Using input layer as output in PyQGIS for QGIS Modeler?

Question

I have written a simple Processing script to rename a given field. It works successfully, changing the input layer's field name. However, I have not been able to produce an Output. This is a problem because I would like to use this script as an interim step in a Graphical Modeler tool, which requires an Output to pass on to the next step.

def processAlgorithm(self, parameters, context, feedback):

    source = self.parameterAsVectorLayer(
        parameters,
        self.INPUT,
        context
    )
    input_field = self.parameterAsString(
            parameters,
            self.INPUT_FIELD,
            context)
    new_name = self.parameterAsString(
            parameters,
            self.NEW_FIELD_NAME,
            context)

    if source is None:
        raise QgsProcessingException(self.invalidSourceError(parameters, self.INPUT))

    # Rename the field
    source.startEditing()
    source.renameAttribute(source.fields().indexFromName(input_field), new_name)
    source.commitChanges()

    return {'MODIFIED_LAYER': source}

The last line does appear to return a Layer reference:

Execution completed in 0.07 seconds
Results:
{'MODIFIED_LAYER': <qgis._core.QgsVectorLayer object at 0x0000022D82850E58>}

However, when I use the script in a Model, it allows me to pipe the output into another algorithm, but at runtime, it does not work.

Input Parameters: # for 'Basic statistics for fields'
{ FIELD_NAME: 'placetype', INPUT_LAYER: None }

I am not sure what approach I should use to produce output which can then be used in the Modeler

Answer 1

I think you have done the script properly. That means, there are inputs and there is also an output (that's why you can chain your script in the modeler).

I guess you are only missing one detail. In the last algorithm (Basic statistics for fields), you need to give the output a name when configuring it in the modeler.

Once you do so, you'll see a green box representing the result.

Running the model now allows you to have the results parameter, which is actually the location where you want to store the statistics. You can leave it empty and run the model with your other parameters entered. You'll get something like this:

Which gives you a temporal location for the statistics.

---

For reference, here are the files I've used:

• Rename Field Script.
• Model with rename field script.