Improve distribution of samples throughout the project

[pitag-ha]

Currently, the number of samples per file is given in absolute terms, e.g. the default is 30 samples per file. For a better distribution of samples throughout the project, it should be relative to the size of the file / the size of the AST.