Introduction

This skill provides some Trello commands that I found useful. It will eventually be expanded to support more commands and use cases.

[BETA WARNING] This is currently in heavy development; most important it currently does not support account linking, thus only interacts with 1 Trello account as defined in the environment variables (see Deployment below).

Features

Below listed features either implemented or planned:

• List boards
• Get / Set board as default
• (TODO) List available lists on board X / default board
• (TODO) List items in list Y on board X / default board
• (TODO) Account Link to Trello OAuth 1.0; see:
  • http://code-coverage.net/building-amazon-alexa-skills-with-net/
  • https://developer.amazon.com/docs/custom-skills/link-an-alexa-user-with-a-user-in-your-system.html
• (TODO) Integrate with Alexa TODO / Shopping lists:
  • https://developer.amazon.com/docs/custom-skills/access-the-alexa-shopping-and-to-do-lists.html
• (TODO) Fully tested

Setup Skill

Use the UI wizard to create the skill through this page:

https://developer.amazon.com/edw/home.html#/skills

We could also create the skill with the ask cli tool:

ask new [-n|--skill-name <name>] [-p| --profile <profile>] [--lambda-name <lambda-name>]

Updating the interaction model

The interaction model can be found in model.json. It defines the invocation name, intents and slots as well as custom types.

This file should be uploaded for each language in the Skill builder.

We can also update the model from the ask cli tool:

# first install the tool from npm and init credentials
npm i -g ask-cli
ask init

ask api update-model -f model.json -s <skill-id> -l <locale>

To update all known languages of skill with $SKILL_ID (must be set in env), use the helper script:

cd trello_skill/
heroku run bash update_model.sh

Deployment

Personally I found it easiest to use heroku, (however any host would work just as fine).

Using heroku means that dependencies are specified in the Pipfile (i.e use pipenv), and certain environment variables must be set in heroku config.

Requirements

The project was developed using new features from Python 3.6, so this is the only supported version, and there are no plans to backport.

• alexandra
• py-trello
• python-dotenv
• gunicorn
• SQLAlchemy
• alembic
• psycopg2

Environment Variables

Configure at least the values below as required:

• TRELLO_API_KEY: Trello API application key
• SKILL_ID: which skill to update when running update_models.py
• DATABASE_URL: Database DMN

To find your TRELLO_API_KEY visit https://trello.com/app-key

For the SKILL_ID, visit your alexa skill page at https://developer.amazon.com/edw/home.html

DATABASE_URL is normally set by heroku-postgres addon in a deployed environment. In development, we default to sqlite:///alexa_trello_skill.db if the variable is unset. Change this to whatever DB or driver you want to run. Follow the SQLAlchemy url documentation to learn more.

Database Setup

Certain information, such as tokens and other user preferences, needs to be stored persistently. The data requirements should be pretty modest, thus a free tier Heroku Postgres addon instance should suffice.

Create Database

When managing our own database, i.e during development, we need to start by creating a database.

$ sudo su - postgres
(postgres) $ psql

> CREATE USER test_alexa WITH PASSWORD '<password>';
> CREATE DATABASE test_alexa OWNER test_alexa;

Database Schema Migrations

We use the alembic package to maintain migrations. Here are some common commands we normally need to run.

Run Migrations

To migrate the DB schema to latest revision (i.e head), run upgrade:

heroku run alembic upgrade head

Create Migration

Create a new migration with the revision command:

alembic revision --autogenerate

Add a Trello token for an alexa user to the Database

We have yet to implement the authentication link to allow alexa users to connect the skill to Trello OAuth 1.0. Thus, we need to manually add users to the application to be able to access a Trello account.

This is easy to done in shell:

# open interactive python shell (if in dev)
pipenv run ipython

# Or, if in production, use heroku's python
heroku run python

>>> from trello_skill.utils import trello_client, save_user_token, setup_tokens
>>> user_id = 'amzn1.ask.account.AABBCC...'  # your alexa ID

>>> # Check if token for user already saved
>>> client = trello_client(user_id=user_id)
AssertionError: User "amzn1.ask.account.AABBCC..." has no known token (OAuth not yet implemented)!

>>> # if the error is raised, we need to save it
>>> token = '4534534...'
>>> api_key, token_map = setup_tokens()
>>> save_user_token(user_id, api_key, token)

Development

Using dotenv with a .env file

Create a .env file with e.g the following, to easily populate the command's environment when running locally:

SKILL_ID=<your-alexa-skill-ID>  # obtain in skill mgmt page
DATABASE_URL=<DMN-to-local-DB-instance>
TRELLO_API_KEY=<a-trello-API-key>

Using pipenv

When running code locally, start by setting up the python environment.

pip install pipenv
pipenv install -d  # install addl packages e.g ipython

The above command may offer to install python 3.6.2 if it's not currently installed and your system has pyenv.

Later, prepend pipenv to any command you'd like to run. Here are some examples

# if you didn't install the package in editable mode, set your python path
export PYTHONPATH=.

# make migration
pipenv run alembic revision --autogenerate

# run migration
pipenv run alembic upgrade head

# run gunicorn webserver
pipenv run gunicorn trello_skill:wsgi

# run debug webserver
pipenv run python trello_skill.py

# update alexa interaction model
pipenv run bash trello_skill/update_model.sh

# run interactive sysmtem shell
pipenv shell

# run interactive python shell (if dev deps installed)
pipenv run ipython