[GraphBolt] FeatureFetcher throws exception if input_nodes does not contain target feature

[Rhett-Ying]

🐛 Bug

dgl/python/dgl/graphbolt/feature_fetcher.py

Lines 91 to 92 in 2da713f

	for type_name, feature_names in self.node_feature_keys.items():
	nodes = input_nodes[type_name]

If input_nodes does not contain any keys(node/edge types) that are listed in node/edge_feature_key, exception is thrown. Should we skip for such case?

Expected behavior

Skip non-existent feature keys.

Environment

• DGL Version (e.g., 1.0): master
• Backend Library & Version (e.g., PyTorch 0.4.1, MXNet/Gluon 1.3):
• OS (e.g., Linux):
• How you installed DGL (conda, pip, source):
• Build command you used (if compiling from source):
• Python version:
• CUDA/cuDNN version (if applicable):
• GPU models and configuration (e.g. V100):
• Any other relevant information:

Additional context

[mfbalin]

Should we fix this before DGL 2.2 release as well? Seems like a big problem for the users.

[mfbalin]

Isn't the fix simple adding a single if expression to check if input_nodes contains the key or not? If it doesn't, we can simply skip.