Upstream Stage Dependencies

When a stage depends on upstream stages, snap_workflow automatically creates soft links to the parent SPoint output directory inside the child runner directory.

For example, if your stage stage_B depends on stage_A, you can load files directly using the stage name as the directory path:

from snap_workflow import SPoint_Manager
import numpy as np

with SPoint_Manager() as sm:
    # Load parent data from the automatically linked directory
    parent_output = np.load("stage_A/state.npy")

    # Run B's computation using A's output
    output = parent_output * 2.0
    np.save("output_B.npy", output)