torchcell.datasets.NodeEmbeddingBuilder

class NodeEmbeddingBuilder

Bases: object

Builder class for creating node embedding datasets based on configuration.

classmethod build(embedding_names: list[str], data_root: str, genome: Any, graph: Any | None = None) → dict[str, Any]

Build node embeddings based on configuration.

Parameters:

• embedding_names – List of embedding names to create

• data_root – Root data directory path

• genome – SCerevisiaeGenome instance

• graph – SCerevisiaeGraph instance (required for graph embeddings)

Returns:

Dictionary mapping embedding names to dataset instances

Raises:

ValueError – If an unknown embedding name is provided or required dependencies are missing

classmethod check_learnable_embedding(embedding_names: list[str]) → bool

Check if learnable embedding is requested.