🆕 random seed support

[GiulioRossetti]

Overview

This PR introduces a centralized mechanism for setting and managing random seeds in cdlib, improving reproducibility across experiments (addressing issue #229 ).

Changes

• Add cdlib.seed(seed) to set a global random seed
• Add cdlib.reset_seed() to reset the seed
• Add warning if multiple seeds are set
• Add cdlib.fixed_seed(seed) context manager for temporary seeding
• Modify algorithms to fallback to the global seed if no seed is passed
• Add comprehensive unit tests
• Update documentation and README

Why

Ensuring reproducibility is crucial in scientific experiments. This update provides users with fine control over randomness, with minimal impact on the existing API.

Notes

• networkx, numpy, scipy, networkit, igraph, scikit-learn, graph-tool are covered as much as possible
• Some external C++ randomness (e.g., deep inside graph-tool) remains outside direct Python control