Lightning Trainer integration in GraphGym

[aniketmaurya]

Third PR from the Roadmap.

• Implements PL Trainer
• Soft import for optional dependencies
• Update tests

[codecov]

Codecov Report

Merging #4689 (94ab45d) into master (70a3760) will decrease coverage by 0.03%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##           master    #4689      +/-   ##
==========================================
- Coverage   82.43%   82.39%   -0.04%     
==========================================
  Files         322      323       +1     
  Lines       17237    17215      -22     
==========================================
- Hits        14209    14184      -25     
- Misses       3028     3031       +3     

Impacted Files	Coverage Δ
torch_geometric/graphgym/config.py	95.04% <ø> (-0.03%)	⬇️
torch_geometric/graphgym/imports.py	100.00% <100.00%> (ø)
torch_geometric/graphgym/logger.py	84.61% <100.00%> (+2.75%)	⬆️
torch_geometric/graphgym/model_builder.py	98.14% <100.00%> (-0.04%)	⬇️
torch_geometric/graphgym/train.py	100.00% <100.00%> (+1.78%)	⬆️
torch_geometric/graphgym/utils/agg_runs.py	46.54% <100.00%> (ø)
torch_geometric/graphgym/utils/epoch.py	57.14% <0.00%> (-42.86%)	⬇️
torch_geometric/graphgym/checkpoint.py	38.29% <0.00%> (-42.56%)	⬇️
torch_geometric/data/lightning_datamodule.py	38.26% <0.00%> (+12.17%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 70a3760...94ab45d. Read the comment docs.

[rusty1s]

This is amazingly clean. Thank you so much! I love it :)