lexstream

lexstream

Alexa Music skill for streaming your own music.

Documentation links

Various documentation links that have been useful when building out this project

• how-a-music-skill-works

Prerequisites

• Go
• Node.js
  • Alexa Skills Kit requires node --version v8.3 or higher.
• AWS Account
  • Configuration and credential file settings
  • These credentials need to have permissions to change S3, IAM, DynamoDB, and Lambda resources
• AWS CLI
• Alexa Developer Console
  • Validate that you can sign in

Initialize Go Modules

• Use go mod tidy in the root of this repository to initialize dependencies

Run preflight check

• go run cmd/preflight.go will:
  • Check that the minimum version of node is installed.
  • Configure and test ask-cli in the .ask directory.
    • This will prompt to run ask configure if no credentials are found
  • Check that AWS credentials and configuration files are set up correctly.
    • This performs read only api calls for S3, IAM, DynamoDB, and Lambda
  • Check that the aws cli command is installed

Warning commands after this point start modifying things, so read warnings closely

Generate Login with Amazon Keys

• Create a new security profile for your Amazon Developer account by following the instructions provided here
• This will generate Client ID and Client Secret keys.
• Save these keys in the file secret.json in this format

  {
    "client_id": "amzn1.application-oa2-client.xxxxx",
    "client_secret": "xxxxxx"
  }
  

• Run go run smapi/lwa.go - this automates a call to ask util generate-lwa-tokens
  • This must be run on a computer with a browser to handle the LWA oauth flow
• The resulting token will be saved to token.json and updated by the token source using the LWA credentials when it expires every hour.

Deploy Skill

TODO - smapi api client has been generated, automation work needed

Deploy Lambda

• Copy settings.json.example to settings.json and update with the appropriate content

WARNING if the provided bucket already exists the permissions will be re-set to no public access

• Run the following command to create the required Bucket, IAM Role, and Lambda

  $ go run infra/sync.go
  

Update Endpoint and Re-Validate

TODO: this should be able to be automated

• Back in the Alexa skills console, update the skill endpoint with the output of the previous command

Design Notes

• Use MusicBrainz_API for identifiers wherever possible.
  • Completed
  • Artist,Album, and Track are all directly supported and would be unique
  • Prefix each since the MB Api uses GUIDs that are opaque

TODOs and other ramblings

• Right now this just plays a single song on loop

• Create a web interface / cli to upload new songs

  • upload audio files and cover art to s3
  • extract metadata from id tags in audio and create catalog files
  • automate upload of catalog data
    • The ask cli does a browser login to get the access token so this may not be possible
    • more research
    • This is possible using LWA tokens and calling the upload catalog api directly
    • Add documentation on how this works

• Metadata notes showing id -> api mappings

  • MUSICBRAINZ_ARTISTID
  • MUSICBRAINZ_ALBUMARTISTID
  • MUSICBRAINZ_TRACKID
  • MUSICBRAINZ_ALBUMID
    • This ID can be used on the cover art archive
    • Redirects to the internet archive
  • MUSICBRAINZ_RELEASEGROUPID
  • MUSICBRAINZ_WORKID
  • MUSICBRAINZ_RELEASETRACKID=a022f133-009d-3e9f-b0c1-d2e30e86e98e
    • I don't think this one can be directly queried, but it shows up under /media/tracks/id in MUSICBRAINZ_ALBUMID
    • It can also be used as an anchor for the release page

• Saving whipper command whipper cd rip -p

Initial readme created by tigwen