A Google Music Alexa Skill
This is a rewrite of the popular GeeMusic Alexa Skill from scratch. It borrows heavily from the original skill but is written entirely from scratch. Born out of a frustration with how complicated it was to get the original skill running and certain bugs that were noticed.
This project is still in development and subject to a lot of change. Basic functionality is working at this time. You are required to run this script on your own server. I personally recommend a cheap VPS or local machine which can be exposed to the web.
This skill is not made or endorsed by Google or the original makers of GeeMusic. This skill uses the gmusicapi by Simon Weber and is dependent on it as a library.
This sill is developed exclusively on an Amazon Echo Dot used in Canada set to English(US). I make no promises or guarentees of compatibility with other versions or setups.
This script is used by saying "Alexa, tell Geemusic to...". Some of the search terms use fuzzy string logic so feel free to experiment. I encourage looking at the sample utterances for samples of what to say.
Currently we have implemented:
Play song Never Gonna Give You Up
Play the song Radioactive by Imagine Dragons
Play music by The Wet Secrets
Play artist Metric
Play songs by Sam Roberts Band
Start genre radio for Indie
Play Rock radio
Start radio for genre Country
You can currently also say "Alexa pause" to stop playback and "Alexa resume" to resume playback.
A lot of things honestly. Probably get all of the existing functionality in the old GeeMusic skill implemented and then a bit more. I've currently implemented my #1 need which was to play a radio station as background music.
The following setup has been tested on Ubuntu 16.04 Xenial on a virtual private server. It should be functional on other UNIX type environments.
This server must be running at all times and the webserver involved must also be run at all times. That is outside the scope of this article, I recommend looking into options such as 'screen' or 'circus' for running jobs in the background with monitoring.
Begin by cloning the repository to your server:
$ git clone https://github.com/litobro/GoogMusic.gitEnsure you have Python 3 installed and cd into the directory. I highly recommend running this script inside of a virtual environment.
# Run this to create a virtual environment and activate it
$ python3 -m venv .
$ source bin/activate
# Continue here, or begin here to skip creating a virtual environment
$ pip3 install -r requirements.txtThis may take some time, allow the requirements to install, it is required to setup a configuration file after it has completed. Also in the root of the directory create a file called config.py using touch config.py. It should contain these variables:
# Google Credentials
GOOGLE_EMAIL='YOUR EMAIL HERE'
GOOGLE_PASSWORD='YOUR PASSWORD HERE'
ANDROID_ID='A VALID AND REGISTERED MOBILE_DEVICE ID HERE'
As with the original GeeMusic skill, I highly recommend you enable 2-factor authentication on your Google Account and insert an application specific password. The password will unfortunately at this time be stored in plaintext on your server.
Your Android_ID can be extracted from your google music webpage which shows registered devices. Simply inspect each device listed in your web browser and view the source, the ID will make itself evident in the list.
At this time, you can try running the server using python3 server.py. It should launch without issue using flask.
Open up the Alexa Dashboard, click "Get Started" in the Alexa Skills Kit box. Then click on the yellow "Add a New Skill" button in the top right corner.
Going through the various sections
| Field | Value |
|---|---|
| Skill Type | Custom Interaction Model |
| Language | Select US English |
| Name | GoogMusic |
| Invocation Name | gee music |
| Audio Player | Yes |
On the "Interaction Model" step, paste in the contents of speech_assets/intent_schema.json to the intent schema field and the contents of speech_assets/sample_utterances.txt to the sample utterances field.
We'll point our skill at the URL for our development server. Select HTTPS as the endpoint type and enter your server's URL in the corresponding box. Remember that your development server must be publicly accessible AND using HTTPS in order for Amazon to be able to connect/interact with it.
If your development server is running on a server that is already available on the internet, type its URL (such as https://geemusic.example.com/alexa). Make sure you include the /alexa, otherwise this won't work!
If you are running the server on a computer behind a firewall we'll need to expose the server via a tunnel in order for this to work. I usually use ngrok for these situations and have used it to develop this project. To start a tunnel run ngrok http 5001 in a console window. You should then see a few URLs, one of which being a publicly accessible HTTPS link to your development server. Copy this URL, being sure to append /alexa so the final result looks something like https://[some-code].ngrok.io/alexa.
You'll also want to select "No" for the "Account Linking" field before moving on.
If using ngrok or heroku select the second option "My development endpoint is a subdomain of a domain that has a wildcard certificate from a certificate authority."
Scroll down to the "Service Simulator" section, the check the Skill is talking to Alexa correcty enter the word help "help" then click "Ask Gee Music", and you'll ideally see some resulting JSON in the Service Response box. You can then try testing phrases like Play artist Arctic Monkeys
Do not fill these in and make sure you never click "SUBMIT FOR CERTIFICATION". You are NOT submitting the Skill to Amazon to include in the public Skill store.