TwoRavens Machine Learning Interface
Generally, our TwoRavens Machine Learning Interface consists of two main components: Front-end and Back-end. Each contains several web applications.
This documentation provides a simple yet complete tutorial for the user and a detailed API introduction for the programmer.
Our TwoRavens Machine Learning interface is built upon Django and we assume the user know the basic of Django, please refer to Django for more information about this framework.
Table of Contents
- Tutorial and User Guide
- API Documentation (Web Application)
The TwoRavens for Machine Learning application has been built as part of the Data Driven Discovery of Models (D3M) program. Please refer to this Gitlab for more information about the API and the json schema we use for communication.
Above is the architecture of our D3M Interface. The function of each component is listed below:
- Content Pages : Render the web page presented to user;
- User Workspace : Render the user workspace in each web page;
- D3M Configuration : (Core Component) Contains all the information of a D3M task;
- R Service : Handle the dataset pre-process using R. (We are using flask as wrapper now). Notably, this component is only used by D3M Configuration, thus it can be considered as part of it.
- TA2 Interface : (Core Component) Handle the communication with remote TA2 Server;
- Datamart : (Core Component) Handle the communication with remote data repo and provide dataset filtering;
- Auth : Handles the user login/logout procedure;
- Admin: Default application provided by Django framework, provides an user-friendly control panel for the webpage;