Document example implementations of metadata extractors

[mih]

http://docs.datalad.org/projects/metalad/en/latest/user_guide/writing-extractors.html nicely documents the purpose of each API component. However, the provided example implementations (which people like me will C&P) are undocumented. This will lead to the duplication of work (everyone needs to add comments), or -- much more likely -- the proliferation of undocumented code.

I think the information from the docs should also be present in the example implementations.