cteodorski / pylexa

Python library for building an Alexa Skills Kit

pylexa

pylexa is a library that aims to ease development of an Alexa Skills Kits.

pylexa allows one to define a simple Flask application that will be able to accept requests and return appropriate responses to the Alexa service.

Example

Let's say you want to define an Alexa Skill that echoes whatever the user says. So far, you've:

• created a skill in the Amazon Developer Console

• Added an Echo intent with a single slot, message:

  {
    "intent": "Echo",
    "slots": [
      {
        "name": "message",
        "type": "AMAZON.LITERAL"
      }
    ]
  }

• Added an utterance to allow users to interact with the skill:

  Echo echo { something | message }

Now, you're ready to create a server that will accept the request and return a response echoing the input. Using pylexa, we'd need only the following code to accomplish this:

from flask import Flask

from pylexa.app import alexa_blueprint
from pylexa.intent import handle_intent
from pylexa.response import PlainTextSpeech


app = Flask(__name__)
app.config['app_id'] = 'amzn1.echo-sdk-ams.app.my_app_id'
app.register_blueprint(alexa_blueprint)


@handle_intent('Echo')
def handle_echo_intent(request):
    return PlainTextSpeech(request.slots.get('message', 'Nothing to echo'))

And that's it! You can push the above code, configure the skill to point to the server running the flask app and use the service simulator to test your skill.

Testing

After installing requirements with pip install -r requirements.pip, tests can be run with nosetests.

Alexa Configuration

Configuration of an Alexa Skill is done in three parts in the developer console:

• the intent schema
• list of utterances
• custom slot definitions

pylexa comes with a command line tool that aims to simplify this configuration by allowing one to define a YAML file with the necessary information and generate the intent schema, utterances, and custom slots from that.

For example, let's say you had the following YAML schema defined:

intents:
    - TestIntent:
        foo: AMAZON.NUMBER
        bar: CUSTOM_SLOT
    - OtherIntent
    - AMAZON.YesIntent

utterances:
    TestIntent:
        - 'do something with {foo} and {bar}'
        - '{foo} {bar}'
    OtherIntent:
        - 'do something else'

slots:
    CUSTOM_SLOT:
        - value 1
        - value 2

This defines a skill that handles three intents (TestIntent, OtherIntent, and AMAZON.YesIntent), specifies utterances for TestIntent and OtherIntent, and contains a custom slot definition.

If we have that YAML definition in conf/schema.yml:

$ tree conf
conf
└── schema.yml

0 directories, 1 file

We can then run the command line tool generate-alexa-conf to create the requisite files:

$ generate-alexa-conf conf/schema.yml

We now have the intent schema, utterances, and slots defined in their own files:

$ tree conf
conf
├── intent_schema.json
├── schema.yml
├── slots
│   ├── CUSTOM_SLOT
└── utterances.txt

1 directory, 4 files

The contents of each file can then be copied + pasted in to the appropriate sections of the Alexa Skill configuration.