Expose and rename properties for accessing underlying graphs and indices

[SimonHeybrock]

This will be required for a variety of functionality that we plan to add in Sciline or elsewhere. A simple example is #83:

import networkx as nx
import pandas as pd

g = workflow.networkx_graph
nodes = [node for node in g.nodes() if len(g.in_edges(node)) == 0]
nodes = [node for node in nodes if 'provider' not in g.nodes[node]]

not_set = '(not set)'
df = pd.DataFrame(
    {'node': nodes, 'value': [g.nodes[node].get('value', not_set) for node in nodes]}
)
df

I don't know if this (without the Pandas bit) is something that should be added as a helper function in Sciline, or more specifically in our applications, e.g., in ESSreduce?

[jl-wynen]

Do we need this property? I don't see where it is used. And it adds one more concept to the interface. Users need to know what is the difference between all the different graphs.