Skip to content
This repository has been archived by the owner on Feb 8, 2019. It is now read-only.

Bug fixes and extended functionality #60

Open
wants to merge 21 commits into
base: master
Choose a base branch
from
Open

Conversation

drftg
Copy link

@drftg drftg commented Apr 20, 2018

[1.1.2] - 2018-04-20

  • Added timeout for screen
  • Added screen section in config to specify timeout and screen on/off commands

[1.1.1] - 2018-04-19

  • Changed handling of config.js so it is now backwards compatible
  • fixed some more GA bugs. It should be stable and robust now and handle conversations correctly

[1.1.0] - 2018-04-17

  • Changed handling of commands in config
  • Changed format of models in config.js so an old one will not work (see: assets/config.txt)
  • Added commands for screen on and off
  • Added commands for showing/hiding a single named module
  • Added mapping of spoken module name to actual module name (really helps non-english speakers)
  • Added command to play .wav
  • Added a single hotword can fire a list of commands and stay in listening mode
    for example, a spoken "wake up" can turn on the screen (like PIR) without activating GA or voice command
  • Added model specific confirmation sound added in config (parameter "confirm")
  • Added translate/nl.json for dutch language

drftg and others added 20 commits April 14, 2018 17:13
Added wake word. A single hotword can wake up MM
solved for execute and notification
…ist of commands (1 or more) instead of a hotword parameter.

Commands can be executed diectly via one hotword but you can also start a voice assistant an speak the commands.
Commands have been extended to show or hide a single module by name. Als commands to turn screen on or off have been added.
There is now a parameter "modulemap" that containsa mapping from spoken names to actual module names. This is to
make it easier for other language STT usage but can also help with mapping "assistant"->"MMM-Assistant".
It is advisable to train this model yourself via snowboy.kitt.ai
Trying to solve weird GA conversation error
@drftg
Copy link
Author

drftg commented Apr 20, 2018

[1.1.3] - 2018-04-20

  • Added Google TTS. Many more languages available than pico2tts
    Command conversation can now be in languages other than en-GB,de-DE,es-ES,fr-FR,it-IT like nl-NL
  • Added config switch to choose between pico and Google TTS
  • Fixed one more bug in GA. Saying "nevermind" just stopped the conversation without returning to listening mode

@E3V3A
Copy link
Collaborator

E3V3A commented Apr 22, 2018

@drftg
Great. I'll have a closer look at this in a few days.

Let's consider not putting more commands in the config, and instead try to separate out and put them in a file... I'll create an issue about this.

@drftg
Copy link
Author

drftg commented Apr 22, 2018

@E3V3A Agreed, the config is becoming quite bloated. I hope you have some ideas about how to organize this.

@drftg
Copy link
Author

drftg commented Apr 22, 2018

@E3V3A Btw @eouia set up quite an elaborate command structure for his different modules to communicate to each other but it is too complex for me to grasp at this moment. Things like screen handling are also solved in MMM-Tools but the only way I know of now to reach that command is to send a language specific string (like a spoken command) via notification. I would be very interested in a way to do that via the module defined functions themselves.

Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants