A module to easily use stt and tts in a project. Features stt triggering via a wake word.
Development on this module has been temporary suspended, focus is currently on genie-router.
This module currently is an empty shell. I will work on it in the coming weeks. The stt part will mostly be a wrapper around the sonus project. Text-to-speech should be available online (Google, Amazon Polly, etc.) and offline (espeak, mimic, etc.).
sudo apt-get install libatlas-base-dev sox
The project contains a Dockerfile which will install the required packages via apt-get, copy package.json and then run npm install. This image is meant to be used
during development, it has not been tested for production use.
To easily create the required volume mappings you can use docker-compose to map
the code from the host into the container. Simply execute:
docker-compose up
This will build the Dockerfile and start running dev.js using nodemon. Nodemon will monitor file changes and restart the process. This command makes 2 assumptions:
- There is a Google project with id genie-ai
- There is a file googlekey.json in the root of the folder that contains an API key to invoke the Google Speech to Text API. (Be careful not to commit this file!)
You can create a container for runing the test by executing:
docker build -t matueranet/genie-ai:dev .
docker run --name=genie-ai-test -v `pwd`:/home/app -v /home/app/node_modules -w /home/app matueranet/genie-ai:dev npm test
After this you can simply execute:
docker start -a genie-ai-test