The All Voice

Python package for handling requests from Amazon and API.AI. Supoorts Alexa, Google Home, Google Assistant, Slack and Web Apps.

Installation

pip install git+https://github.com/TheLampshady/all_voice.git

Usage

The AllVoice Base Skill allows a developer to extend a single class and inherit the functionality of the skill associated with the service making the call. By uses dynamic parenting from different types of base skill (e.g. Alexa, Google Home) the request and response are handling implicitly while providing support for multiple devices.

The developer can extend one class and just focus on developing actions with easy access to slots/parameters or session attributes. With session support, the Amazon attribute is used and a default case is added for API.AI

#TODO Add Alexa context support.

#TODO Add Alexa yes support to handle context of the question being answered

Creating a Skill

Developers can create there skill by extending the AllVoice and adding their Intents as functions.

from all_voice.models import AllVoice

class TalkToMeSkill(AllVoice):
    """My Skill"""

    def AwesomeIntent(self):
        """My Intent/Action"""
        return self.build_response("I am awesome.")

Explicit Skills

Developers can also use a single skill class such as just Alexa

from all_voice.models import AlexaSkill

class TalkToMeSkill(AlexaSkill):
    """My Skill"""

    def AwesomeIntent(self):
        """My Intent/Action"""
        return self.build_response("I am awesome.")

User Objects

User objects are created by extending the BaseUser class. By default an AllVoice user provides simple logging powered by a python dict. Do not rely on this for webhooks that scale.

from all_voice.models import BaseUser

class UserClass(BaseUser): pass

Once created, the class can be passes a a parameter to your skills init.

def webhook(event):
    skill = TalkToMeSkill(event, UserClass)

Testing

Developers can extend the unittest class and utils for generating JSON events fomr various services.

import unittest
from all_voice.tests.utils import AllVoiceTestUtils

class TestBaseIntent(AllVoiceTestUtils, unittest.TestCase):

    def setUp(self):
        super(TestBaseIntent, self).setUp()

Caveats

Since the parents are dynamic, references to class fields / methods will be correct but IDEs will reference the unimplemented code and not the dynamic classes (Alexa / Google Home)

Hosting

Fulfillments can be hosted on various platforms with this library.

App Engine

This exmaple will go through hosting your webhook in Google App Engine with Python an WebApp2 Review this link for a quickstart into google app engine. Google App Engine Quickstart

1. Create your

   • config app.yaml
   • application webapp2.WSGIApplication
   • handler class class Webhook(BaseHandler)

2. Link your app.yaml to the webapp2.WSGIApplication

- url: .*
  script: routes._APP
  secure: always

3. Add a route to your web hook handler

   • webapp2.Route("/webhook", handler=Webhook, name='webhook'),

4. Add your skill to handler

import json 
# ....
def post(self):
    event = json.loads(self.request.body)
    
    request = TalkToMeSkill(event)
    response = json.dumps(request.response())
    
    self.response.headers['Content-Type'] = 'application/json'
    return self.response.out.write(response)

5. Install this library for upload pip install -t libs git+https://github.com/TheLampshady/all_voice.git

Lambda

This example will provide the code for hosting a webhook on lambda functions. A brief tutorial on creating a lmbda function Alexa on Lambda Function

1. Create a a function that takes the parameters below in a file called. lambda_function.py

from my_skill import TalkToMeSkill

def lambda_handler(event, context={}):
    return TalkToMeSkill(event=event).response()

2. Install this library for upload pip install -t libs git+https://github.com/TheLampshady/all_voice.git