Objects without a cross-chunk edge fail during edit

[ceesem]

Some merges with small objects fail with the error: int() argument must be a string, a bytes-like object or a number, not 'NoneType'

Looking at the logs, it seems to come from:

"
7: "  File "./pychunkedgraph/graph/operation.py", line 438, in execute
    timestamp=override_ts if override_ts else timestamp,
"
8: "  File "./pychunkedgraph/graph/operation.py", line 631, in _apply
    rows=rows,
"
9: "  File "./pychunkedgraph/graph/edits.py", line 243, in add_edges
    return create_parents.run(), new_l2_ids, rows + create_parents.create_new_entries()
"
10: "  File "./pychunkedgraph/graph/edits.py", line 504, in run
    self._create_new_parents(layer)
"
11: "  File "./pychunkedgraph/graph/edits.py", line 468, in _create_new_parents
    layer_node_ids, layer, lower_layer_ids
"
12: "  File "./pychunkedgraph/graph/edits.py", line 418, in _get_connected_components
    cross_edges = get_sv_parents(sv_cross_edges)
"
13: "  File "/usr/local/lib/python3.7/site-packages/numpy/lib/function_base.py", line 2163, in __call__
    return self._vectorize_call(func=func, args=vargs)
"
14: "  File "/usr/local/lib/python3.7/site-packages/numpy/lib/function_base.py", line 2249, in _vectorize_call
    res = asanyarray(outputs, dtype=otypes[0])
"
15: "TypeError: int() argument must be a string, a bytes-like object or a number, not 'NoneType'
"

Which makes it seem like the offending line is here, where the np.vectorize method tries to return the cross-chunk edge list as an array of ints and has no way to handle the case where no cross-chunk edges exist for the object:
https://github.com/seung-lab/PyChunkedGraph/blob/82ae75c091fd1a7fc41c0c73a94f9a8179acf1a2/pychunkedgraph/graph/edits.py#L409

[ceesem]

(This is for pcgv2)