📦 Pull Request: Add Implementations and Example Scripts for Tables 3–8 (Zhou et al. 2024)

Thank you for reviewing this contribution!
This PR adds the full experimental implementations and lightweight example scripts reproducing Tables 3–8 from Zhou et al. (2024), “Revisiting Black-box Ownership Verification for Graph Neural Networks,” following the PyGIP project conventions.

📋 Summary

This PR introduces complete experiment pipelines and aligned example scripts, structured according to PyGIP’s dataset, model, and device conventions.

Changes introduced:

• Implemented all core experiments under the implementation/ folder:
  • run_bboxve.py — Table 3 (BBoxVe)
  • run_bgrove.py — Table 4 (BGrOVe)
  • run_table5_full.py — Table 5 & Figure 3
  • adversial.py — Tables 6 & 7
  • adversial_table8.py — Table 8
• Added reviewer-friendly example scripts under examples/, one for each table, to demonstrate how to invoke the corresponding implementation.
• Organized experiment results under results/ (CSV outputs per table).

Structural improvements:

• Updated dataset wrappers in pygip/datasets/ for compatibility with the experiment structure.
• Refactored model backbones under pygip/models/nn/pyg_backbones.py to ensure consistent import paths (e.g., from pygip.models.nn.pyg_backbones import GCN, GAT, GraphSAGE, ...).
• These changes only improve consistency and modularity; they do not modify the experimental logic or results.

✅ Checklist

• My code follows the project's coding style
• I have tested the changes and verified that they work
• I have added necessary documentation (example scripts + folder structure)
• I have linked related issues above (if any)
• The PR is made from a feature branch (feat/implementation_iqra)

🧠 Additional Context (Optional)

• All experiments replicate reported behavior from Zhou et al. (2024).
• Example scripts provide simple, reproducible entry points for reviewers. In case of an error, you can just run the script directly.
• Dataset and backbone structure now align fully with PyGIP’s import hierarchy.