Create a sequencer that processes Java source files and creates a representative graph of the structure of the code. This should include the types, fields, methods, and annotations (each with descriptions taken from the JavaDoc), but probably should not include any method implementation detail (maybe much later).

There are a couple of obvious approaches for parsing: Antlr 3 grammar, Eclipse JDT library, or JavaCC (discouraged).

The schema for the output should ideally follow as closely as possible the constructs in the Java specification.