should_run_policies.py

import nx_parallel as nxp


__all__ = [
    "default_should_run",
    "should_skip_parallel",
    "should_run_if_large",
    "should_run_if_nodes_none",
    "should_run_if_sparse",
]


def should_skip_parallel(*_):
    return "Fast algorithm; skip parallel execution"


def should_run_if_large(G, *_):
    if len(G) <= 200:
        return "Graph too small for parallel execution"
    return True


def default_should_run(*_):
    n_jobs = nxp.get_n_jobs()
    print(f"{n_jobs=}")
    if n_jobs in (None, 0, 1):
        return "Parallel backend requires `n_jobs` > 1 to run"
    return True


def should_run_if_nodes_none(G, nodes=None, *_):
    if nodes is None:
        return True
    return "`nodes` should be None for parallel execution"


def should_run_if_sparse(G, *args, threshold=0.3, **kwargs):
    if hasattr(G, "graph_object"):
        G = G.graph_object

    nodes = len(G)
    # Handle trivial graphs separately to avoid division by zero
    if nodes <= 1:
        return "Empty graph" if nodes == 0 else "Single-node graph"

    density = 2 * G.number_of_edges() / (nodes * (nodes - 1))
    return (
        True
        if density <= threshold
        else "Graph too dense to benefit from parallel execution"
    )