[Bug] Sparse example graph_transformer.py crashed

[chang-l]

🐛 Bug

Hi, I found one of the SparseAPI -related example crashed due to this function:

dgl/python/dgl/transforms/functional.py

Line 3614 in 617979d

def lap_pe(g, k, padding=False, return_eigval=False):

.

Details attached below...

To Reproduce

Steps to reproduce the behavior:

python examples/sparse/graph_transformer.py 

Output error:

Traceback (most recent call last):
  File "/workspace/examples/SparseAPI/graph_transformer.py", line 184, in <module>
    g.ndata["PE"] = dgl.lap_pe(g, k=pos_enc_size, padding=True)
  File "/usr/local/lib/python3.10/dist-packages/dgl/transforms/functional.py", line 3703, in lap_pe
    eigvals = F.tensor(topk_EigVal, dtype=F.float32)
UnboundLocalError: local variable 'topk_EigVal' referenced before assignment

Environment

• DGL Version (e.g., 1.0): 1.1
• Backend Library & Version (e.g., PyTorch 0.4.1, MXNet/Gluon 1.3): PyTorch
• OS (e.g., Linux): Ubuntu 22.04
• How you installed DGL (conda, pip, source): source
• GPU models and configuration (e.g. V100): A5000

Additional context

This may be due to missing to define topk_EigVal in the following conditional block:

dgl/python/dgl/transforms/functional.py

Lines 3684 to 3693 in 617979d

	if k + 1 < n - 1:
	# Use scipy if k + 1 < n - 1 for memory efficiency.
	EigVal, EigVec = scipy.sparse.linalg.eigs(
	L, k=k + 1, which="SR", tol=1e-2
	)
	topk_indices = EigVal.argsort()[1:]
	# Since scipy may return complex value, to avoid crashing in NN code,
	# convert them to real number.
	topk_eigvals = EigVal[topk_indices].real
	topk_EigVec = EigVec[:, topk_indices].real