Reputation: 1315
Saving plots in igraph as eps format.
I need save a plots in igraph as image, but eps format.
I have this:
def _plot(g, membership=None):
visual_style = {}
visual_style["vertex_size"] = 24
visual_style["layout"] = g.layout("kk")
visual_style["bbox"] = (400, 300)
visual_style["margin"] = 20
for vertex in g.vs():
vertex["label"] = vertex.index + 1
if membership is not None:
for vertex in g.vs():
if(membership[vertex.index] == 0):
vertex["color"] = "gray"
else:
vertex["color"] = "white"
if(membership[vertex.index] == 0):
vertex["shape"] = "circle"
else:
vertex["shape"] = "rectangle"
visual_style["vertex_color"] = g.vs["color"]
visual_style["vertex_shape"] = g.vs["shape"]
visual_style["vertex_label_color"] = "black"
visual_style["vertex_label_dist"] = -0.4
igraph.plot(g, **visual_style)
if __name__ == "__main__":
g = igraph.Nexus.get("karate")
cl = g.community_fastgreedy()
membership = cl.as_clustering(2).membership
_plot(g, membership)
I try this:
(1)
igraph.plot(g,'graph.eps',**visual_style)
(2)
import matplotlib.pyplot as plt
igraph.plot(g,'graph.eps',**visual_style)
plt.savefig('graph.eps', format='eps', dpi=1000)
But (1) and (2) not work, can someone help me?
Upvotes: 3
Views: 2687
Reputation: 48071
EPS is not very different from PS; basically, EPS (encapsulated PostScript) files are just PS files with a few additional restrictions. One such restriction is that the EPS file must contain a BoundingBox DSC comment that specifies the "bounds" of the drawing. You can usually convert a PS file to an EPS file with an appropriate converter that "infers" the bounds and places the additional comment in the file. So, I would first try plotting to PS and then convert it to EPS using ps2eps
. An alternative is to produce an SVG drawing and then ask an SVG-aware tool (such as Inkscape) to convert it to EPS. Inkscape even has command line arguments that allow you to do this in a batch script.
Upvotes: 1
Reputation: 23530
Unfortunately, igraph
's plot
only supports PDF, PNG, SVG, and PS. So that is why your method #1 fails.
As what comes to your method #2, there is nothing to draw the igraph
graph onto the matplotlib
drawing areas. It is, however, possible to draw an igraph
graph onto a Cairo SVG canvas created by matplotlib
(the renderer context needs to be given as an argument to plot
).
The latter trick is outlined here: https://lists.nongnu.org/archive/html/igraph-help/2010-06/msg00098.html
So, it is possible while not very straightforward. Another question is if you could avoid the use of EPS, it is usually possible with modern tools.
Upvotes: 2