Input nodes

An input node holds the values you supply to a pipeline, such as input file paths and scalar parameters. It runs no tool; its values are passed to the downstream nodes it is connected to.

rf_in = g.add_input_node(
    input_pdb="/vol/inputs/1l9h.pdb",
    contigs="[150-150]",
    num_designs="2",
)

Keys must match the downstream tool

Each key must match one of the input names declared by the downstream tool. A key that matches no input is ignored, which is the usual cause of a parameter not taking effect. See Tools and the registry for how to look up a tool’s inputs.

Values set on an input node take priority over values produced upstream for the same key, so set a key only when you intend to override.

Multiple input nodes

A node may take several input nodes alongside an upstream tool, which is useful for keeping shared parameters separate from the data being processed:

mpnn_in = g.add_input_node(num_seq_per_target="2", sampling_temp="0.1")
g.add_edge((rfdiffusion, proteinmpnn), (mpnn_in, proteinmpnn))

Reference

See the API reference for the full method and parameter reference.