Edgelist Returned from dataset.get_edgelist() is global

[alexbarghi-nv]

Currently, when calling get_edgelist(), a global edgelist is returned. This causes issues when performing operations such as:

df_unweighted = datasets.karate.get_edgelist()
df.drop('wgt', axis=1, inplace=True)
...
df_weighted = datasets.karate.get_edgelist()
print(df_weighted)

Expected:

src  dst  wgt
...

Actual:

src  dst
...

To fix this issue, datasets should return a copy of the global edgelist so the edgelist can be used in multiple places without affecting the original object.

[rlratzel]

We can also address this with documentation stating that a reference is being returned and that the caller should make a copy if needed. This makes it more of a documentation bug.