refactor graph lifting & enhancements

[ekarais]

Objective

This branch began with the aim to refactor graph lifting functionality but evolved to encompass various improvements and fixes across the repository.

Key Changes

Refactoring:

• Refactored graph lifting by separating the cell generation algo (e.g. rips_lift, clique_lift) from the final sample generation.
• Consolidated utility functions by integrating simplicial_data.py contents into utils.py.
• Streamlined legacy code management by relocating old simplicial_data/ code to the newly created legacy/.

Testing & Development:

• Introduced testing via pytest and pre-commit.
• Added comprehensive tests for the refactoring efforts.
• Addressed several bugs, including import errors and an issue inherited from the parent branch, to stabilize the codebase.

Usability & Documentation:

• Added documentation to simplicial_data/
• Improved the development environment with updated setup scripts,and a data subsampling option.
• Simplified and fixed issues within the Conda environment setup to streamline developer onboarding and project setup.

Performance Monitoring & Code Optimization:

• Introduced logging of duration per epoch to WandB for better performance tracking.
• Added torch.compile(model)

Additional Notes

The collective updates not only address the initial goal of refactoring but also introduce significant enhancements that improve the development workflow, code maintainability, and project scalability.

Request for Review

Feedback is highly encouraged to ensure these changes align with our project objectives and meet our quality standards.

[ekarais]

I decided to remove [all] after noticing that it would install the following:

doc = [
    "jupyter",
    "nbsphinx",
    "nbsphinx_link",
    "sphinx <= 5.3.0",
    "sphinx_gallery",
    "pydata-sphinx-theme"
]
lint = [
    "black < 23.0",
    "black[jupyter]",
    "flake8",
    "flake8-docstrings",
    "Flake8-pyproject",
    "isort",
    "pre-commit"
]
test = [
    "pytest",
    "pytest-cov",
    "coverage",
    "jupyter"
]

These are really only needed if we were to contribute to TopoNetX so I think it is fine to leave them out. We can consider adding black, isort etc. as pre-commit hooks but this would be a different issue. @mauriciogtec can you try creating the environment on your end before we merge this?