Algorithm Development Kit (ADK), Java edition

package com.algorithmia.algorithm;

import com.algorithmia.development.*;
import com.algorithmia.*;

import java.util.HashMap;


// This class defines your algorithm.
class Algorithm extends AbstractAlgorithm<Algorithm.ExampleInput, String>{

    // This class defines the input to your algorithm, the algorithmia platform will attempt to deserialize JSON into this type.

    class ExampleInput {
        //If you flag a field with the @Required annotation, we will only validate the deserialization operation if the field is present.
        @Required String first_name;
        //If the @Required annotation is not present, then we will use the default / null value for that type if the field isn't present, consider it "optional".
        String last_name;
        ExampleInput(String first_name, String last_name){
            this.first_name = first_name;
            this.last_name = last_name;
        }
    }


    // This apply function defines the primary motive driver of your algorithm. Please ensure that the types defined in
    // your algorithm are the same as those defined in as generic variables in your concrete class defined above.

    public String apply(ExampleInput input){
        if(input.last_name != null){
            return "hello " + input.first_name + " " + input.last_name;
        }
        else {
            return "hello " + input.first_name;
        }
    }


    public static void main(String[] args) {
        Algorithm algorithm = new Algorithm();
        ADK algo = new ADK<>(algorithm);
        algo.init("world!");
    }
}

Focus

This document will describe the following:

• What is an Algorithm Development Kit
• Changes to Algorithm development
• Example workflows you can use to create your own Algorithms.

What is an Algorithm Development Kit

An Algorithm Development Kit is a package that contains all of the necessary components to convert a regular application into one that can be executed and run on Algorithmia. To do that, an ADK must be able to communicate with langserver. To keep things simple, an ADK exposes some optional functions, along with an apply function that acts as the explicit entrypoint into your algorithm. Along with those basics, the ADK also exposes the ability to execute your algorithm locally, without langserver; which enables better debuggability.

This kit, when implemented by an algorithm developer - enables an easy way to get started with your project, along with well defined hooks to integrate with an existing project.

Changes to Algorithm Development

Algorithm development does change with this introduction:

• An additional import (import com.algorithmia.development.*;)
• An AbstractAlgorithm class to extend, which contains two functions to implement, apply and (optionally) load.
• Finally, your algorithm is now executed directly instead of being a library, make sure to add the following at the bottom of your Algorithm.py file.

  public static void main(String[] args) {
  Algorithm algorithm = new Algorithm();
  ADK algo = new ADK<>(algorithm);
  algo.init();
  }

  • Now your algorithm is Executable, rather than just a library
    • Which will interact with the langserver service on Algorithmia
    • But is debuggable via stdin/stdout when executed locally / outside of an Algorithm container
      • When a payload is provided to init(), that payload will be directly provided to your algorithm when executed locally, bypassing stdin parsing and simplifying debugging!
    • This includes being able to step through your algorithm code in your IDE of choice! Just execute your src/Algorithm.java script and try stepping through your code with your favorite IDE

Example workflows

Check out these examples to help you get started:

hello world example

package com.algorithmia.algorithm;

import com.algorithmia.development.*;
import com.algorithmia.*;

import java.util.HashMap;


// This class defines your algorithm.
class Algorithm extends AbstractAlgorithm<Algorithm.ExampleInput, String>{

    // This class defines the input to your algorithm, the algorithmia platform will attempt to deserialize JSON into this type.

    class ExampleInput {
        //If you flag a field with the @Required annotation, we will only validate the deserialization operation if the field is present.
        @Required String first_name;
        //If the @Required annotation is not present, then we will use the default / null value for that type if the field isn't present, consider it "optional".
        String last_name;
        ExampleInput(String first_name, String last_name){
            this.first_name = first_name;
            this.last_name = last_name;
        }
    }


    // This apply function defines the primary motive driver of your algorithm. Please ensure that the types defined in
    // your algorithm are the same as those defined in as generic variables in your concrete class defined above.

    public String apply(ExampleInput input){
        if(input.last_name != null){
            return "hello " + input.first_name + " " + input.last_name;
        }
        else {
            return "hello " + input.first_name;
        }
    }


    public static void main(String[] args) {
        Algorithm algorithm = new Algorithm();
        ADK algo = new ADK<>(algorithm);
        algo.init("world!");
    }
}

Readme publishing

To compile the template readme, please check out embedme utility and run the following:

npm install -g npx
npx embedme --stdout README_template.md > README.md

To publish a new version

First, deploy locally on the Algorithmia Deep Purple worker using the following: sbt publishM2 then, load into an environment and validate functionality using the Environment Validator After tests are complete, merge PR into master and let the automatic CI/CD system deploy the new version to maven.

and you're done :)