Test for to_networkx conversion fails when enabled

## 🐛 Bug

The [existing test](https://github.com/dmlc/dgl/blob/master/tests/python/common/function/test_basics.py#L202) for the `to_networkx` conversion is disabled. When it is enabled, it fails with `dgl._ffi.base.DGLError: Expect edge_id_attr_name and edge_attrs to be None when nx_graph is undirected, got None and ['h']`.

## To Reproduce

Smallest code snippet that reproduces the behaviour:

```
nxg = nx.cycle_graph(5)
nxg.remove_nodes_from([0, 4])
for u in nxg.nodes():
nxg.nodes[u]["h"] = F.tensor([u])
for u, v, d in nxg.edges(data=True):
d["h"] = F.tensor([u, v])

g = dgl.from_networkx(nxg, node_attrs=["h"], edge_attrs=["h"])
```

## Expected behavior

Either the error should not be raised, or a better explanation of why this happens is needed in the docs. The test should then be updated accordingly.

## Environment

- macOS 13.3
- Python 3.10
- Latest DGL version, installed with `pip install dgl -f https://data.dgl.ai/wheels/repo.html` together with PyTorch (`pip install torch`)