Add detection for legacy peak filter IDs in process methods

[Mailaender]

The IDs for process methods use a hardcoded ID based on class name with inconsistent prefixes/suffixes to avoid collisions, which seems like a bad idea in the first place. In https://github.com/eclipse-chemclipse/chemclipse/pull/1093/ I accidentally changed the IDs of most peak filters with a simple namespace change. This adds a method to auto-detect and migrate those previous IDs, in the same vein as https://github.com/eclipse-chemclipse/chemclipse/issues/1891 did for the settings content. It also allows us to clean them up with less volatile IDs in the future.