A lightweight framework for automated Model-to-Code transformation

Model Driven Engineering (MDE) focuses on the automatic generation of code from system models, referred to as the Model-to-Code (M2C) transformation. Currently, M2C transformation relies on the transformation rules, metamodels and transformation languages, making the transformation process complex and lengthy. In this paper, we propose a novel framework for M2C transformation utilizing the available set of transformation examples. The examples are expressed as predicates and are used to train the system regarding automatic code generation. After the training, a set of transformations are extracted from the available training data, which are eventually used to transform the input models into implementation code. The quality of the generated code is ensured by the correctness of the available transformation examples. This framework uses transformation examples to utilize the transformation knowledge on hand and evades the definition of transformation rule set from the M2C transformation process.