Initial Roadmap and Structure

General Ideas For Each Part:

• Core API
  • Speech Recognition
  • Speech-to-text
  • Text-to-speech
  • Text-to-intent
• Extensions
  • Mad Libs
  • Mood Recognition
  • Weather
  • Talk to Bash Bot and plot world domination
  • Send messages to people
  • Open other apps ex spotify
• Interfaces
  • App with GUI
  • App with speech
  • Discord chat
  • Raspberry pi
  • App with text/keyboard input
  • Video input

---

Process of build and runtime

• Build
  • Combine commands from extensions into global list
    • Includes tracking which commands go with which extensions
    • Combine test data
  • Check for collisions in commands
• Runtime
  • Command list
    • Global list of commands to choose from
  • OLGA Output Object Generator

---

Interface Definitions
This is meant to be a sample of what each component (Core, extensions, front ends) might implement.

• Core API
  • OLGA.run(string knownCommand)
    • Input = string that is a known command from the global list
    • Return = OLGA Output Object (ÖÖÖ)
  • OLGA.parse(string unknownCommand)
    • Input = string that is intended by the user to be a command but which is not a word-for-word duplicate of a command from the global list
    • Return = string representation of a command from the global list which most closely matches the input string (arrived at most likely via neural network)
  • OLGA.listen(audio voiceInput)
    • Input = audio that needs to be interpreted
    • Output = string representation of the voice input
  • OLGA.output (OLGA Output Object)
    • Text property
    • Audio property
    • Image property
    • Video
    • Link/URL
    • Program to Open/Commands to execute
• Extension
  • EXT.install()
    • Sets commands in the global config
      • Check for command conflicts
    • Check for dependencies
  • EXT.init()
    • Start the extension asynchronously
  • EXT.listen(string commandID)
    • Wait for commands from OLGA core
    • Return = OLGA Output Object
• Front End
  • FNT.processOutput(OLGA.output output)
    • Handle the OLGA.output object returned from the Core API
  • very dependent on the platform for handling input