Dispenser

class pyLambdaFlows.dispenser.Dispenser

This class will explain data dependancies over two layers.

For instance, if you have 3 lambda instance incoming, pylambdaflow will use a dispenser class in order create the dependency graph.

distribute(3) -> [[0,1], [2]]

In this example the dependency graph will be like :

X1----N1
      |
X2----+

X3----N2

with Xn the previous op and N the new one.

__call__(size)

Call self as a function.

distribute(size)

Function computing the next layer dependancies.

Parameters:size (int) – Previous instance size.

class pyLambdaFlows.dispenser.DHardReduce

Reduce dispenser that produce only one instance with all inputs.

distribute(size)

Function computing the next layer dependancies.

Parameters:size (int) – Previous instance size.

class pyLambdaFlows.dispenser.DMap

Map dispenser that create one lambda per previous instance.

distribute(size)

Function computing the next layer dependancies.

Parameters:size (int) – Previous instance size.