Creating a UDF Using Annotations

Development Process

The following figure shows the process of developing a UDF using annotations.

Creating a UDF

The following uses a function that implements the add function as an example to describe how to create a UDF using the dataflow.pyflow and dataflow.method annotations.

import dataflow as df

# Use pyflow annotations to decorate a common function and build the common function as a UDF.
@df.pyflow()
def add(a, b):
    return a + b

# Decorate the class with the pyflow annotation and its methods with the method annotation to construct the class methods as UDFs
@df.pyflow
class Add():
    @df.method()
    def add(self, a, b):
        return a + b

Constructing a FlowGraph

After UDFs are constructed using annotations, use the fnode method to construct FlowNodes and FlowGraphs. The following is an example:

# Define inputs.
data0 = df.FlowData()
data1 = df.FlowData()

# Define FlowNodes.
flow_node0 = add.fnode()
flow_node1 = Add.fnode()

# Construct edge connections.
flow_node0_out = flow_node0(data0, data1)
flow_node1_out = flow_node1.add(flow_node0_out, data1)

# Construct FlowGraphs.
dag = df.FlowGraph([flow_node1_out])

Parent topic: UDF Development Process