通过python过滤出对称线

Question

I have a graphviz file automatically generated, which sometimes produces double connections, eg

"test textanalysis" -> "development" [color="white"];
"development" -> "test textanalysis" [color="white"];

I'd like to know

1) if it is possible to show this as a single connection with double arrows at start&end in Graphviz

2) or in alternative if you know a Python way to filter out one of those lines (I dont mind losing that information): I cannot find a regex able to do it!

Answer 1

You can try something like this to filter repeating edges:

import re

edges = {}

with open(dot_file) as fr:
    for line in fr:
        key = tuple(sorted(re.findall('"([a-z ]+)"', line)[:2]))
        edges.setdefault(key, []).append(line.strip())

for v in edges.values():
    if len(v) > 1:
        print re.sub("\[(.+)\]", '[\\1, dir="both"]', v[0])
    else:
        print v[0]

You can also try using concentrate=true option (See: Dot graph language - how to make bidirectional edges? ) for details.