation of this function should return the node, if any, that appear on axis axis, from origin.
If no such node is available, a default constructed QXmlNodeModelIndex is returned.
QSimpleXmlNodeModel removes the need to handle redundant corner cases by guaranteeing that it will never ask for:
A typical implementation does a switch over axis:
QXmlNodeModelIndex MyTreeModel::nextFromSimpleAxis(SimpleAxis axis, const QXmlNodeModelIndex &origin) const
{
// Convert the QXmlNodeModelIndex to a value that is specific to what we represent.
const MyValue value = toMyValue(ni);
switch(axis)
{
case Parent:
return toNodeIndex(value.parent());
case FirstChild:
case PreviousSibling:
case NextSibling:
// and so on
}
}
Returns the elements and/or attributes that an IDREF value equal to idref. XQuery's idref() function ends up calling this.
The implementor guarantees the returned nodes are elements or attributes.
In effect it's not sufficient for an attribute or element to merely be called idref in order to be of type IDREF. Elements must be typed as xs:IDREF or xs:IDREFS, or in the case of attributes only, as IDREF or IDREFS in the schema.
In idref the namespace URI and the prefix is undefined, while the local name is the ID that should be looked up.
See also XQuery 1.0 and XPath 2.0 Functions and Operators, 15.5.3 fn:idref.
Returns the root node of the tree that n is part of. This is typically a document node. This function is used among other things for fn:root() and the root expression, such as seen in the expression /html.
The caller guarantees that n is not null and that it belong to this QAbstractXmlNodeModel instance.
If n is a direct child of the QXmlNodeModelIndex returned from this function, parent() would return the same QXmlNodeModelIndex.
Returns the string value for node n.
The caller guarantees that n is not null and that it belong to this QAbstractXmlNodeModel instance.
This function maps to the dm:string-value() accessor, which the specification completely specifies. Here's a summary:
See also XQuery 1.0 and XPath 2.0 Data Model (XDM), 5.13 string-value Accessor.
Returns the typed value for node node.
The typed value is an atomic value, which an element or attribute contains.
The caller guarantees that node is either an element or an attribute. The implementor guarantees that the returned QVariant has a value which is supported in XQuery. It cannot be an arbitrary QVariant value. The implementor also guarantees that stringValue() returns a lexical representation of typedValue()(this is guaranteed by QSimpleXmlNodeModel::stringValue()).
If the return QVariant is a default constructed variant, it signals that node has no typed value.
| Copyright © 2008 Trolltech | Trademarks | Qt 4.4.0 |