Development of Dream Builder

[oserikov]

difficulty: medium
duration: 350 hours
mentor: Maria, @mablin7 , @danielkornev , @dilyararimovna
requirements:

1. Fluency in TypeScript, Python
2. Good understanding of CSS
3. React + redux or similar state management library + FastAPI
4. SQL
5. Machine learning basics
   useful links: Spring Initializr Deepy demo

tl;dr: Dream Builder is a graphical tool for creating and configuring components of the DeepPavlov Dream AI Assistant platform. It is implemented as a web app with a React frontend and a Python backend.
The task is to embed our tool for creating scenario-driven skills (DF Designer) into Dream Builder.

Idea Description:

Dream Builder is a graphical tool for creating and configuring components of the Dream bot framework.

Dream bot framework

Our company develops Dream.
Dream is an open-source platform for creating multi-skill AI assistants.

• The first version was created in September 2019. It was actively developed during the Alexa Prize Challenge 3 and 4.
• The Dream dialog system has a modular structure.
• Dream contains dozens of annotators (NLU) and skills (NLG).
• Dream uses the open-source DeepPavlov Agent library for orchestration by various components.

Dream release article
DeepPavlov web-site
Dream demo

Dream Builder

Dream Builder is a graphical tool for creating and configuring components of the Dream bot framework. It is implemented as a web app with a React frontend and a python backend. The latter is responsible for generating configurations for the different Dream components based on user settings, as well as training and running these components for testing. The backend is mostly functional, with only minor adaptations being required to work well with Dream, but the frontend is in a prototype state: only two editors are functional, one of which is deprecated, plus basic navigation.

Status of the project at the beginning of the internship

By the beginning of the internship, the following functionality will be implemented in Dream Builder:
Skills and DP annotators can be added/removed in Dream Builder to form a custom distribution
Users will be able to “compile” their custom distribution, talk to it within the interface of the system, and make it up-n-running for the external users
Users will be able to develop scenario-driven skills via a VS Code extension called DF Designer

Tasks for the Internship

We are working on the development of DF Designer.
DF Designer is a tool for developing scenario-driven skills for multi-skill AI assistants and simple bots in a visual environment.
DF Designer includes basic skill creation capabilities and advanced functionality such as recommender systems in scenario-driven skills, statistical information, etc.
The task is to embed our tool for creating scenario-driven skills (DF Designer) into Dream Builder.

Coding Challenge:

Create a web based editor for the pipeline_conf.json config file. Docs