Ing. Josef Vogel, CSc.

Ing. Josef Vogel, CSc.

Ing. Pavel Náplava, Ph.D.

Department of Software Engineering

Ing. Josef Vogel, CSc.

MUDr. Jan Wolf

Department of Software Engineering

Ing. Josef Vogel, CSc.

Ing. Tomáš Zahradnický, Ph.D.

Department of Software Engineering

In this bachelor thesis I am solving the design and development of a system for searching and aggregation of log files and generating events based on the data found. In the first part is described the architecture of this system, examples of use cases, requirements for this system, description of the system users and the selected technologies. Based on the results of the first part, the second part describes the process of implementation of the individual components of the system, the interface of their communication and a description of the search and processing of the results of the indexing tool. Developed components are the RESTful Java API that handles client application requests, creates search objects for an indexing tool, and generates events based on the data found. In addition, a client application was developed to interact with the user and process the results returned by the system's RESTful API. System RESTful API was implemented in Java, the client application in AngularJS 2. Elasticsearch was chosen as the indexing tool. Proces of the deployment of the system to the production or testing environment is described in Chapter 4. In this chapter is described how to containerize each component of the system with the Docker and link their communications. In the end, the future of the system and other possible extensions are discussed.

Thesis on DSpace

Ing. Josef Vogel, CSc.

Ing. Ivan Halaška

Department of Software Engineering

Ing. Josef Vogel, CSc.

Ing. Kamil Foltin

Department of Software Engineering

The goal of this thesis is to create an interactive web application. The application analysis and visualizes patients data. These patients undergo allogeneic hematopoiesis transplantation. Visualizations and analysis are divided into 3 different categories and each category follows up different problems. The thesis describes in detail, how individual statistical elements used in the application work. The application provides a user interface and server logic. Both of these are written in programming language R with help of its framework Shiny. Processed data are being stored in MySQL database in ÚHKT's local network. This application will serve ÚHKT staff for quality monitoring and it will be also used for scientific purposes.

Thesis on DSpace

Ing. Josef Vogel, CSc.

Department of Software Engineering

Ing. Josef Vogel, CSc.

Ing. Josef Kokeš

Department of Software Engineering

This master thesis deals with the design and development of a system for automated deployment of neural networks. The aim of this work is to create a system that train a neural network model based on user-submitted data, deploys it automatically to the Kubernetes cluster and build an API for communication with the user. In order to allow the user to add additional functionality to the neural network models which already created, the system offers the possibility to create user modules whose source code is entered by the user within the front end application. Those modules are linked to the neural network model through classification classes. The first part is focused on the description of neural networks. The second part describes the architecture of this system, examples of use, requirements for this system, description of system users and selected methods of machine learning. Based on the results from the second part, the third part describes the process of implementation of the individual components of the system, the interface of their communication and the description of the implementation of measurement of selected neural networks. The developed components are REST Kotlin API, which handles client application requests, communicates with the Kubernetes cluster and also handles the developed REST Kotlin API, which creates Docker images from the user-specified code. In addition, a client application was developed to interact with the user and process the returned REST API results, and three models of neural networks for image processing, log classification, and written test processing for information chatbots. The entire system has been tested and a description of the testing process and the selected test method, both automatic and manual testing, are described in Chapter Four. The process of deploying the system to the production or test environment is described in Chapter 5. In the end, the future of the system and other possible extensions are discussed.

Thesis on DSpace