Generates syntax trees from the given source and input using ANTLR, then measures and produces similarity scores using tree kernel functions.