Question

I recently started rewriting an application and I’m trying to port it to model/view to reduce the number of kludges I have there.

So far I was able to succcessfully make a read-only model inheriting from QAbstractTableModel. This model is something like this:

class MyModel(QtCore.QAbstractTableModel):
    def __init__(self, data, parent=None):
        super(MyModel, self).__init__(parent)
        self.data = data

data is a list that contains a number of objects. These are then accessed in the data() method:

   def data(self, index, role):

       # much stuff omitted for clarity
       return QtCore.QVariant(self.data[index.column()].id)

Now this is fine if I work with a predefined data element. But in fact data changes programmatically (it expands when certain signals are received). How can I keep the model aware of this, so that then my view can also react to these changes?

I’ve been reading about read-write models but they also allow the user to edit and change things, while in my view I want things to be un-editable: in short, the model would need to be changed “behind the scenes” only, and the view adapt to that.

What is the best appproach in this case? Implement a read-write model with setData() and so on, or is there a simpler solution?

Answer 1

My logview application uses QAbstractTableModel with a dynamic list – of log records, which increases dynamically as records are received across the network. You can download the source code and take a look at LogRecordModel, which is a QAbstractTableModel subclass, as is PropertySheetModel. Both of these are read-only models.

It’s a pretty simple application using Qt model/view APIs, so you should be able to adapt its techniques for your own application.