nvidia-riva / sample-apps Goto Github PK

JavaScript Socket.IO version: 4.5.4
python-socketio version: 4.4.0
python-engineio version: 3.11.1

https://python-socketio.readthedocs.io/en/latest/intro.html#version-compatibility

I am utilizing the latest version of riva_quickstart (v2.11.0), but I am unable to locate the "riva_api-<x.y.z>-py3-none-any.whl" file within the directory. Consequently, I proceeded to install nvidia-riva-client; however, I encountered an error upon attempting to execute the command "python3 main.py".

Traceback (most recent call last): File "main.py", line 12, in <module> from client.webapplication.start_web_application import start_web_application File "/root/riva_quickstart_v2.11.0/sample-apps/virtual-assistant/client/webapplication/start_web_application.py", line 8, in <module> from client.webapplication.server.server import * File "/root/riva_quickstart_v2.11.0/sample-apps/virtual-assistant/client/webapplication/server/server.py", line 22, in <module> from riva.chatbot.chatbots_multiconversations_management import create_chatbot, get_new_user_conversation_index, get_chatbot File "/root/riva_quickstart_v2.11.0/sample-apps/virtual-assistant/riva/chatbot/__init__.py", line 8, in <module> from .chatbot import * File "/root/riva_quickstart_v2.11.0/sample-apps/virtual-assistant/riva/chatbot/chatbot.py", line 10, in <module> from riva.asr.asr import ASRPipe File "/root/riva_quickstart_v2.11.0/sample-apps/virtual-assistant/riva/asr/__init__.py", line 8, in <module> from .asr import * File "/root/riva_quickstart_v2.11.0/sample-apps/virtual-assistant/riva/asr/asr.py", line 11, in <module> import riva_api.riva_audio_pb2 as ra ModuleNotFoundError: No module named 'riva_api'

Hi team,
We are trying to run this virtual-assistant-rasa application pipeline on a remote GCP ubuntu server but we are facing some issues in the pipeline.
For testing and exploration purposes we have ran the webapp without ssl-context by allowing a particular remote server endpoint in the chrome://flags/#unsafely-treat-insecure-origin-as-secure to enable the device-mic usage to stream the audio to the server.
It is perfectly able to send the audio-bytes to the Flask-SocketIO backend and the Riva Server for transcription. And we are also able to see the transcripts on the webapp logs (as we're running in Verbose mode). But the transcripts are not received to the UI's text-field, which then get's submitted and sent to the Rasa DM + Riva NLP.

So, require some help in debugging this issue. It would great if any contributor will be able to help us resolving this.

Interesting fact:
When I am forwarding the remote server ports to localhost, it is working perfectly alright.