[Documentation] make documentation of GATConv and GATv2Conv more explicit and accurate

[SZiesche]

What?

I explicitly named the different \Thetas used in the different layers by an index s and t (like it is already done in the docs when talking about input sizes). I did the same for a in the GATconv layer. Finally removed the unclear || operator and just wrote out the few additions.

Why?

To solve the issues discussed in #8200

Testing?

No tests needed, as only documentation was changed.

[codecov]

Codecov Report

Merging #8201 (3b7ff8c) into master (7085a3a) will decrease coverage by 0.02%.
The diff coverage is n/a.

❗ Current head 3b7ff8c differs from pull request most recent head ae15bc1. Consider uploading reports for the commit ae15bc1 to get more accurate results

@@            Coverage Diff             @@
##           master    #8201      +/-   ##
==========================================
- Coverage   87.40%   87.39%   -0.02%     
==========================================
  Files         473      473              
  Lines       28613    28613              
==========================================
- Hits        25010    25006       -4     
- Misses       3603     3607       +4     

Files	Coverage Δ
torch_geometric/nn/conv/gat_conv.py	100.00% <ø> (ø)
torch_geometric/nn/conv/gatv2_conv.py	94.73% <ø> (ø)

... and 1 file with indirect coverage changes

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

[SZiesche]

ok great, should we make the two failing checks pass? I had a look at the changelog, but the PR doesn't seem to contain something worth mentioning there. The test for previous pytorch version seems not to be failing because of this PR.

How can the stuff be merged in the end?

[SZiesche]

hey guys, I wanted to ask again, who can push the merge button (once this is updated with master) and if we need to make the changelog happy.

[EdisonLeeeee]

@rusty1s would take care of it :)

[rusty1s]

Thank you, this is indeed much clearer :)

[SZiesche]

Thank you, this is indeed much clearer :)

It was a pleasure. I always like to share such improvements. Documentation is always a difficuilt point, because the author of the code is often hugely biased and has a large knowledge advantage, which makes it pretty hard for him to write the docs.