Development of Dialogue Tools with Automatic Assessment for MIEM Digital Environment

The article explores the questions of API implementation to provide access to the database and chat-bot implementation based on Zulip messenger for automatic assessment of student’s tasks. The first section consists of a comparative analysis of methods of natural language processing in the context of text normalization to create an effective tool as word transformation to the normal form is one of the main features to avoid spelling mistakes from the student. The second section describes libraries to process text, its functionality, advantages, disadvantages, usage for text normalization. The third section explains how the chat-bot was implemented to interact with students and check answers, describes the database design for automatic text assessment tasks, architecture of the project, features to implement, the process of interaction with the bot and the process of text normalization. The fourth section describes planning, implementation, and usage of the API to interact with the third users, the choice of REST architecture paradigm, and the use of Swagger user interface to test API handlers. The fifth section contains information about the public deployment of the chatbot and API on the virtual machine, containerization within Docker, features of preparing the Flask application for the production deployment, and configuration details on the server. The last section of the article explains the main user scenarios for working with the bot and API, application behavior under critical loads and security breach, and also minimum requirements for the machine where chat-bot and API will be deployed. The article consists of 65 pages. There are 25 illustrations and 3 tables in the paper. The bibliography consists of 36 sources.