CODEWITHSUNDEEP
pythonpython-2.7networkx
Nobi
Nobi

Reputation: 1109

Connecting opposite nodes

I have two line graphs of equal length and parallel to each other in networkx, how can i loop over the graphs and connect opposite nodes with two edges ie connect 2 & 2a, 3 & 3a etc 

nodes1 = [2,5,7,8,9,10]

nodes2 = ['2a','5a','7a','8a','9a','10a']

I tried 

for k in range (0, len(nodes1)):
    for i in range (0, len(nodes2)):
        if nodes1.index(kk) == nodes2.index(i):
            X.add_edge(k,i)

But it dosent give the desired output please can anyone correct me. Thanks

               This is what I have 
          2---------3---------4----------5----------6



          2a---------3a-------4a--------5a---------6a


                 This is what I want 
            2---------3---------4----------5-----------6
            ||        ||        ||         ||         ||
            ||        ||        ||         ||         ||
            ||        ||        ||         ||         ||
           2a--------3a---------4a---------5a---------6a

Sorry if the previous post was not clear

Upvotes: 1

Views: 97

Answers (2)

EdChum
EdChum

Reputation: 393883

The reason your code does not work as is, is because index returns the index in the list where the value resides, so you are looping from 0 to the length of each list and the first value 0 does not exist in your list which throws a ValueError. See the online docs: http://docs.python.org/2/tutorial/datastructures.html

If you just wanted to create the tuple pairs and assuming lengths and ordering were already correct then if you modify your code with this line:

for k in range (0, len(nodes1)):
    X.add_edge(nodes1[k],nodes2[k])

would achieve what you want as you want to traverse both lists in parallel, your original code even if index didn't raise a ValueError would have resulted in 5 x 5 entries as you were adding an edge for every entry in nodes2 for each node in node1 which is not what you wanted.

Assuming your ordering matches for both lists then you can do a list comprehension to create the edge list and create your graph from this

In [171]:

edge_list = [(nodes1[i], nodes2[i]) for i in range(len(nodes1))]
# I'm using python 3 so you could do the following also
# edge_list = list(zip(nodes1, nodes2))
G=nx.DiGraph()
G.add_edges_from(edge_list)
G.edges()
Out[171]:
[(2, '2a'), (5, '5a'), (7, '7a'), (8, '8a'), (9, '9a'), (10, '10a')]

EDIT

If you want to add edges in both directions then you just create a secondary edge_list like so by swapping the order of the lists around:

In [194]:

edge_list_reversed = list(zip(nodes2, nodes1))
edge_list_reversed
G.add_edges_from(edge_list_reversed)
G.edges()
Out[194]:
[(2, '2a'),
 (5, '5a'),
 (7, '7a'),
 (8, '8a'),
 (9, '9a'),
 (10, '10a'),
 ('2a', 2),
 ('8a', 8),
 ('9a', 9),
 ('5a', 5),
 ('10a', 10),
 ('7a', 7)]

Upvotes: 2

pbacterio
pbacterio

Reputation: 1152

Use zip() method

>>>> zip(nodes1, nodes2)
[(2, '2a'), (5, '5a'), (7, '7a'), (8, '8a'), (9, '9a'), (10, '10a')]

Upvotes: 0

Related Questions