pyasista 1.1.2

Arcelik Asista Smart Assistant AWS Lambda Python Library

Arcelik Asista Smart Assistant AWS Lambda Python Library

Simple Python API wrapper for Arcelik Asista Smart Assistant AWS Lambda handlers.

---

pyasista is an easy to use Python package to help developers implement their AWS lambda function with ease.

Installation

To install pyasista, simply use pip.

pip install pyasista

Usage

Just import pyasista into your project and use the wrapper classes for easily implementing your lambda functions.

from pyasista import AsistaInput
from pyasista import AsistaOutput
from pyasista import Command
from pyasista import CommandType


# Sample lambda handler to make Asista say `Hello, my name is Asista.`.
def lambda_handler(event, context):
    asista_input = AsistaInput(event)
    asista_output = AsistaOutput()
    command = Command(CommandType.PLAY_ANNOUNCE, 'Hello, my name is Asista.')
    asista_output.push_command(command)
    return dict(asista_output)

Documentation

AsistaInput

Arcelik Asista Smart Assistant Lambda function has a standard event format. AsistaInput class is a wrapper for the lambda function input. Just use the event of the lambda function while initializing the object.

...

def lambda_handler(event, context):
    asista_input = AsistaInput(event)
    ...

AsistaInput has the following properties and methods:

• customer_confirmed: True if the customer responded yes; False if the customer responded no; None if the customer has not responded to a Yes/No question.
• environment: Environment (DEVELOPMENT, TESTING or PRODUCTION). (See: Environment)
• ext_node_id: Node id of the detected NLP tree.
• serial_number: Serial number of the physical Asista device.
• session: A dictionary of the session variables.
• slot_values: A dictionary of detected dynamic variables. Slot values stored in lists, with the same order as it appeared in the event.
• sr_text: Actual speech detected with speech recognition.
• timeout: While waiting for user to respond your query, the Asista device timeouts after 10 seconds if it could not detect any speech. In that case, lambda function will be fed with timeout parameter is set to True.
• token: A helper object for holding OAuth2 token and refreshing the token. (See: AsistaToken)
• user_guide(): Retrieves the user guide of the skill. Returns None if the method encounters an error.

All properties of AsistaInput are immutable.

...

def lambda_handler(event, context):
    asista_input = AsistaInput(event)
    print(asista_input.serial_number)  # Print the serial number of the device
    asista_input.serial_number = 'QWERTY123'  # Raise an AttributeError

If your lambda strictly requires more of those parameters, you can pass them to the constructor and raise an error if the event lack any of those arguments.

...

def lambda_handler(event, context):
    # If serial_number and token keys are missing in the event, raise an error.
    asista_input = AsistaInput(event, required_parameters=['serial_number', 'token'])
    ...

AsistaOutput

A wrapper for Arcelik Asista Smart Assistant Lambda function standard output.

from pyasista import AsistaInput
from pyasista import AsistaOutput


def lambda_handler(event, context):
    asista_input = AsistaInput(event)
    asista_output = AsistaOutput()
    return dict(asista_output)

Due to the non-serializable structure of AsistaOutput, lambda functions should return dict(asista_output). [Under revision]

AsistaOutput has the following methods:

• push_command(command): Add a command for Asista to perform. (See: Command)
• session(dict): A dictionary containing the session information. [default=dict()]
• session_continue(bool): True if the session should continue; elseFalse. [default=False]

AsistaToken

For the applications requiring OAuth2 access token, Asista lambda input contains the access token and all the required information on refreshing them. You can access the token from the AsistaInput.

AsistaToken has the following property and methods:

• access_token: The property holds the actual access token. It is an immutable variable, trying to set it will raise an AttributeError.
• is_valid(): Checks whether the AsistaToken class structually valid. It is important to note that this method does not check whether the token itself is valid or not. It only checks whether the AsistaToken instance has access_token set and all the data present to refresh it.
• refresh(): Refreshs the access token. Return True if the token is refreshed successfully, else False.

Example usage of the AsistaToken as follows:

from pyasista import AsistaInput
from pyasista import AsistaToken


def lambda_handler(event, context):
    asista_input = AsistaInput(event)
    asista_token = asista_input.token  # The AsistaToken object

    if asista_token.is_valid():  # Checks access_token is structually valid
        print(asista_token.access_token)  # Print the access_token
        if asista_token.refresh():  # Refresh the access_token
            print(asista_token.access_token)  # Print the new access_token
        else:
            print('Cannot refresh access token')  # Error in refresh
    else:
        print('Event does not have the token object')

    ...

Command

Arcelik Asista Smart Assistant is capable of converting a text to speech using a TTS engine and play it out loud, and playing an audio stream for the customer. AsistaOutput has an array of commands for Asista to perform. Command is a wrapper class for this purpose.

Basically initialize the class with type (See: CommandType) and the data (either the text to be spoken or the stream to be streamed) and add it to the output using AsistaOutput.push_command(command) method.

An example usage of Command is below:

from pyasista import AsistaOutput
from pyasista import Command
from pyasista import CommandType


def lambda_handler(event, context):
    # Parse event using AsistaInput
    # Implement your business logic
    ...
    asista_output = AsistaOutput()

    # Create a command to make Asista say 'Hello, my name is Asista.'
    command = Command(CommandType.PLAY_ANNOUNCE, 'Hello, my name is Asista.')
    asista_output.push_command(command)
    return dict(asista_output)

CommandType

An enumeration of supported command types that Asista can perform:

• CommandType.PLAY_ANNOUNCE: Read a text out loud using TTS engine.
• CommandType.PLAY_STREAM: Stream an audio from a URL.
• CommandType.STOP: Stop Asista.
• CommandType.NOACTION: Make Asista take no action at all.

Environment

An enumeration of supported environments that current event of the lambda function is executed.

Supported environments are:

• Environment.DEVELOPMENT: Development environment.
• Environment.TESTING: Test environment.
• Environment.PROD: Production environment.

Accessing the environment is easy:

from pyasista import AsistaInput
from pyasista import Environment


def lambda_handler(event, context):
    asista_input = AsistaInput(event)
    environment = asista_input.environment
    if environment == Environment.DEVELOPMENT:
        # Do tasks for development
    elif environment == Environment.TESTING:
        # Do tasks for test
    elif environment == Environment.PROD:
        # Do tasks for production
    else:
        # Undetermined environment

EXAMPLES

Looking for more examples? Check out the examples folder.

Release History

1.1.2 (2018-11-22)

• user_guide() method is added to the AsistaInput.

1.1.1 (2018-10-22)

• timeout property is added to the AsistaInput.

1.1.0 (2018-09-25)

• Environment string for Environment.PROD changed from PROD to PRODUCTION. If you are using Environment you should update your pyasista version to keep things working.

1.0.1 (2018-08-13)

• Following static methods are added to AsistaOutput to create lambda outputs easier:
  • AsistaOutput.with_stream(data): Creates an instance of AsistaOutput with command PLAY_STREAM.
  • AsistaOutput.with_announce(data): Creates an instance of AsistaOutput with command PLAY_ANNOUNCE.
  • AsistaOutput.with_stop(): Creates an instance of AsistaOutput with command STOP.
  • AsistaOutput.with_noaction(): Creates an instance of AsistaOutput with command NOACTION.

1.0.0 (2018-08-10)

• Birth!