ADA

Autonomous Digital Assistant

Suspended until an alternative to Google Speech is found (first one closed, second one need an API key and seems limited to 50 requests a day)

General Idea :

The general idea is to make a vocal assistant taking advantage of all of API and service now available out there. It will mainly be built in Python with easy modularity in mind.

Process :

Voice input -> STT -> Google -> Match module -> Exec module -> TTS Answer -> Output answer

Workflow :
-> Test recognition + print input/output
Test voice AND print output
Test 2 phase recognition ("Name to wait then execute order or timeout)
Test interpretation command and execution
Test interpretation from AIML file
Test GUI (fullscreen, printed output, voice animation)

Module Ideas:

• Easy
  • Say something
  • Ask time
  • Translation (Google Translate)
  • Calculation
    
• Medium
  • Meteo
  • Lights (hardware need)
    
• Hard
  • Music (+TV/Mediacenter/XBMC server)
  • eMail
    

Improvements:

• Use PyGSR for the GoogleAPI access
• GUI: Tkinter (+Tksnack for sound display)
• Create Client/Server interface to control TV/XBMC through HDMI (CEC)
• AIML: Based on Pandorabots ?
• Support for W-Zave/X10 infrastructure
• Voice: Modify before playing with Tksnack ?