Grapute

PM> Install-Package Grapute -Version 1.0.0

The library helps to define a graph of computations or a pipeline by connecting computation Nodes to eachother, where an output of one node becomes an input for the next Node.

Node always returns an array of results. Later these results can be either consumed one by one by another Node or all together by a SinkNode.

You can define your processing Node by inheriting from Node class and implementing process method. Another option is to provide your own process Func to FuncNode.

You can find a sample code that demonstrates how to implement a simple MapReduce algorithm using Grapute.

Example

var produceThreeOutputsNode = new FuncNode<int, int>(x =>
    {
        //
        return new[] { x , x + 1, x + 2 };
    });

var pipeline = produceThreeOutputsNode
    .ForEachOutput(x =>
        {
            // As an output we have a doubled input
            return new []{x, x};
        })
    .ForEachOutput(x =>
        {
            // Increase the input
            return new []{x + 1};
        })
    .CollectAllOutputsToOneArray()
    .ForArray(all =>
        {
            // Split the input array and convert to a String
            var p1 = all
                .Take(all.Length / 2)
                .Aggregate((a ,b) => $"{a}{b}");

            var p2 = all
                .Skip(all.Length / 2)
                .Aggregate((a ,b) => $"{a}{b}");

            return new []{p1, p2};
        });

produceThreeOutputsNode.SetInput(5);
var result = pipeline.Process().Output;

The code above represents the following graph of computations: