Remove adding num_edges directly to store in the Pad transform

[bwroblew]

Handling num_edges as an attribute requires couple more changes and doesn't work well with the current implementation. For example, when num_edges is set on the Data objects and those objects are merged into a Batch the num_edges attribute is a tensor with values from all the objects. It works well when implemented as a property of EdgeStorage and not as a direct attribute, so removing it for now.

[codecov]

Codecov Report

Merging #6758 (5bd5957) into master (e9dc546) will decrease coverage by 1.58%.
The diff coverage is n/a.

❗ Current head 5bd5957 differs from pull request most recent head e85243d. Consider uploading reports for the commit e85243d to get more accurate results

@@            Coverage Diff             @@
##           master    #6758      +/-   ##
==========================================
- Coverage   90.78%   89.21%   -1.58%     
==========================================
  Files         425      425              
  Lines       23152    23113      -39     
==========================================
- Hits        21019    20620     -399     
- Misses       2133     2493     +360     

Impacted Files	Coverage Δ
torch_geometric/transforms/pad.py	95.58% <ø> (-0.05%)	⬇️
torch_geometric/nn/models/dimenet_utils.py	0.00% <0.00%> (-75.52%)	⬇️
torch_geometric/nn/models/dimenet.py	17.66% <0.00%> (-49.78%)	⬇️
torch_geometric/visualization/graph.py	68.25% <0.00%> (-26.99%)	⬇️
torch_geometric/nn/conv/utils/typing.py	83.75% <0.00%> (-15.00%)	⬇️
torch_geometric/nn/pool/asap.py	91.89% <0.00%> (-8.11%)	⬇️
torch_geometric/resolver.py	82.75% <0.00%> (-6.90%)	⬇️
torch_geometric/nn/dense/linear.py	88.97% <0.00%> (-5.92%)	⬇️
torch_geometric/nn/models/attentive_fp.py	95.83% <0.00%> (-4.17%)	⬇️
torch_geometric/sampler/neighbor_sampler.py	91.56% <0.00%> (-3.22%)	⬇️
... and 29 more

📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more

[rusty1s]

Makes sense, there is some custom logic in Batch to obtain the real number of nodes, which I assume we would need to copy?

[bwroblew]

Yes, I can work on that later and extend that logic. I guess for now it's not necessary and it was a mistake on my side to set it explicitly.
There are couple of places where num_nodes has a separate logic (for example when creating a subgraph) and similar logic would be probably needed for num_edges.
I guess it's best to remove it now and I can work on that when it's really needed.