Empirical evaluation for JLAMP 2022

- Updated README.md
- New module for benchmarks (sdg-bench)
- Added representations for the original JSysDG and for the Allen&Horwitz SDG (graph and slicing algorithm)
- Fixed bugs:
  * Accept abstract methods (without a body)
  * Avoid generating constructors for interfaces
  * Handle calls outside methods
  * Handle methods inside methods
  * Locate and handle all method calls (including on the left-hand side of assignments and in return statements).
  * Correctly determining the type of -output- nodes
  * Various errors related to object tree location and traversal