Type annotations for boto3 SageMaker

Warnings

• breaking Starting with `mypy-boto3-builder` version 8.12.0 (which generates these stubs), Python 3.8 support has been removed across all packages. Additionally, packages migrated to follow PEP 561 distribution standards.
  Fix: Ensure your project uses Python 3.9 or higher. For older Python versions, use an older `mypy-boto3-sagemaker` release compatible with Python 3.8 (e.g., from `mypy-boto3-builder < 8.12.0`).
• breaking Breaking changes in TypeDef naming conventions were introduced in `mypy-boto3-builder` 8.9.0. This includes shorter names for packed method arguments (e.g., `CreateDistributionRequestRequestTypeDef` -> `CreateDistributionRequestTypeDef`) and moving `Extra` postfixes in conflicting TypeDefs (e.g., `CreateDistributionExtraRequestTypeDef` -> `CreateDistributionRequestExtraTypeDef`).
  Fix: Review and update explicit `TypeDef` imports and usage if you were relying on the older naming conventions. Refer to the specific service module's `type_defs` for the updated names.
• gotcha Pylint may issue `undefined-variable` warnings when using `TYPE_CHECKING` blocks to conditionally import `mypy-boto3-*` types. This is a known issue with Pylint's handling of `TYPE_CHECKING`.
  Fix: To fix this, explicitly set types to `object` in the `else` branch of the `TYPE_CHECKING` block. Example: `if TYPE_CHECKING: from mypy_boto3_sagemaker.client import SagemakerClient else: SagemakerClient = object`.
• gotcha PyCharm users might experience slow performance with Literal overloads in `boto3-stubs`. This is a known IDE issue (PY-40997).
  Fix: If experiencing performance issues in PyCharm, consider installing the 'lite' version of `boto3-stubs` (e.g., `pip install boto3-stubs-lite[sagemaker]`) which provides type annotations without session.client/resource overloads but requires more explicit type annotations.
• deprecated The `sms-voice` service was deprecated and removed from `mypy-boto3-builder` in version 8.11.0. Users of this service should migrate to `pinpoint-sms-voice`.
  Fix: If you were using stubs for `sms-voice`, switch to using stubs for `pinpoint-sms-voice` instead.

Install

• pip install mypy-boto3-sagemaker Install standalone SageMaker stubs
• pip install 'boto3-stubs[sagemaker]' Install SageMaker stubs via boto3-stubs extras

Imports

• SageMakerClient

  from mypy_boto3_sagemaker.client import SageMakerClient

• SagemakerClient

  from mypy_boto3_sagemaker import SagemakerClient

  While 'from mypy_boto3_sagemaker.client import SagemakerClient' technically works, the top-level import is idiomatic and often sufficient for the primary client type.

Quickstart

This quickstart demonstrates how to instantiate a type-hinted SageMaker client using `boto3` and `mypy-boto3-sagemaker`. The `SageMakerClient` type provides comprehensive type checking and auto-completion for all available SageMaker API operations and their parameters.

import boto3
from boto3.session import Session
from mypy_boto3_sagemaker.client import SagemakerClient

def get_sagemaker_client() -> SagemakerClient:
    # boto3 automatically picks up credentials from environment variables or AWS config.
    # For local testing without actual AWS calls, ensure 'boto3' is configured.
    session = Session()
    client: SagemakerClient = session.client("sagemaker")
    return client

# Example usage (will only run if AWS credentials are set up)
if __name__ == "__main__":
    try:
        sagemaker_client = get_sagemaker_client()
        # Access type-hinted methods, e.g., sagemaker_client.list_training_jobs()
        print(f"Successfully created SageMaker client: {sagemaker_client.meta.region_name}")
        # Uncomment below to make an actual API call (requires permissions)
        # response = sagemaker_client.list_training_jobs(MaxResults=1)
        # print(f"Training jobs: {response.get('TrainingJobSummaries')}")
    except Exception as e:
        print(f"Error creating SageMaker client or making call: {e}")