Using temporary layer as input for other algorithm in Processing script

Question

My script has a single point input layer that will be processed with some logic to extract just some points.

I'd like to pass these extracted points as input of another algorithm (simple buffer) within the same scripts.

I define the final output as QgsProcessingParameterFeatureSink:

self.addParameter(QgsProcessingParameterFeatureSink(
    self.OUTPUT,
    "Extracted points",
    QgsProcessing.TypeVectorPoint
    )
)

....

   (sink, dest_id) = self.parameterAsSink(
        parameters,
        self.OUTPUT,
        context,
        source.fields(),
        source.wkbType(),
        source.sourceCrs()
    )

running some logic to extract some points:

features = source.getFeatures()
for current, feat in enumerate(features):
    if feat.geometry().asPoint() == maximum:
        sink.addFeature(feat, QgsFeatureSink.FastInsert)

then I'm trying to call the buffer algorithm:

   buffer = processing.run(
        "native:buffer", {
            'INPUT': sink,
            'DISTANCE': distance,
            'SEGMENTS': 5,
            'END_CAP_STYLE': 0,
            'JOIN_STYLE': 0,
            'MITER_LIMIT': 2,
            'DISSOLVE': False,
            'OUTPUT': 'memory:'
        }
    )

but it does not accept the input:

Incorrect parameter value for INPUT

Answer 1

There's two things you need to fix:

• Pass dest_id, not the sink as the parameter value, i.e.

param = { 'INPUT': dest_id, 'DISTANCE': distance, 'SEGMENTS':5, 'END_CAP_STYLE':0, 'JOIN_STYLE':0, 'MITER_LIMIT':2, 'DISSOLVE':False, 'OUTPUT':'memory:' }

• Pass the context and feedback objects to processing.run, so that the child algorithm can access project layers and other objects from the wider context, and can pass feedback messages from the child algorithm (and handle cancelation and progress reports):

buffer = processing.run("native:buffer", param, context=context, feedback=feedback)