Design Specification

Proposed API for Reporting from Di/Multi/Graph

• nodes() - iterator over nodes
• neighbors(n, data=None) - iterator over neighbors of n.
  • If data is True: iterator over (nbr, eattr_dict)
  • if data not in (None,True): iterator over (nbr, eattr_dict[data])
• degree(nbunch=None, weight=None) - iterator over (n, deg(n))
  • if nbunch is not None: only iterate over nodes in nbunch… single node->single value
  • if weight is not None: deg(n) is sum of weights on edges instead of number of edges
• edges(nbunch=None, data=None, keys=None) - iterator of 2-tuples (undirected only once)
  • if nbunch is not None: iterator of edges from nodes in nbunch (undirected only once)
  • if data is True: iterator of edge info (n,nbr,eattr_dict)
  • if data not in (None,True): 3-tuple (n,nbr,eattr_dict[data]
  • MultiGraph
    • if data/key is True/None: iterator of edge info (n,nbr,keydict)
    • if data/key is None/True: iterator of 3-tuple edge (n,nbr,key)
    • if data/key is True/True: iterator of 4-tuple edge (n,nbr,key,eattr_dict)
    • if data not in (True,None): 4-tuple (n,nbr,key,eattr[data])
• get_edge_data(u, v, data=None, key=None) eattr_dict for edge between u,v
  • if data is not None: eattr[data] for edge between u,v
  • MultiGraph
    • if key is None, return eattr_keydict, otherwise look at edge u,v,key
• adjacency_iter(data=None) - iterator of (n, {nbr:eattr, ...}) This is dict-of-dict... better way??
  • if data is not None: eattr replaced by eattr[data]