Migration Issues Random Walks

[M-Lampert]

As far as I am aware, we reused the PathPy3 code for the implementation of the random walks. I noticed the following problems that need to be fixed:

• Docstrings in the wrong format (fixed in #189)
• Bug in transition_probabilities(...) (fixed in #189)
• Documentation out of date. E.g. the example uses rw.plot(...) which does not work in pathpyG as far as I know.
• Bug in HigherOrderRandomWalk.first_order_stationary_state(...): v.relations[-1] throws an exception since each node is an id (tuple or string) and not a node object as in pathpy3.
• TODOs in process.py: More than half of the file is commented out and marked as TODO.

There are probably some more problems that I haven't noticed so far, so I propose some additional unit-tests for each file in processes/:

• Tests for process.py
• Additional tests for random_walk.py
• Tests for sampling.py