Graph Class
A class to define a pipeline run graph.
Initialize Graph.
Constructor
Graph(name, context)Parameters
| Name | Description |
|---|---|
|
name
Required
|
Name of the graph. |
|
context
Required
|
<xref:azureml.pipeline.core._GraphContext>
The current graph context. |
|
name
Required
|
Name of the graph. |
|
context
Required
|
<xref:azureml.pipeline.core._GraphContext>
The current graph context. |
Methods
| add_datasource_node |
Add a datasource node to the graph. |
| add_module_node |
Add a module node to the graph. |
| connect |
Connect two ports and creates an Edge. |
| delete_node |
Delete a node from the graph. |
| finalize |
Finalize resources for nodes in the graph. |
| generate_yaml |
Generate the yaml representation of the graph. |
| get_node |
Get a node by ID. |
| sequence |
Configure a list of nodes to run in a sequence following the first node in the list. |
| submit |
Submit the graph to run in the cloud. |
| validate |
Validate graph. Returns a list of errors. |
add_datasource_node
Add a datasource node to the graph.
add_datasource_node(name, datasource=None, datasource_builder=None, datapath_param_name=None)Parameters
| Name | Description |
|---|---|
|
name
Required
|
Name of the node. |
|
datasource
|
Datasource for this node. Default value: None
|
|
datasource_builder
|
<xref:azureml.pipeline.core._DatasourceBuilder>
_DatasourceBuilder for this node. Default value: None
|
|
datapath_param_name
|
Datapath parameter name. Default value: None
|
Returns
| Type | Description |
|---|---|
|
node |
add_module_node
Add a module node to the graph.
add_module_node(name, input_bindings, output_bindings=None, param_bindings=None, module=None, module_builder=None, module_wiring=None, runconfig=None, cloud_settings=None)Parameters
| Name | Description |
|---|---|
|
name
Required
|
Name of the node |
|
input_bindings
Required
|
List of input port bindings. |
|
output_bindings
|
List of output port bindings. Default value: None
|
|
param_bindings
|
Dictionary of name-value pairs for parameter assignments. Default value: None
|
|
module
|
Module for this node Default value: None
|
|
module_builder
|
<xref:azureml.pipeline.core._ModuleBuilder>
_ModuleBuilder for this node Default value: None
|
|
module_wiring
|
A mapping between the node's inputs/outputs and the module inputs/outputs. Holds two keys, inputs and outputs, each mapped to a dict whose keys are the module's inputs/outputs names, and the values are the node's ports. Default value: None
|
|
runconfig
|
Runconfig that will be used for python_script_step Default value: None
|
|
cloud_settings
|
<xref:azureml.pipeline.core._restclients.aeva.models.CloudSettings>
Settings that will be used for clouds Default value: None
|
Returns
| Type | Description |
|---|---|
|
node |
connect
Connect two ports and creates an Edge.
connect(source_port, dest_port)Parameters
| Name | Description |
|---|---|
|
source_port
Required
|
Output port from the node that is the source of the connection |
|
dest_port
Required
|
Input port from the node that is the destination of the connection |
Returns
| Type | Description |
|---|---|
|
edge |
delete_node
Delete a node from the graph.
delete_node(node_id)Parameters
| Name | Description |
|---|---|
|
node_id
Required
|
node id |
finalize
Finalize resources for nodes in the graph.
finalize(dry_run=None, regenerate_outputs=False)Parameters
| Name | Description |
|---|---|
|
dry_run
|
Set to True to verify that the graph can be built without making any external API calls to Azure ML service. Default value: None
|
|
regenerate_outputs
|
Set to True to force a new run (disallows module/datasource reuse). Default value: False
|
Returns
| Type | Description |
|---|---|
|
Dictionary of {node_id, (resource_id, is_new_resource)} |
generate_yaml
Generate the yaml representation of the graph.
generate_yaml()Returns
| Type | Description |
|---|---|
|
Return the yaml dict |
get_node
sequence
Configure a list of nodes to run in a sequence following the first node in the list.
sequence(nodes)Parameters
| Name | Description |
|---|---|
|
nodes
Required
|
The list of nodes. |
submit
Submit the graph to run in the cloud.
submit(pipeline_parameters=None, continue_on_step_failure=False, regenerate_outputs=False, parent_run_id=None, **kwargs)Parameters
| Name | Description |
|---|---|
|
pipeline_parameters
|
Parameters for pipeline execution. Optional. Default value: None
|
|
continue_on_step_failure
|
Indicates whether to let the experiment continue executing if one step fails. If True, only steps that have no dependency on the output of the failed step will continue execution. Default value: False
|
|
regenerate_outputs
|
Set to True to force a new run (disallows module/datasource reuse) Default value: False
|
|
parent_run_id
|
Optional run ID to set for the parent run of this pipeline run, which is reflected in RunHistory. The parent run must belong to same experiment as this pipeline is being submitted to. Default value: None
|
|
kwargs
Required
|
Custom keyword arguments, reserved for future development. |
Returns
| Type | Description |
|---|---|
|
A PipelineRun. |
validate
Attributes
datasource_nodes
node_dict
node_name_dict
params
Get a dictionary containing all graph parameters. Values are literal types or data reference as JSON string.
Returns
| Type | Description |
|---|---|
|
Dictionary of {param name, param value} |