Transformation of UML State Machine Diagram into Graph Database to Generate Test Cases

Abstract: The manual approach in software testing is considered as expensive, error-prone, and time-consuming activity since it depends highly on the test engineers. As well, the process of software testing requires proper planning and resources to design the test cases. For this reason, any approach that can be used to enhance or automate the current testing process is necessary. This thesis introduces an approach to transform the Unified Modeling Language (UML) behavioral state machines diagram into a graph database inside Neo4j, a graph database software. Moreover, a framework is proposed that fetches test data from the graph database. Based on the similarity between the state machine notation and the nodes and edges (with properties) in graph databases, a set ofrules for representation is presented in this thesis. Along with a framework based on GRANDstack (full-stack framework), the framework should use the pre-built graph database together with other technologies to generate test cases from the inserted requirements specification. A proof-of-concept is implemented to demonstrate the proposed framework. By using a dedicated schema, the fetched data is matched with the expected results. The results prove that the transformation method and the proposed framework have a good potential to be developed and evaluated with a realistic test from the industry practice. Keywords: requirement, test data, UML, state machine diagram, black-box testing, test case generation, Neo4j, GraphQL API