Passing selected layer-feature inside QgisAlgorithm (in code)

Question

In our algorithm, we do three steps:

1. call Algorithm A AlgoA,
2. takeResultLayer into lyr from AlgoA output,
3. for each feature of lyr we want to run Algorithm B AlgoB .

Code of my processAlgorithm looks like this:

 alg_params = {
 'Input': parameters['INPUT'],
 'native:intersection_1:OUT': 'memory:'
 }
 zta = processing.run('model:AlgoA', alg_params, context=context, feedback=feedback, is_child_algorithm=True)['native:intersection_1:OUT']
 
 lyr = context.takeResultLayer(zta) 
 features = lyr.getFeatures()
 
 for i, f in enumerate(features):
 # FeaturesIntersectionOperation
 alg_params = {
 'One_Feature': f,
 'native:intersection_1:out': 'memory:'
 }
 
 out = processing.run('model:AlgoB', alg_params, context=context, feedback=feedback, is_child_algorithm=True)['native:intersection_1:out']

The error I get is:

Unable to execute algorithm Could not load source layer for One_Feature: invalid value

If I pass an entire layer lyr as 'One_Feature': lyr it all runs Ok. If I use my algorithm from QGis UI with iterate over layer features selected like this: enter image description here it also runs Ok.

So how to implement for each layer feature inside QGIS 3+ processAlgorithm script?

Answer 1

Instead of

 for i, f in enumerate(features):
 # FeaturesIntersectionOperation
 alg_params = {
 'One_Feature': f,
 'native:intersection_1:out': 'memory:'
 }
 
 out = processing.run('model:AlgoB', alg_params, context=context, feedback=feedback, is_child_algorithm=True)['native:intersection_1:out']

do

 # List features ids from layer
 ids = [f.id() for f in features]
 # Loop to select by id the layer
 for id in ids:
 lyr.selectByIds([id])
 # Use the selected layer with the following (see https://qgis.org/pyqgis/master/core/QgsProcessingFeatureSourceDefinition.html#qgis.core.QgsProcessingFeatureSourceDefinition for params)
 sourceDef = QgsProcessingFeatureSourceDefinition(lyr.id(), True)
 # FeaturesIntersectionOperation
 alg_params = {
 'One_Feature': sourceDef,
 'native:intersection_1:out': 'memory:'
 }
 
 out = processing.run('model:AlgoB', alg_params, context=context, feedback=feedback, is_child_algorithm=True)['native:intersection_1:out']

• DuckQueen: Sadly getting same Could not load source layer for One_Feature: output_32b4b3b6_f977_4ffd_9d19_5e366c1e0c41 not found error
• ThomasG77: How did you define your model:AlgoB? In particular what does processing.algorithmHelp("model:AlgoB") returns? Depending of your algo, our approach could be wrong (for both of us)