marcelengelmann / conversation-engine Goto Github PK

A dialog management system framework

Report Bug · Request Feature

Conversation-Engine is a framework that is supposed to be used as a dialog management system. It is designed to simplify the creation of chatbots by combining skills and a single NLP-Component} into an all in one system.
The system is based on a finite state machine to create an accessible testing environment.

To get a local copy up and running follow these steps.

To install Java you can choose one of the following options

• OpenJDK

  Please visit https://openjdk.java.net/install/ for installation instructions

• Oracle

  Please visit https://docs.oracle.com/en/java/javase/13/install/overview-jdk-installation.html for installation instructions

• Please visit https://maven.apache.org/install.html for installation instructions

1. Clone the Conversation-Engine Project

   git clone https://github.com/death-truction/Conversation-Engine.git

2. Change the directory to the project folder

   cd Conversation-Engine

3. Install the package to your local maven repository (replace the version number with the latest release)

   mvn clean install

The project includes a playground to try the Conversation-Engine with some example skills.

1. To start the playground simple run the following command in your cloned github repository:

   mvn exec:java

2. These languages are currently supported for the playground:
   • German
   • English
3. For a list of supported request checkout
   • Current Skills German
   • Current Skills English

1. To use the Conversation-Engine Framework simply include the installed dependency in your pom (replace the version number with the latest release)

   ...
     <dependencies>
       ...
       <dependency>
         <groupId>de.dai-labor</groupId>
         <artifactId>conversation-engine</artifactId>
         <version>1.0.0</version>
       </dependency>
       ...
     </dependencies>
     ...

2. Implement the interfaces found in the interfaces package at one of the following locations:
   • github
   • javadocs
   • For an implementation example checkout these test classes
     • NLPComponent
     • NLPAnswer
     • Skill
     • SkillAnswer
3. Create a valid state machine JSON-File
   • by using the Dialog Modeling Tool inside the project
     • You can run the tool by navigating inside the folder and running the following command AFTER installing the conversation engine:

     mvn javafx:run

   • Create your own JSON-File by following this JSON-Schema
4. Create a new Conversation-Engine object and add your NLPComponent and your skills

   NLPComponent nlp = new NLPComponent();
   ConversationEngine conversationEngine = new ConversationEngine(nlp);
   MySkill mySkill = new MySkill();
   String mySkillStateMachineJsonFile = loadJsonFileAsString("MySkill.json");
   stateMachine.addSkill(mySkill, mySkillStateMachineJsonFile);

5. Now you can simply send your inputs to the Conversation-Engine and get the answers in a List

   List<String> answers = conversationEngine.userInput("Hello world!");

The project includes two Logging-Facades named "DeveloperLogger" and "ConversationLogger"
To use the output of the loggers you have to add your own logging framework to the project and configure the loggers.
Checkout the tests of the project, which are using the logging dependency logback with a simple configuration.

• Overview
• Javadoc

See the open issues for a list of proposed features (and known issues).

Distributed under the MIT License. See License for more information.

• Best-README-Template