Source code updated in SVN r18: Fixes and working towards similarity methods for protein ligand graphs

The source code has been updated from our internal repo and the repo here at SF is now at revision 18. Many fixes and improvements everywhere. Implementations of some classical algorithms from the field of bioinformatics (Needleman-Wunsch, Smith-Waterman) have been added. They are used in the new similarity methods that are intended to allow searching for graphs in the database which are similar to a given template graph in the future. Other major changes include code for spatial ordering of SSEs in graphs. Minor changes to DB layout, changed some default settings, etc.

Most of the new features (especially the new '--similar' command line option) are still very experimental.