Haha, apologies for screwing up the review process. To continue from the
previous posts, if you first re-order the embedding to actually be clockwise,
then the trace_faces() method works as expected:
import numpy
def reorder_embedding(emb, locs):
new_emb = {}
for i,neighbors in
On Wednesday, March 26, 2014 5:34:53 PM UTC-6, Christa Brelsford wrote:
for a simple graph, trace_faces() gives the expected answer for the faces of
a planar graph, as shown below.
import networkx as nx
lat = nx.Graph()
lat.add_edge(1,2)
lat.add_edge(2,3)
lat.add_edge(2,5)
On Wednesday, March 26, 2014 5:34:53 PM UTC-6, Christa Brelsford wrote:
for a simple graph, trace_faces() gives the expected answer for the faces
of a planar graph, as shown below.
import networkx as nx
lat = nx.Graph()
lat.add_edge(1,2)
lat.add_edge(2,3)
lat.add_edge(2,5)
The problem is not in the trace_faces() method, but in the is_planar()
calculation. The embedding of the second graph is not correct:
S = Graph(lat)
S.show(vertex_size = 600, pos = nodes_dict)
S.is_planar(set_embedding = True)
s_emb = S.get_embedding()
print s_emb
{1: [2], 2: [1, 3, 5], 3: [4,
The problem is not in the trace_faces() method, but in the is_planar()
calculation. The embedding of the second graph is not correct:
The is_planar method does not take the nodes_dict dictionary as an
input. Why do you say that its output is incorrect ?
Nathann
--
You received this message
Oh, I forgot : it is indeed weird that faces may not exist in your
version of Sage even though trace_faces is already deprecated,but
everything seems to be fine in Sage 6.2.beta5. Sorry for that, I don't
know where it comes from, but it will be fixed if you update your
install :-)
Nathann
On 28
Hello !
for a simple graph, trace_faces() gives the expected answer for the faces
of a planar graph
That's a good news :-D
the face that should exist between just nodes 3,4 and 5 is not found, it's
replaced with a face around nodes 1,2,3,4,5. Any ideas what is wrong, or
Nathann,
Thanks for your help! I'm fairly new to both Sage and graph theory, but I
understand the difference you point out, and it looks like the trace faces
function is giving me accurate faces for some valid planar embedding- just
not the one I thought it was working on. I spoke with a
Christa,
The problem is not with the code, but your expectations of it (which
may be valid, but that would be a feature request and not a bug). You
expect the code to look at your planar position dictionary, and gin up
an embedding from that. That is not a bad idea, and possibly a good
feature