Step Functions examples using SDK for Python (Boto3) - AWS SDK Code Examples

There are more AWS SDK examples available in the AWS Doc SDK Examples GitHub repo.

Step Functions examples using SDK for Python (Boto3)

The following code examples show you how to perform actions and implement common scenarios by using the AWS SDK for Python (Boto3) with Step Functions.

Basics are code examples that show you how to perform the essential operations within a service.

Actions are code excerpts from larger programs and must be run in context. While actions show you how to call individual service functions, you can see actions in context in their related scenarios.

Scenarios are code examples that show you how to accomplish specific tasks by calling multiple functions within a service or combined with other AWS services.

Each example includes a link to the complete source code, where you can find instructions on how to set up and run the code in context.

Get started

The following code examples show how to get started using Step Functions.

SDK for Python (Boto3)
Note

There's more on GitHub. Find the complete example and learn how to set up and run in the AWS Code Examples Repository.

import boto3 def hello_stepfunctions(stepfunctions_client): """ Use the AWS SDK for Python (Boto3) to create an AWS Step Functions client and list the state machines in your account. This list might be empty if you haven't created any state machines. This example uses the default settings specified in your shared credentials and config files. :param stepfunctions_client: A Boto3 Step Functions Client object. """ print("Hello, Step Functions! Let's list up to 10 of your state machines:") state_machines = stepfunctions_client.list_state_machines(maxResults=10) for sm in state_machines["stateMachines"]: print(f"\t{sm['name']}: {sm['stateMachineArn']}") if __name__ == "__main__": hello_stepfunctions(boto3.client("stepfunctions"))
  • For API details, see ListStateMachines in AWS SDK for Python (Boto3) API Reference.

Basics

The following code example shows how to:

  • Create an activity.

  • Create a state machine from an Amazon States Language definition that contains the previously created activity as a step.

  • Run the state machine and respond to the activity with user input.

  • Get the final status and output after the run completes, then clean up resources.

SDK for Python (Boto3)
Note

There's more on GitHub. Find the complete example and learn how to set up and run in the AWS Code Examples Repository.

Run an interactive scenario at a command prompt.

class StateMachineScenario: """Runs an interactive scenario that shows how to get started using Step Functions.""" def __init__(self, activity, state_machine, iam_client): """ :param activity: An object that wraps activity actions. :param state_machine: An object that wraps state machine actions. :param iam_client: A Boto3 AWS Identity and Access Management (IAM) client. """ self.activity = activity self.state_machine = state_machine self.iam_client = iam_client self.state_machine_role = None def prerequisites(self, state_machine_role_name): """ Finds or creates an IAM role that can be assumed by Step Functions. A role of this kind is required to create a state machine. The state machine used in this example does not call any additional services, so it needs no additional permissions. :param state_machine_role_name: The name of the role. :return: Data about the role. """ trust_policy = { "Version": "2012-10-17", "Statement": [ { "Sid": "", "Effect": "Allow", "Principal": {"Service": "states.amazonaws.com"}, "Action": "sts:AssumeRole", } ], } try: role = self.iam_client.get_role(RoleName=state_machine_role_name) print(f"Prerequisite IAM role {state_machine_role_name} already exists.") except ClientError as err: if err.response["Error"]["Code"] == "NoSuchEntity": role = None else: logger.error( "Couldn't get prerequisite IAM role %s. Here's why: %s: %s", state_machine_role_name, err.response["Error"]["Code"], err.response["Error"]["Message"], ) raise if role is None: try: role = self.iam_client.create_role( RoleName=state_machine_role_name, AssumeRolePolicyDocument=json.dumps(trust_policy), ) except ClientError as err: logger.error( "Couldn't create prerequisite IAM role %s. Here's why: %s: %s", state_machine_role_name, err.response["Error"]["Code"], err.response["Error"]["Message"], ) raise self.state_machine_role = role["Role"] def find_or_create_activity(self, activity_name): """ Finds or creates a Step Functions activity. :param activity_name: The name of the activity. :return: The Amazon Resource Name (ARN) of the activity. """ print("First, let's set up an activity and state machine.") activity_arn = self.activity.find(activity_name) if activity_arn is None: activity_arn = self.activity.create(activity_name) print( f"Activity {activity_name} created. Its Amazon Resource Name (ARN) is " f"{activity_arn}." ) else: print(f"Activity {activity_name} already exists.") return activity_arn def find_or_create_state_machine( self, state_machine_name, activity_arn, state_machine_file ): """ Finds or creates a Step Functions state machine. :param state_machine_name: The name of the state machine. :param activity_arn: The ARN of an activity that is used as a step in the state machine. This ARN is injected into the state machine definition that's used to create the state machine. :param state_machine_file: The path to a file containing the state machine definition. :return: The ARN of the state machine. """ state_machine_arn = self.state_machine.find(state_machine_name) if state_machine_arn is None: with open(state_machine_file) as state_machine_file: state_machine_def = state_machine_file.read().replace( "{{DOC_EXAMPLE_ACTIVITY_ARN}}", activity_arn ) state_machine_arn = self.state_machine.create( state_machine_name, state_machine_def, self.state_machine_role["Arn"], ) print(f"State machine {state_machine_name} created.") else: print(f"State machine {state_machine_name} already exists.") print("-" * 88) print(f"Here's some information about state machine {state_machine_name}:") state_machine_info = self.state_machine.describe(state_machine_arn) for field in ["name", "status", "stateMachineArn", "roleArn"]: print(f"\t{field}: {state_machine_info[field]}") return state_machine_arn def run_state_machine(self, state_machine_arn, activity_arn): """ Run the state machine. The state machine used in this example is a simple chat simulation. It contains an activity step in a loop that is used for user interaction. When the state machine gets to the activity step, it waits for an external application to get task data and submit a response. This function acts as the activity application by getting task input and responding with user input. :param state_machine_arn: The ARN of the state machine. :param activity_arn: The ARN of the activity used as a step in the state machine. :return: The ARN of the run. """ print( f"Let's run the state machine. It's a simplistic, non-AI chat simulator " f"we'll call ChatSFN." ) user_name = q.ask("What should ChatSFN call you? ", q.non_empty) run_input = {"name": user_name} print("Starting state machine...") run_arn = self.state_machine.start(state_machine_arn, json.dumps(run_input)) action = None while action != "done": activity_task = self.activity.get_task(activity_arn) task_input = json.loads(activity_task["input"]) print(f"ChatSFN: {task_input['message']}") action = task_input["actions"][ q.choose("What now? ", task_input["actions"]) ] task_response = {"action": action} self.activity.send_task_success( activity_task["taskToken"], json.dumps(task_response) ) return run_arn def finish_state_machine_run(self, run_arn): """ Wait for the state machine run to finish, then print final status and output. :param run_arn: The ARN of the run to retrieve. """ print(f"Let's get the final output from the state machine:") status = "RUNNING" while status == "RUNNING": run_output = self.state_machine.describe_run(run_arn) status = run_output["status"] if status == "RUNNING": print( "The state machine is still running, let's wait for it to finish." ) wait(1) elif status == "SUCCEEDED": print(f"ChatSFN: {json.loads(run_output['output'])['message']}") else: print(f"Run status: {status}.") def cleanup( self, state_machine_name, state_machine_arn, activity_name, activity_arn, state_machine_role_name, ): """ Clean up resources created by this example. :param state_machine_name: The name of the state machine. :param state_machine_arn: The ARN of the state machine. :param activity_name: The name of the activity. :param activity_arn: The ARN of the activity. :param state_machine_role_name: The name of the role used by the state machine. """ if q.ask( "Do you want to delete the state machine, activity, and role created for this " "example? (y/n) ", q.is_yesno, ): self.state_machine.delete(state_machine_arn) print(f"Deleted state machine {state_machine_name}.") self.activity.delete(activity_arn) print(f"Deleted activity {activity_name}.") self.iam_client.delete_role(RoleName=state_machine_role_name) print(f"Deleted role {state_machine_role_name}.") def run_scenario(self, activity_name, state_machine_name): print("-" * 88) print("Welcome to the AWS Step Functions state machines demo.") print("-" * 88) activity_arn = self.find_or_create_activity(activity_name) state_machine_arn = self.find_or_create_state_machine( state_machine_name, activity_arn, "../../../resources/sample_files/chat_sfn_state_machine.json", ) print("-" * 88) run_arn = self.run_state_machine(state_machine_arn, activity_arn) print("-" * 88) self.finish_state_machine_run(run_arn) print("-" * 88) self.cleanup( state_machine_name, state_machine_arn, activity_name, activity_arn, self.state_machine_role["RoleName"], ) print("-" * 88) print("\nThanks for watching!") print("-" * 88) if __name__ == "__main__": logging.basicConfig(level=logging.INFO, format="%(levelname)s: %(message)s") try: stepfunctions_client = boto3.client("stepfunctions") iam_client = boto3.client("iam") scenario = StateMachineScenario( Activity(stepfunctions_client), StateMachine(stepfunctions_client), iam_client, ) scenario.prerequisites("doc-example-state-machine-chat") scenario.run_scenario("doc-example-activity", "doc-example-state-machine") except Exception: logging.exception("Something went wrong with the demo.")

Define a class that wraps state machine actions.

class StateMachine: """Encapsulates Step Functions state machine actions.""" def __init__(self, stepfunctions_client): """ :param stepfunctions_client: A Boto3 Step Functions client. """ self.stepfunctions_client = stepfunctions_client def create(self, name, definition, role_arn): """ Creates a state machine with the specific definition. The state machine assumes the provided role before it starts a run. :param name: The name to give the state machine. :param definition: The Amazon States Language definition of the steps in the the state machine. :param role_arn: The Amazon Resource Name (ARN) of the role that is assumed by Step Functions when the state machine is run. :return: The ARN of the newly created state machine. """ try: response = self.stepfunctions_client.create_state_machine( name=name, definition=definition, roleArn=role_arn ) except ClientError as err: logger.error( "Couldn't create state machine %s. Here's why: %s: %s", name, err.response["Error"]["Code"], err.response["Error"]["Message"], ) raise else: return response["stateMachineArn"] def find(self, name): """ Find a state machine by name. This requires listing the state machines until one is found with a matching name. :param name: The name of the state machine to search for. :return: The ARN of the state machine if found; otherwise, None. """ try: paginator = self.stepfunctions_client.get_paginator("list_state_machines") for page in paginator.paginate(): for state_machine in page.get("stateMachines", []): if state_machine["name"] == name: return state_machine["stateMachineArn"] except ClientError as err: logger.error( "Couldn't list state machines. Here's why: %s: %s", err.response["Error"]["Code"], err.response["Error"]["Message"], ) raise def describe(self, state_machine_arn): """ Get data about a state machine. :param state_machine_arn: The ARN of the state machine to look up. :return: The retrieved state machine data. """ try: response = self.stepfunctions_client.describe_state_machine( stateMachineArn=state_machine_arn ) except ClientError as err: logger.error( "Couldn't describe state machine %s. Here's why: %s: %s", state_machine_arn, err.response["Error"]["Code"], err.response["Error"]["Message"], ) raise else: return response def start(self, state_machine_arn, run_input): """ Start a run of a state machine with a specified input. A run is also known as an "execution" in Step Functions. :param state_machine_arn: The ARN of the state machine to run. :param run_input: The input to the state machine, in JSON format. :return: The ARN of the run. This can be used to get information about the run, including its current status and final output. """ try: response = self.stepfunctions_client.start_execution( stateMachineArn=state_machine_arn, input=run_input ) except ClientError as err: logger.error( "Couldn't start state machine %s. Here's why: %s: %s", state_machine_arn, err.response["Error"]["Code"], err.response["Error"]["Message"], ) raise else: return response["executionArn"] def describe_run(self, run_arn): """ Get data about a state machine run, such as its current status or final output. :param run_arn: The ARN of the run to look up. :return: The retrieved run data. """ try: response = self.stepfunctions_client.describe_execution( executionArn=run_arn ) except ClientError as err: logger.error( "Couldn't describe run %s. Here's why: %s: %s", run_arn, err.response["Error"]["Code"], err.response["Error"]["Message"], ) raise else: return response def delete(self, state_machine_arn): """ Delete a state machine and all of its run data. :param state_machine_arn: The ARN of the state machine to delete. """ try: response = self.stepfunctions_client.delete_state_machine( stateMachineArn=state_machine_arn ) except ClientError as err: logger.error( "Couldn't delete state machine %s. Here's why: %s: %s", state_machine_arn, err.response["Error"]["Code"], err.response["Error"]["Message"], ) raise else: return response

Define a class that wraps activity actions.

class Activity: """Encapsulates Step Function activity actions.""" def __init__(self, stepfunctions_client): """ :param stepfunctions_client: A Boto3 Step Functions client. """ self.stepfunctions_client = stepfunctions_client def create(self, name): """ Create an activity. :param name: The name of the activity to create. :return: The Amazon Resource Name (ARN) of the newly created activity. """ try: response = self.stepfunctions_client.create_activity(name=name) except ClientError as err: logger.error( "Couldn't create activity %s. Here's why: %s: %s", name, err.response["Error"]["Code"], err.response["Error"]["Message"], ) raise else: return response["activityArn"] def find(self, name): """ Find an activity by name. This requires listing activities until one is found with a matching name. :param name: The name of the activity to search for. :return: If found, the ARN of the activity; otherwise, None. """ try: paginator = self.stepfunctions_client.get_paginator("list_activities") for page in paginator.paginate(): for activity in page.get("activities", []): if activity["name"] == name: return activity["activityArn"] except ClientError as err: logger.error( "Couldn't list activities. Here's why: %s: %s", err.response["Error"]["Code"], err.response["Error"]["Message"], ) raise def get_task(self, activity_arn): """ Gets task data for an activity. When a state machine is waiting for the specified activity, a response is returned with data from the state machine. When a state machine is not waiting, this call blocks for 60 seconds. :param activity_arn: The ARN of the activity to get task data for. :return: The task data for the activity. """ try: response = self.stepfunctions_client.get_activity_task( activityArn=activity_arn ) except ClientError as err: logger.error( "Couldn't get a task for activity %s. Here's why: %s: %s", activity_arn, err.response["Error"]["Code"], err.response["Error"]["Message"], ) raise else: return response def send_task_success(self, task_token, task_response): """ Sends a success response to a waiting activity step. A state machine with an activity step waits for the activity to get task data and then respond with either success or failure before it resumes processing. :param task_token: The token associated with the task. This is included in the response to the get_activity_task action and must be sent without modification. :param task_response: The response data from the activity. This data is received and processed by the state machine. """ try: self.stepfunctions_client.send_task_success( taskToken=task_token, output=task_response ) except ClientError as err: logger.error( "Couldn't send task success. Here's why: %s: %s", err.response["Error"]["Code"], err.response["Error"]["Message"], ) raise def delete(self, activity_arn): """ Delete an activity. :param activity_arn: The ARN of the activity to delete. """ try: response = self.stepfunctions_client.delete_activity( activityArn=activity_arn ) except ClientError as err: logger.error( "Couldn't delete activity %s. Here's why: %s: %s", activity_arn, err.response["Error"]["Code"], err.response["Error"]["Message"], ) raise else: return response

Actions

The following code example shows how to use CreateActivity.

SDK for Python (Boto3)
Note

There's more on GitHub. Find the complete example and learn how to set up and run in the AWS Code Examples Repository.

class Activity: """Encapsulates Step Function activity actions.""" def __init__(self, stepfunctions_client): """ :param stepfunctions_client: A Boto3 Step Functions client. """ self.stepfunctions_client = stepfunctions_client def create(self, name): """ Create an activity. :param name: The name of the activity to create. :return: The Amazon Resource Name (ARN) of the newly created activity. """ try: response = self.stepfunctions_client.create_activity(name=name) except ClientError as err: logger.error( "Couldn't create activity %s. Here's why: %s: %s", name, err.response["Error"]["Code"], err.response["Error"]["Message"], ) raise else: return response["activityArn"]
  • For API details, see CreateActivity in AWS SDK for Python (Boto3) API Reference.

The following code example shows how to use CreateStateMachine.

SDK for Python (Boto3)
Note

There's more on GitHub. Find the complete example and learn how to set up and run in the AWS Code Examples Repository.

class StateMachine: """Encapsulates Step Functions state machine actions.""" def __init__(self, stepfunctions_client): """ :param stepfunctions_client: A Boto3 Step Functions client. """ self.stepfunctions_client = stepfunctions_client def create(self, name, definition, role_arn): """ Creates a state machine with the specific definition. The state machine assumes the provided role before it starts a run. :param name: The name to give the state machine. :param definition: The Amazon States Language definition of the steps in the the state machine. :param role_arn: The Amazon Resource Name (ARN) of the role that is assumed by Step Functions when the state machine is run. :return: The ARN of the newly created state machine. """ try: response = self.stepfunctions_client.create_state_machine( name=name, definition=definition, roleArn=role_arn ) except ClientError as err: logger.error( "Couldn't create state machine %s. Here's why: %s: %s", name, err.response["Error"]["Code"], err.response["Error"]["Message"], ) raise else: return response["stateMachineArn"]

The following code example shows how to use DeleteActivity.

SDK for Python (Boto3)
Note

There's more on GitHub. Find the complete example and learn how to set up and run in the AWS Code Examples Repository.

class Activity: """Encapsulates Step Function activity actions.""" def __init__(self, stepfunctions_client): """ :param stepfunctions_client: A Boto3 Step Functions client. """ self.stepfunctions_client = stepfunctions_client def delete(self, activity_arn): """ Delete an activity. :param activity_arn: The ARN of the activity to delete. """ try: response = self.stepfunctions_client.delete_activity( activityArn=activity_arn ) except ClientError as err: logger.error( "Couldn't delete activity %s. Here's why: %s: %s", activity_arn, err.response["Error"]["Code"], err.response["Error"]["Message"], ) raise else: return response
  • For API details, see DeleteActivity in AWS SDK for Python (Boto3) API Reference.

The following code example shows how to use DeleteStateMachine.

SDK for Python (Boto3)
Note

There's more on GitHub. Find the complete example and learn how to set up and run in the AWS Code Examples Repository.

class StateMachine: """Encapsulates Step Functions state machine actions.""" def __init__(self, stepfunctions_client): """ :param stepfunctions_client: A Boto3 Step Functions client. """ self.stepfunctions_client = stepfunctions_client def delete(self, state_machine_arn): """ Delete a state machine and all of its run data. :param state_machine_arn: The ARN of the state machine to delete. """ try: response = self.stepfunctions_client.delete_state_machine( stateMachineArn=state_machine_arn ) except ClientError as err: logger.error( "Couldn't delete state machine %s. Here's why: %s: %s", state_machine_arn, err.response["Error"]["Code"], err.response["Error"]["Message"], ) raise else: return response

The following code example shows how to use DescribeExecution.

SDK for Python (Boto3)
Note

There's more on GitHub. Find the complete example and learn how to set up and run in the AWS Code Examples Repository.

def describe_run(self, run_arn): """ Get data about a state machine run, such as its current status or final output. :param run_arn: The ARN of the run to look up. :return: The retrieved run data. """ try: response = self.stepfunctions_client.describe_execution( executionArn=run_arn ) except ClientError as err: logger.error( "Couldn't describe run %s. Here's why: %s: %s", run_arn, err.response["Error"]["Code"], err.response["Error"]["Message"], ) raise else: return response
  • For API details, see DescribeExecution in AWS SDK for Python (Boto3) API Reference.

The following code example shows how to use DescribeStateMachine.

SDK for Python (Boto3)
Note

There's more on GitHub. Find the complete example and learn how to set up and run in the AWS Code Examples Repository.

class StateMachine: """Encapsulates Step Functions state machine actions.""" def __init__(self, stepfunctions_client): """ :param stepfunctions_client: A Boto3 Step Functions client. """ self.stepfunctions_client = stepfunctions_client def describe(self, state_machine_arn): """ Get data about a state machine. :param state_machine_arn: The ARN of the state machine to look up. :return: The retrieved state machine data. """ try: response = self.stepfunctions_client.describe_state_machine( stateMachineArn=state_machine_arn ) except ClientError as err: logger.error( "Couldn't describe state machine %s. Here's why: %s: %s", state_machine_arn, err.response["Error"]["Code"], err.response["Error"]["Message"], ) raise else: return response

The following code example shows how to use GetActivityTask.

SDK for Python (Boto3)
Note

There's more on GitHub. Find the complete example and learn how to set up and run in the AWS Code Examples Repository.

class Activity: """Encapsulates Step Function activity actions.""" def __init__(self, stepfunctions_client): """ :param stepfunctions_client: A Boto3 Step Functions client. """ self.stepfunctions_client = stepfunctions_client def get_task(self, activity_arn): """ Gets task data for an activity. When a state machine is waiting for the specified activity, a response is returned with data from the state machine. When a state machine is not waiting, this call blocks for 60 seconds. :param activity_arn: The ARN of the activity to get task data for. :return: The task data for the activity. """ try: response = self.stepfunctions_client.get_activity_task( activityArn=activity_arn ) except ClientError as err: logger.error( "Couldn't get a task for activity %s. Here's why: %s: %s", activity_arn, err.response["Error"]["Code"], err.response["Error"]["Message"], ) raise else: return response
  • For API details, see GetActivityTask in AWS SDK for Python (Boto3) API Reference.

The following code example shows how to use ListActivities.

SDK for Python (Boto3)
Note

There's more on GitHub. Find the complete example and learn how to set up and run in the AWS Code Examples Repository.

class Activity: """Encapsulates Step Function activity actions.""" def __init__(self, stepfunctions_client): """ :param stepfunctions_client: A Boto3 Step Functions client. """ self.stepfunctions_client = stepfunctions_client def find(self, name): """ Find an activity by name. This requires listing activities until one is found with a matching name. :param name: The name of the activity to search for. :return: If found, the ARN of the activity; otherwise, None. """ try: paginator = self.stepfunctions_client.get_paginator("list_activities") for page in paginator.paginate(): for activity in page.get("activities", []): if activity["name"] == name: return activity["activityArn"] except ClientError as err: logger.error( "Couldn't list activities. Here's why: %s: %s", err.response["Error"]["Code"], err.response["Error"]["Message"], ) raise
  • For API details, see ListActivities in AWS SDK for Python (Boto3) API Reference.

The following code example shows how to use ListStateMachines.

SDK for Python (Boto3)
Note

There's more on GitHub. Find the complete example and learn how to set up and run in the AWS Code Examples Repository.

Find a state machine by name by searching the list of state machines for the account.

class StateMachine: """Encapsulates Step Functions state machine actions.""" def __init__(self, stepfunctions_client): """ :param stepfunctions_client: A Boto3 Step Functions client. """ self.stepfunctions_client = stepfunctions_client def find(self, name): """ Find a state machine by name. This requires listing the state machines until one is found with a matching name. :param name: The name of the state machine to search for. :return: The ARN of the state machine if found; otherwise, None. """ try: paginator = self.stepfunctions_client.get_paginator("list_state_machines") for page in paginator.paginate(): for state_machine in page.get("stateMachines", []): if state_machine["name"] == name: return state_machine["stateMachineArn"] except ClientError as err: logger.error( "Couldn't list state machines. Here's why: %s: %s", err.response["Error"]["Code"], err.response["Error"]["Message"], ) raise
  • For API details, see ListStateMachines in AWS SDK for Python (Boto3) API Reference.

The following code example shows how to use SendTaskSuccess.

SDK for Python (Boto3)
Note

There's more on GitHub. Find the complete example and learn how to set up and run in the AWS Code Examples Repository.

class Activity: """Encapsulates Step Function activity actions.""" def __init__(self, stepfunctions_client): """ :param stepfunctions_client: A Boto3 Step Functions client. """ self.stepfunctions_client = stepfunctions_client def send_task_success(self, task_token, task_response): """ Sends a success response to a waiting activity step. A state machine with an activity step waits for the activity to get task data and then respond with either success or failure before it resumes processing. :param task_token: The token associated with the task. This is included in the response to the get_activity_task action and must be sent without modification. :param task_response: The response data from the activity. This data is received and processed by the state machine. """ try: self.stepfunctions_client.send_task_success( taskToken=task_token, output=task_response ) except ClientError as err: logger.error( "Couldn't send task success. Here's why: %s: %s", err.response["Error"]["Code"], err.response["Error"]["Message"], ) raise
  • For API details, see SendTaskSuccess in AWS SDK for Python (Boto3) API Reference.

The following code example shows how to use StartExecution.

SDK for Python (Boto3)
Note

There's more on GitHub. Find the complete example and learn how to set up and run in the AWS Code Examples Repository.

class StateMachine: """Encapsulates Step Functions state machine actions.""" def __init__(self, stepfunctions_client): """ :param stepfunctions_client: A Boto3 Step Functions client. """ self.stepfunctions_client = stepfunctions_client def start(self, state_machine_arn, run_input): """ Start a run of a state machine with a specified input. A run is also known as an "execution" in Step Functions. :param state_machine_arn: The ARN of the state machine to run. :param run_input: The input to the state machine, in JSON format. :return: The ARN of the run. This can be used to get information about the run, including its current status and final output. """ try: response = self.stepfunctions_client.start_execution( stateMachineArn=state_machine_arn, input=run_input ) except ClientError as err: logger.error( "Couldn't start state machine %s. Here's why: %s: %s", state_machine_arn, err.response["Error"]["Code"], err.response["Error"]["Message"], ) raise else: return response["executionArn"]
  • For API details, see StartExecution in AWS SDK for Python (Boto3) API Reference.

Scenarios

The following code example shows how to create an AWS Step Functions messenger application that retrieves message records from a database table.

SDK for Python (Boto3)

Shows how to use the AWS SDK for Python (Boto3) with AWS Step Functions to create a messenger application that retrieves message records from an Amazon DynamoDB table and sends them with Amazon Simple Queue Service (Amazon SQS). The state machine integrates with an AWS Lambda function to scan the database for unsent messages.

  • Create a state machine that retrieves and updates message records from an Amazon DynamoDB table.

  • Update the state machine definition to also send messages to Amazon Simple Queue Service (Amazon SQS).

  • Start and stop state machine runs.

  • Connect to Lambda, DynamoDB, and Amazon SQS from a state machine by using service integrations.

For complete source code and instructions on how to set up and run, see the full example on GitHub.

Services used in this example
  • DynamoDB

  • Lambda

  • Amazon SQS

  • Step Functions

The following code example shows how to build and orchestrate generative AI applications with Amazon Bedrock and Step Functions.

SDK for Python (Boto3)

The Amazon Bedrock Serverless Prompt Chaining scenario demonstrates how AWS Step Functions, Amazon Bedrock, and https://docs.aws.amazon.com/bedrock/latest/userguide/agents.html can be used to build and orchestrate complex, serverless, and highly scalable generative AI applications. It contains the following working examples:

  • Write an analysis of a given novel for a literature blog. This example illustrates a simple, sequential chain of prompts.

  • Generate a short story about a given topic. This example illustrates how the AI can iteratively process a list of items that it previously generated.

  • Create an itinerary for a weekend vacation to a given destination. This example illustrates how to parallelize multiple distinct prompts.

  • Pitch movie ideas to a human user acting as a movie producer. This example illustrates how to parallelize the same prompt with different inference parameters, how to backtrack to a previous step in the chain, and how to include human input as part of the workflow.

  • Plan a meal based on ingredients the user has at hand. This example illustrates how prompt chains can incorporate two distinct AI conversations, with two AI personas engaging in a debate with each other to improve the final outcome.

  • Find and summarize today's highest trending GitHub repository. This example illustrates chaining multiple AI agents that interact with external APIs.

For complete source code and instructions to set up and run, see the full project on GitHub.

Services used in this example
  • Amazon Bedrock

  • Amazon Bedrock Runtime

  • Amazon Bedrock Agents

  • Amazon Bedrock Agents Runtime

  • Step Functions