Functional Testing of Graph Query Frameworks

OData support
Semeráth Oszkár
Department of Measurement and Information Systems

Nowadays model-driven development is used to design fault-tolerant systems where documentation, system configuration or even complete source code of components is derived from engineering models by using model transformations to improve the efficiency of the development process. The most challenging task in model transformation frameworks is the graph pattern-matching. This means selecting elements from a satisfies a specific property.

However the error of Graph Query Frameworks are invalidating results of model transformations so functional testing and extensive validation of these transformations is one of the most important task. Graph Query Frameworks should be completely reliable because a singular error may invalidate the developed system thus restarting the certification process.

The aim of my thesis is to support the developement of the EMF IncQuery by providing a framework for functional testing. This is challenging work because (i) need to generate several various model and (ii) need to compare results of matches but before this we must mark the model elements by identification. (iii) It needs using of several technologies.

Through a case study I present the approach and working of the tesing, modelling and Graph Query Frameworks. I provide a solution for automatic generation of test models and test pattern. I discribe how it is possible to map EMF-IncQuery patterns to the input of the Alloy SAT-Solver. Moreover I represent which method can be used for comparing the two query results. I describe the technical details of my framework as well and I make measurements in connection with the efficiency of my solution.


Please sign in to download the files of this thesis.