Ensure that given the same input data and fixed random number generators, the results are actually the same (reproducible)

[1kastner]

What has been done
One idea is that ConFlowGen should generate reproducible results. The first half of the solution has been discussed and realised in #9 and #191. The second half came with #196. All random number generators need to be fixed, all SQL results need to be sorted, and any kind of shuffling needs to work with a (known) random seed.

What remains to be done
Thoroughly check whether all this works.