wdlindmeier / magpiradio

Twitter Radio for Raspberry Pi

Magpi Radio

A Twitter Radio for Raspberry Pi and Arduino. Featured on Hack-a-day, Adafruit and Make!

A Brief Overview: http://itp.nyu.edu/~wdl225/work/?p=275

Some Technical Details: http://itp.nyu.edu/~wdl225/work/?p=286

The Twitter Stream: http://twitter.com/magpiradio

Requirements

Magpi Radio was designed to run on a Raspberry Pi running Debian Wheezy, but it may also work on OS X and other systems.

Software requirements:

• SDL library (SMPEG must be installed for mp3 playback): http://www.libsdl.org/
• SDL_mixer library. This is the SDL module that controls audio playback.
• ruby (tested using ruby 1.9.3)
• The ruby-sdl-ffi gem: http://github.com/jacius/ruby-sdl-ffi
• The twitter ruby gem: http://sferik.github.com/twitter/
• The tweetstream gem: http://github.com/intridea/tweetstream
• ffmpeg (optionally for the OS X say server): http://www.ffmpegx.com/

API requirements:

• A Google API Key. Enable Google Maps Geolocation API in your console. (https://code.google.com/apis/console)
• Twitter API OAuth Credentials. Read and write permissions should be enabled.

Configure

1. Once you clone the source, copy the file api_config_sample.rb to api_config.rb. Update the value of the global variables with your API credentials and Twitter account name.

2. Create a "tweets" folder within audio/. This is where the network audio files will be stored. $ mkdir audio/tweets

3. Update Magpi_Content::configure to use your own list channels. This could also be handled in a config file, or a channel could be created from all of the account lists. Suggestions and pull-requests are welcome.

4. Upload the Arduino code to your Uno.

5. Connect your hardware components based on the system diagram found here: http://itp.nyu.edu/~wdl225/work/wp-content/uploads/2012/12/system_diagram.jpg

6. (Optional) Install the say.rb script on your web-server. I've dropped it into /Library/WebServer/CGI-Executables on my Mac with Apache running. This will allow you to use your own machine to generate the audio files. The tts will default to using tts-api.com if you don't specify one. If tts-api is unavailable, the script will try to use espeak on the Pi.

Running

Without a custom text-to-speech server:

$ ./radio

With a custom text-to-speech server (e.g. using the say.rb script on 10.0.1.4):

$ ./radio.rb "http://10.0.1.4/cgi-bin/say.rb?q=__QUERY__&v=__VOICE__"

When you run the script, you may get these SDL warnings:

Warning: Could not load SDL_image. Warning: Could not load SDL_ttf. Warning: Could not load SDL_gfx.

These can be ignored, because Magpi Radio doesn't use the graphics and imaging components of SDL.

Twitter Interface

If the radio is tuned to the "Any Requests?" channel, you can change the search query with the following tweet:

@magpiradio $ request #new_search

Replace @magpiradio with your own Twitter account and #new_search with whatever string you'd like to listen to.

If you're running the say.rb tts script, you can also change the voice of the Request channel with the following tweet:

@magpiradio $ voice zarvox

Again, replace @magpiradio with your account and "zarvox" can be any of the following voices:

agnes, albert, alex, bad news, bahh, bells, boing, bruce, bubbles, cellos, deranged, fred, good news, hysterical, junior, kathy, pipe organ, princess, ralph, trinoids, vicki, victoria, whisper, zarvox

Contact

Magpi Radio was created as a student project by William Lindmeier at ITP/NYU. Email me at [e-mail]