def AgentPredictorWorker.transform_dual(self, inputs) [source]
Synchronized dual inference: train + test in one call.
Executes transform on both the training and test
sets, returning paired predictions. Used for pre-flight
validation before go-live deployment — ensures that
the scoring pipeline produces consistent results across
both data sources.
Parameters
inputs : AgentPredictorInput
Must contain both
(ML-ready data).
train_set and test_set(ML-ready data).
Returns
AgentPredictorResultDual
Paired predictions:
data (train) andtest_data (test), with full feature lineage.