snorkelflow.client.nodes.add_node

snorkelflow.client.nodes.add_node(application, input_node_uids=None, expected_op_type=None, node_config=None, output_node_uid=None, output_node_uids=None, node_cls='ApplicationNode', op_type=None, op_config=None, add_to_parent_block=False)

Adds a node to the graph. Must specify one of input_node_uids or output_node_uids to determine the location of the node.

Parameters

Parameters

Name	Type	Default	Info
application	Union[str, int]		The name or UID of the application.
input_node_uids	Optional[List[int]]	None	The list of input nodes. Use [-1] if your input node is the dataset node.
expected_op_type	Optional[str]	None	The type of node.
node_config	Optional[Dict[str, Any]]	None	A dictionary config for the node.
output_node_uids	Optional[List[int]]	None	An optional list of output node ids.
output_node_uid	Optional[int]	None	Deprecated. An optional output node id.
node_cls	str	'ApplicationNode'	The name of the node class.
op_type	Optional[str]	None	The operator type to be committed, e.g., TruncatePreprocessor.
op_config	Optional[Dict[str, Any]]	None	The operator config, e.g., {“field”: “text”, “by”: “chars”, “length”: 5}. Use get_default_operator() to check required fields for built-in operators.
add_to_parent_block	Optional[bool]	False	The boolean value to determine if the new node is added to the parent block. Note: in an application with a single existing block with input_node_uids=[-1], setting this parameter to True will result in a node added ahead of the block, not within it.

Returns

Returns

Information about the created node object

Return type

Return type

Dict[str, Any]

Examples

>>> # Add a place holder node
>>> sf.add_node(
>>>     application=APP_NAME,
>>>     input_node_uids=[input_node],
>>>     expected_op_type="Featurizer",
>>>     output_node_uids=[output_node],
>>> )
{'node_uid': <node_uid>, 'op_version_uid': None}

>>> # Add a committed node
>>> sf.add_node(
>>>     application=APP_NAME,
>>>     input_node_uids=[input_node],
>>>     op_type="TruncatePreprocessor",
>>>     op_config={"field": "text", "by": "chars", "length": 5},
>>>     output_node_uids=[output_node],
>>> )
{'node_uid': <node_uid>, 'op_version_uid': <op_version_uid>}