cscanlin / alexa-casatunes

Control CasaTunes with Alexa

alexa-casatunes

Control CasaTunes with Alexa

Available Commands

• Play / Pause
• Previous / Next Song
• Turn Room On / Off
• Set Volume in Room
• Increase / Decrease Volume in Room
• Now Playing Info
• Search for Track / Artist / Album / Playlist

Planned Commands (With Conversational Support)

• Wake Alarms / Sleep Timers
• Better Search (source selection + conversational)
• Playlist / Queue Management

Deploying

In order to start you must have a CasaTunes server (obviously), and it must have Windows 10 installed on it.

This project is still in the early stages, and the deployment process is still very manual / unpolished for new users. If you get stuck or have any feedback, please open an issue!

Also, I'm working on building this out so it can be properly deployed as a real Alexa skill that can be installed by any user. If you're interested in contributing, please leave a comment here!

CasaTunes Server Setup

1. Install Bash on Windows by following the instructions here: https://msdn.microsoft.com/en-us/commandline/wsl/install_guide

2. Setup SSH on your CasaTunes server, following the instructions here: https://superuser.com/a/1114162/401859

   • Note that you should not use the default port (22), use a different unused port instead (22222 recommended). Record which port you use, as it will be need later

   • You will also create a password at some point in the setup. Remember this as well.

3. Forward the port above through your router.

4. Create an ssh key called casa_rsa and add the public key to the server in the authorized_keys file (usually found in C:\Users\<user>\.ssh\)

5. Upload the casa_rsa private key to a directory called keys in your s3 bucket.

Alexa / Lambda Setup

1. Fork / clone this repository

2. Run mapping_generator.py MY_LOCAL_CASA_IP, using your local CasaTunes IP address

3. Change the s3_bucket name in zappa_settings.json to one you control.

4. Install Docker if not already installed

5. Build the docker image:

   docker build -t zappa .

6. Deploy to AWS Lambda using Zappa

   sh deploy_zappa.sh -d # you only need the `-d` flag for initial deploy

   • This will take a few minutes to run; record the url generated after it finishes

7. Go to https://developer.amazon.com/alexa-skills-kit and create a new Alexa skill

8. Update configuration to use https, with the path generated in the command line from step 6

9. In the interaction model builder, navigate to code editor (top of the left side menu) and paste everything from interaction_model.json

10. Set environment variables on Lambda in the "Code" tab (and locally if desired) for the following:

    • ALEXA_USER_ID - can be found in the "Test" portion of the Alexa skill dashboard using the service simulator. Multiple user ids can be added, separated by a semicolon
    • CASA_SERVER_IP - Your public IP address, can be found by googling "what is my ip"
    • CASA_SERVER_PORT - From earlier ssh setup (22222 recommended)
    • CASA_S3_BUCKET_NAME - Your s3 bucket (likely same as in zappa_settings.json)
    • CASA_SSH_PASSWORD - From earlier ssh setup

alexa-casatunes's Issues

Deploy Skill Publicly

A place to discuss ideas for releasing this skill publicly.

Necessary components:

1. The Alexa Skill - The interaction model must be flexible for rooms from all users.

2. The Lambda Function - must be able to handle account linking, tokens, and authentication

3. The CasaTunes server will need to run software so it is able to authenticate a request/token received from Lambda, and route the request.

For part one, room names are currently stored in the interaction model, but it doesn't appear to really make a difference in validating what rooms names are parsed from the Users request. If that continues to be the case, this part can be considered pretty much done.

I found this blog post which demonstrates how to accomplish part 2 in Rails with a gem called doorkeeper: https://echobind.com/blog/alexa-custom-skills-with-account-linking/

I'm working on a proof of concept application that leverages Flask-OAuthlib to try to get basic account linking working, but have been unsuccessful so far (See the README in that repo for a description of the problem):
https://github.com/cscanlin/ask-auth

I haven't started part 3 yet, but it should be relatively straightforward once the second part is done.

If anyone has any ideas on ways to proceed (especially on part 2) please let me know!