This service will take care of receiving the voice message and transcribing it using the Mozilla's DeepSpeech model. To learn more about DeepSpeech, you can access to the documentation here. In the stt.js file you'll find how the received audioBuffer is processed.
npm install to install all dependencies.
If you don't want to use docker, please make sure, that you installed deepspeech models by executing the following command ./download_models.sh
In the gdo_voicebot you'll find a lint bash that will check the integrality of the syntax in each microservice. Execute ./lint.sh to symply check the syntax and ./lint.sh --fix to correct syntax errors automaticaly.
Locally you can execute npm run lintornpm run lint:fix.
Use npm run start:dev if you want the server to be refreshed automatically at any modification in the code. Else you can directly use npm run start
In the gdo_voicebot, you'll find a docker-compose file. Here you can run docker-compose up do-voice-stt-service and that's it.
In the docs folder, you'll find javascript documentation about the Speech-to-Text service.
To generate your own documentation, please use jsdoc <file_1.js> <file_2.js> indicating all documented javascripts files.