sagemaker.core.tools.shapes_codegen

A class for generating class structure from Service Model JSON.

To run the script be sure to set the PYTHONPATH export PYTHONPATH=<sagemaker-code-gen repo directory>:$PYTHONPATH

Classes

ShapesCodeGen()

Generates shape classes based on an input Botocore service.json.

class sagemaker.core.tools.shapes_codegen.ShapesCodeGen

Bases: object

Generates shape classes based on an input Botocore service.json.

Parameters:

service_json (dict) – The Botocore service.json containing the shape definitions.

service_json

The Botocore service.json containing the shape definitions.

Type:

dict

shapes_extractor

An instance of the ShapesExtractor class.

Type:

ShapesExtractor

shape_dag

Shape DAG generated from service.json

Type:

dict

build_graph()

Builds a directed acyclic graph (DAG) representing the dependencies between shapes.

topological_sort()

Performs a topological sort on the DAG to determine the order in which shapes should be generated.

generate_data_class_for_shape(shape)

Generates a data class for a given shape.

_generate_doc_string_for_shape(shape)

Generates the docstring for a given shape.

generate_imports()

Generates the import statements for the generated shape classes.

generate_base_class()

Generates the base class for the shape classes.

_filter_input_output_shapes(shape)

Filters out shapes that are used as input or output for operations.

generate_shapes(output_folder)

Generates the shape classes and writes them to the specified output folder.

build_graph()

Builds a directed acyclic graph (DAG) representing the dependencies between shapes.

Steps: 1. Loop over the Service Json shapes.

1.1. If dependency(members) found, add association of node -> dependency.

1.1.1. Sometimes members are not shape themselves, but have associated links to actual shapes.

In that case add link to node -> dependency (actual)

CreateExperimentRequest -> [ExperimentEntityName, ExperimentDescription, TagList]

1.2. else leaf node found (no dependent members), add association of node -> None.

Returns:

A dict which defines the structure of the DAG in the format: {key : [dependencies]} Example input:

{‘CreateExperimentRequest’: [‘ExperimentEntityName’, ‘ExperimentEntityName’,

’ExperimentDescription’, ‘TagList’],

’CreateExperimentResponse’: [‘ExperimentArn’], ‘DeleteExperimentRequest’: [‘ExperimentEntityName’], ‘DeleteExperimentResponse’: [‘ExperimentArn’]}

generate_base_class()

Generates the base class for the shape classes.

Returns:

The generated base class as a string.

generate_data_class_for_shape(shape)

Generates a data class for a given shape.

Parameters:

shape – The name of the shape.

Returns:

The generated data class as a string.

generate_imports()

Generates the import statements for the generated shape classes.

Returns:

The generated import statements as a string.

generate_license()

Generates the license string.

Returns:

The license string.

Return type:

str

generate_shapes(output_folder='/home/docs/checkouts/readthedocs.org/user_builds/sagemaker/checkouts/5737/docs/src/sagemaker/core', file_name='shapes.py') → None

Generates the shape classes and writes them to the specified output folder.

Parameters:

output_folder – The path to the output folder.

topological_sort()

Performs a topological sort on the DAG to determine the order in which shapes should be generated.

Returns:

A list of shape names in the order of topological sort.