| Abstract [eng] |
This thesis presents a test case generation methodology designed to automate the creation of manual test cases from UML use case diagrams and their specifications. Although manual test case creation is still widely used in software testing processes, this activity requires significant time and human resources and largely depends on the tester’s experience and interpretation. As a result, test case creation becomes increasingly complex in projects that involve a large number of use cases and scenarios. An analysis of software testing, UML use case diagrams, and existing test case generation solutions revealed that most existing approaches use other types of UML diagrams, such as activity or sequence diagrams, for test case generation, while use case diagrams are typically applied only as an additional source of information. It was also determined that existing solutions do not evaluate the content of use case specification tables and the scenarios described within them, and therefore are unable to generate detailed manual test cases based solely on UML use case diagrams and their specifications. To address these issues, the aim of the thesis was defined as facilitating the software testing process by automating test case creation based on UML use case diagrams and their specifications. To achieve this goal, a test case generation methodology was designed, defining the principles of UML use case diagram and specification analysis, scenario identification, and test case generation. To implement the methodology, a tool was developed that analyzes the provided UML use case diagram and use case specification tables, identifies possible testing scenarios, and generates manual test cases in Microsoft Excel format. The tool was implemented using the Python programming language, the Tkinter graphical user interface framework, the python-docx library for reading data from Word documents, and the openpyxl library for exporting generated test cases. The generative artificial intelligence model OpenAI GPT-4.1-mini was applied for use case analysis and feature identification. An experimental study consisting of two parts was conducted to evaluate the developed solution. In the first part, data from three different real-world projects obtained from an IT company were used, and the generated test cases were compared with manually created test cases. In the second part, a survey of IT specialists was carried out to evaluate the clarity, usefulness, and practical applicability of the tool. The results of the experimental study demonstrated that the developed solution can be applied to facilitate the test case creation process in projects that use standardized UML use case specifications. |
