Add base path in AnonymizationConfig

[pounard]

We need to document and implement a stable way to resolve relative file names, for later implementation of data source files using csv or plain text files: the user needs to have a predictable and stable way of identifying its files.

This will be required for #228 implementation to be completed.

First step is to add the internal logic and documentation logic for this, without any usage of it.

[pounard]

Closing this, I decided to include the documentation update directly in #228 and remove the whole configuration changes, it's too much hard work for something that's not quite right.