hoops_ai.create_flow

hoops_ai.create_flow(name, tasks, flows_outputdir, max_workers=None, ml_task=None, debug=False, auto_dataset_export=True, export_visualization=True)

Module-level flow creation function with simplified parameters.

Parameters:

  • name (str) – Name of the flow

  • tasks (List[Any]) – List of decorated functions to execute

  • flows_outputdir (str) – Output directory for flow results

  • max_workers (int) – Number of parallel workers (None = auto-detect CPU count)

  • ml_task (str) – Optional ML task description

  • debug (bool) – Enable debug logging

  • auto_dataset_export (bool) – Automatically inject dataset export task

  • export_visualization (bool) – Export visualization resources (stream cache, PNG) for each processed file

Returns:

Configured Flow instance

Example:

import hoops_ai
from hoops_ai.flowmanager import flowtask

@flowtask.transform(name="encode", parallel_execution=True)
def my_encoder(cad_file, cad_loader, storage):
    # transformation logic
    return face_count, edge_count

flow = hoops_ai.create_flow(
    name="my_flow",
    tasks=[my_encoder],
    flows_outputdir="./output",
    max_workers=4
)