Dynamic Graph Supported Object Mapping in Python

Abstract

A wide variety of tools and data models are used in the aerospace industry to design and simulate spacecrafts and rockets. Moving data between these tools is often done by hand, which is prone to errors and time consuming. Due to the complexity of the applications used, traditional object mappers are not suitable for this task. This thesis presents a new concept of moving data between applications and data models. Graph theory and graph rewriting systems are used to map between applications and data models by using pattern detection instead of static mapping code. The concept is implemented in Python and its performance and usability is tested using profiling tools and user tests. The tests show that the concept of using graph rewriting can be used to map objects between different applications and data models, although the performance of graph rewriting libraries has to be improved to be suitable for mapping large applications. A further limitation is the inability to create conditional rules, which makes it impossible to map based on the context of an object outside of its type and name. These limitations can be eliminated by extending existing graph rewriting libraries to allow conditional rules and by implementing more efficient pattern matching and rule application algorithms.