deepspeech binding not available for node 16 on ubuntu

[johannbarbie]

Specs

• Leon version: 1.0.0-beta.5
• OS (or browser) version: ubuntu 20.04
• Node.js version: 16.13.2
• Complete "npm run check" output:

  ➡ Here is the diagnosis about your current setup
  ✔ Run
  ✔ Run modules
  ✔ Reply you by texting
  ✔ Amazon Polly text-to-speech
  ❗ Google Cloud text-to-speech
  ❗ Watson text-to-speech
  ✔ Offline text-to-speech
  ❗ Google Cloud speech-to-text
  ❗ Watson speech-to-text
  ❗ Offline speech-to-text

Expected Behavior

after following instructions on https://docs.getleon.ai/offline, and configuring Deepspeach, it should be possible to run the server and issue command.

Actual Behavior

npm run setup:offline seems to complete fine:

  ➡ Setting up offline hotword detection...
  ➡ Installing dependencies...
  ➡ Executing the following command: sudo apt-get install sox libsox-fmt-all -y
  ✔ System dependencies downloaded
  ➡ Installing hotword dependencies...
  ✔ Offline hotword detection installed
  ➡ Setting up offline text-to-speech...
  ✔ Offline text-to-speech is already installed
  ➡ Setting up offline speech-to-text...
  ✔ Offline speech-to-text is already installed
  ✔ Hooray! Offline components are installed!

then, after npm run start, server throws error like this:

.: DEEPSPEECH PARSER :.

➡ GPU version not found, trying to get the CPU version...
✖ No DeepSpeech library found:
GPU: Error: Cannot find module 'deepspeech-gpu'
Require stack:
- /home/ubuntu/leon/server/dist/stt/deepspeech/parser.js
- /home/ubuntu/leon/server/dist/stt/stt.js
- /home/ubuntu/leon/server/dist/core/server.js
- /home/ubuntu/leon/server/dist/index.js
CPU: Error: Cannot find module '/home/ubuntu/leon/node_modules/deepspeech/lib/binding/v0.9.3/linux-x64/node-v93/deepspeech.node'
Require stack:
- /home/ubuntu/leon/node_modules/deepspeech/index.js
- /home/ubuntu/leon/server/dist/stt/deepspeech/parser.js
- /home/ubuntu/leon/server/dist/stt/stt.js
- /home/ubuntu/leon/server/dist/core/server.js
- /home/ubuntu/leon/server/dist/index.js
➡ Loading model from file bin/deepspeech/deepspeech.pbmm...
/home/ubuntu/leon/server/dist/stt/deepspeech/parser.js:75
    model = new DeepSpeech.Model(args.model);
            ^

TypeError: DeepSpeech.Model is not a constructor
    at Object.parser.init (/home/ubuntu/leon/server/dist/stt/deepspeech/parser.js:75:13)
    at Stt.init (/home/ubuntu/leon/server/dist/stt/stt.js:55:27)
    at Socket.<anonymous> (/home/ubuntu/leon/server/dist/core/server.js:163:20)
    at Socket.emit (node:events:390:28)
    at Socket.emitUntyped (/home/ubuntu/leon/node_modules/socket.io/dist/typed-events.js:69:22)
    at /home/ubuntu/leon/node_modules/socket.io/dist/socket.js:466:39
    at processTicksAndRejections (node:internal/process/task_queues:78:11)

How Do We Reproduce?

follow instruction on ubuntu 20.04.

Extra (like a sample repo to reproduce the issue, etc.)

deepspeech misses node v16 binding. how was this ever working, if leon requires node version > 16 ?

[johannbarbie]

deepspeech has ben discontinued by mozilla. Apparently coqui.ai is a fork by original devs with maintained codebase. I guess a new integration needs to be written.

[johannbarbie]

already on the todo list here: #329
closing this issue.

[JRMeyer]

We've happy to help get Speech-to-Text in Leon!

The Coqui STT folks hang out here: gitter.im/coqui-ai/STT

don't hesitate to let us know how we might help 💚