Simple DOM Model Example

The Simple DOM Model example shows how an existing class can be adapted for use with the model/view framework.

../_images/simpledommodel-example.png

Qt provides two complementary sets of classes for reading XML files: The classes based around QXmlReader provide a SAX-style API for incremental reading of large files, and the classes based around QDomDocument enable developers to access the contents of XML files using a Document Object Model (DOM) API.

In this example, we create a model that uses the DOM API to expose the structure and contents of XML documents to views via the standard QAbstractModel interface.

Design and Concepts

Reading an XML document with Qt’s DOM classes is a straightforward process. Typically, the contents of a file are supplied to QDomDocument , and nodes are accessed using the functions provided by QDomNode and its subclasses.

The aim is to use the structure provided by QDomDocument by wrapping QDomNode objects in item objects similar to the TreeItem objects used in the Simple Tree Model example.

DomModel Class Definition

Let us begin by examining the DomModel class:

class DomModel : public QAbstractItemModel
{
    Q_OBJECT

public:
    explicit DomModel(const QDomDocument &document, QObject *parent = nullptr);
    ~DomModel();

    QVariant data(const QModelIndex &index, int role) const override;
    Qt::ItemFlags flags(const QModelIndex &index) const override;
    QVariant headerData(int section, Qt::Orientation orientation,
                        int role = Qt::DisplayRole) const override;
    QModelIndex index(int row, int column,
                      const QModelIndex &parent = QModelIndex()) const override;
    QModelIndex parent(const QModelIndex &child) const override;
    int rowCount(const QModelIndex &parent = QModelIndex()) const override;
    int columnCount(const QModelIndex &parent = QModelIndex()) const override;

private:
    QDomDocument domDocument;
    DomItem *rootItem;
};

The class definition contains all the basic functions that are needed for a read-only model. Only the constructor and document() function are specific to this model. The private domDocument variable is used to hold the document that is exposed by the model; the rootItem variable contains a pointer to the root item in the model.

DomItem Class Definition

The DomItem class is used to hold information about a specific QDomNode in the document:

class DomItem
{
public:
    DomItem(const QDomNode &node, int row, DomItem *parent = nullptr);
    ~DomItem();
    DomItem *child(int i);
    DomItem *parent();
    QDomNode node() const;
    int row() const;

private:
    QDomNode domNode;
    QHash<int, DomItem *> childItems;
    DomItem *parentItem;
    int rowNumber;
};

Each DomItem provides a wrapper for a QDomNode obtained from the underlying document which contains a reference to the node, it’s location in the parent node’s list of child nodes, and a pointer to a parent wrapper item.

The parent(), child(), and row() functions are convenience functions for the DomModel to use that provide basic information about the item to be discovered quickly. The node() function provides access to the underlying QDomNode object.

As well as the information supplied in the constructor, the class maintains a cache of information about any child items. This is used to provide a collection of persistent item objects that the model can identify consistently and improve the performance of the model when accessing child items.

DomItem Class Implementation

Since the DomItem class is only a thin wrapper around QDomNode objects, with a few additional features to help improve performance and memory usage, we can provide a brief outline of the class before discussing the model itself.

The constructor simply records details of the QDomNode that needs to be wrapped:

DomItem::DomItem(const QDomNode &node, int row, DomItem *parent)
    : domNode(node),              parentItem(parent),
      rowNumber(row)
{}

As a result, functions to provide the parent wrapper, the row number occupied by the item in its parent’s list of children, and the underlying QDomNode for each item are straightforward to write:

DomItem *DomItem::parent()
{
    return parentItem;
}
int DomItem::row() const
{
    return rowNumber;
}
QDomNode DomItem::node() const
{
    return domNode;
}

It is necessary to maintain a collection of items which can be consistently identified by the model. For that reason, we maintain a hash of child wrapper items that, to minimize memory usage, is initially empty. The model uses the item’s child() function to help create model indexes, and this constructs wrappers for the children of the item’s QDomNode , relating the row number of each child to the newly-constructed wrapper:

DomItem *DomItem::child(int i)
{
    DomItem *childItem = childItems.value(i);
    if (childItem)
        return childItem;

    // if child does not yet exist, create it
    if (i >= 0 && i < domNode.childNodes().count()) {
        QDomNode childNode = domNode.childNodes().item(i);
        childItem = new DomItem(childNode, i, this);
        childItems[i] = childItem;
    }
    return childItem;
}

If a QDomNode was previously wrapped, the cached wrapper is returned; otherwise, a new wrapper is constructed and stored for valid children, and zero is returned for invalid ones.

The class’s destructor deletes all the child items of the wrapper:

DomItem::~DomItem()
{
    qDeleteAll(childItems);
}

These, in turn, will delete their children and free any QDomNode objects in use.

DomModel Class Implementation

The structure provided by the DomItem class makes the implementation of DomModel similar to the TreeModel shown in the Simple Tree Model example.

The constructor accepts an existing document and a parent object for the model:

DomModel::DomModel(const QDomDocument &document, QObject *parent)
    : QAbstractItemModel(parent),
      domDocument(document),
      rootItem(new DomItem(domDocument, 0))
{
}

A shallow copy of the document is stored for future reference, and a root item is created to provide a wrapper around the document. We assign the root item a row number of zero only to be consistent since the root item will have no siblings.

Since the model only contains information about the root item, the destructor only needs to delete this one item:

DomModel::~DomModel()
{
    delete rootItem;
}

All of the child items in the tree will be deleted by the DomItem destructor as their parent items are deleted.

Basic Properties of The Model

Some aspects of the model do not depend on the structure of the underlying document, and these are simple to implement.

The number of columns exposed by the model is returned by the columnCount() function:

int DomModel::columnCount(const QModelIndex &parent) const
{
    Q_UNUSED(parent);
    return 3;
}

This value is fixed, and does not depend on the location or type of the underlying node in the document. We will use these three columns to display different kinds of data from the underlying document.

Since we only implement a read-only model, the flags() function is straightforward to write:

Qt::ItemFlags DomModel::flags(const QModelIndex &index) const
{
    if (!index.isValid())
        return Qt::NoItemFlags;

    return QAbstractItemModel::flags(index);
}

Since the model is intended for use in a tree view, the headerData() function only provides a horizontal header:

QVariant DomModel::headerData(int section, Qt::Orientation orientation,
                              int role) const
{
    if (orientation == Qt::Horizontal && role == Qt::DisplayRole) {
        switch (section) {
            case 0:
                return tr("Name");
            case 1:
                return tr("Attributes");
            case 2:
                return tr("Value");
            default:
                break;
        }
    }
    return QVariant();
}

The model presents the names of nodes in the first column, element attributes in the second, and any node values in the third.

Implementation Notes

Ideally, we would rely on the structure provided by QDomDocument to help us write the parent() and index() functions that are required when subclassing QAbstractItemModel . However, since Qt’s DOM classes use their own system for dynamically allocating memory for DOM nodes, we cannot guarantee that the QDomNode objects returned for a given piece of information will be the same for subsequent accesses to the document.

We use item wrappers for each QDomNode to provide consistent pointers that the model can use to navigate the document structure.

Example project @ code.qt.io