Skip to content

Groggy Documentation

Pipeline API Reference

rollingstorms/groggy

Pipeline API Reference¶

The pipeline module lives at groggy.pipeline and exposes two primary entry points:

`apply(subgraph, algorithm_or_pipeline)`¶

Convenience helper that accepts a Subgraph, plus one of the following:

An AlgorithmHandle (runs the algorithm once)
A list/tuple of handles (runs them sequentially)
A Pipeline object (reuses an existing compiled pipeline)

Returns a new Subgraph with all algorithm results applied. Internally this wraps the same Rust executor used by Subgraph.apply().

`Pipeline`¶

from groggy import pipeline, algorithms

pipe = pipeline([
    algorithms.centrality.pagerank(output_attr="pr"),
    algorithms.pathfinding.bfs(start_attr="is_start", output_attr="dist"),
])

result = pipe(subgraph)
print(len(pipe))        # Number of steps
print(pipe)             # String summary

Methods:

__call__(subgraph) / run(subgraph) — Execute the pipeline
__len__() — Number of algorithm steps
__repr__() — Developer-friendly summary

Pipeline objects automatically free their native handles when the Python object is collected.