How ReflectCode works ?

ReflectCode transformation system internally uses various compiler components and techniques such as pre-processor, scanner, tokenizer, and syntax analyser along with Artificial Intelligence (AI) and Machine Learning (ML) technologies.

ML technology gives ReflectCode an unsupervised self learning capability, which is used to build the Knowledge Base (KB) for entire source code. KB is an organized collection of meta data about all classes, their relations and class members in a source project. This project specific KB is built for each project and discarded once the transformation is complete. The transformation engine also contains platform SDK knowledge base which contains information about both Android / iOS platforms and 3rd party libraries.

With both knowledge base ready, engine then invokes the next part i.e. ReflectCode Expert System (RES).
RES is our Artificial Intelligence core, which performs the actual transformation using its extensive inference rules and pattern matching algorithms. RES contains different set of rules for xml files and Java/Kotlin source code. The end result is –
    1) Android layout transformed into pixel perfect iOS storyboard
    2) Java/Kotlin code transformed into semantically equivalent Swift code
    3) Android assets transformed into Xcode asset catalog

RES also contains special rules to transform the features which are not available in iOS SDK and in few cases it also uses wrapper classes.
While all this happens magically, it still have boundaries because of its limited inference rules and KB. If it encounters any code which is not present in its rules, then it applies generic rules and you may get compiler error for this.

For porting digital assets we use open source tools such as ImageMagick and Inkscape.
 
Our aim is to make sure that migrated code is developer friendly iOS Xcode project and uses of actual iOS classes. RC preserves the signature of user defined classes and variables declared in your project making the generated code a true reflection of source code.

Seems unbelievable? Have a look at our sample projects on GitHub.