来自边缘列表的python图形

Question

I have a list of edges in a text file:

0 1
0 2
0 3
1 637
1 754
1 1319 
1 1350
1 1463
1 1523
2 637
2 754
2 1319
2 1350
2 1463
2 1523
3 499
3 539
3 595
3 637
3 706
3 1128
3 1194
3 1213
3 1319
.. ...

I need to turn it into a dictionary like this:

graph = { "a" : ["c"],
      "b" : ["c", "e"],
      "c" : ["a", "b", "d", "e"],
      "d" : ["c"],
      "e" : ["c", "b"],
      "f" : []
    }

my attempt so far has been:

import numpy as np
file_name='/Volumes/City_University/data_mining/Ecoli.txt'
key_column=0

dat=np.genfromtxt(file_name,dtype=str)
d={i:[] for i in np.unique(dat[:,key_column])}

for row in dat:
    for key in d.keys():
        if row[key_column]==key :d[key].append(row[1])

print (d)

However, this does not work properly inasmuch i don't get a new key when this appear in the values: as an example I get:

'0': ["1", "2", "3"]
'1': ['637', '754', '1319', '1350', '1463', '1523']

in the '1', the "0" is missing.

to make it more simple. If I have a text like this

a b
c d

I should get an outcome like: graph = { "a": ["b"], "b": ["a"], "c": ["d"], "d": ["c"]}

thank you in advance

Answer 1

If you want a bidirectional graph, you need two appends.

Also, you don't really need the for key in d.keys() loop, just append to d[row[0]] instead of d[key] .

for row in dat:
    d[row[0]].append(row[1])
    d[row[1]].append(row[0])

---

Also, consider using a defaultdict, in which case you wouldn't need to initialize d with np.unique . It will also guard against errors that would otherwise occur when a node only appears in the second column.

import numpy as np
from collections import defaultdict

file_name='/Volumes/City_University/data_mining/Ecoli.txt'
dat=np.genfromtxt(file_name,dtype=str)
d=defaultdict(list)

for row in dat:
    d[row[0]].append(row[1])
    d[row[1]].append(row[0])

print (d)