dora-gradio 0.2.0

dora-gradio

Dora Gradio UI Interface

A versatile UI interface for Dora-rs that provides text, audio, and video input capabilities using Gradio.

Features

• Text Input: Direct text input through a chat-like interface
• Audio Input: Real-time audio streaming in 16kHz format
• Video Input: WebRTC camera streaming at 640x480
• Multiple Output Channels:
  • text: For direct text messages
  • audio: For raw audio stream
  • image: For camera feed
• Clean Interface: Simple and intuitive UI with tabbed sections
• Auto Port Management: Automatically handles port conflicts

Installation

Using pip:

python -m venv .venv
source .venv/bin/activate
pip install -e .

Usage

1. Web Interface

The interface will be available at: http://localhost:7860

2. As a Dora Node

Create a YAML configuration:

nodes:
  - id: ui
    build: pip install -e .
    path: dora-gradio
    outputs:
      - text     # Text from chat interface
      - audio    # Raw audio stream
      - image    # Camera feed
    env:
      VIRTUAL_ENV: path to your venv   # comment this if not using venv

Run with Dora:

dora run demo.yml

3. Integration Examples

Video Processing Pipeline

nodes:
  - id: ui
    build: pip install -e .
    path: dora-gradio
    outputs:
      - image    # Camera feed

  - id: video_processor
    build: pip install -e path/to/processor
    path: video-processor
    inputs:
      video: ui/image

Audio Processing Pipeline

nodes:
  - id: ui
    build: pip install -e .
    path: dora-gradio
    outputs:
      - audio    # Raw audio stream

  - id: audio_processor
    build: pip install -e path/to/processor
    path: audio-processor
    inputs:
      audio: ui/audio

4. Demo Example

Here's a complete demo pipeline using the UI with visualization and audio processing:

nodes:
  - id: ui
    build: pip install -e .
    path: dora-gradio
    outputs:
      - text     # Text messages
      - audio    # Raw audio stream
      - image    # Camera feed

  - id: plot
    build: pip install dora-rerun
    path: dora-rerun
    inputs:
      text_input: ui/text
      audio: ui/audio
      image: ui/image

  - id: dora-vad
    build: pip install -e path/to/dora-vad
    path: dora-vad
    inputs:
      audio: ui/audio

  - id: dora-distil-whisper
    build: pip install -e path/to/dora-distil-whisper
    path: dora-distil-whisper
    inputs:
      audio: ui/audio

This demo showcases:

• Real-time visualization with dora-rerun
• Voice Activity Detection with dora-vad
• Speech processing with dora-distil-whisper

Interface Features

Camera Tab

• Real-time WebRTC video streaming
• Fixed 640x480 resolution
• BGR8 color format
• Automatic timestamp synchronization

Audio and Text Input Tab

• Chat-like interface for text input
• Real-time audio streaming (16kHz, mono)
• Status indicators for streaming state
• Immediate output through respective channels

Controls

• Send Text button for chat messages
• Stop Server button for graceful shutdown

System Requirements

• Python ≥ 3.10
• Required ports:
  • 7860 (Gradio interface)

Known Limitations

• Fixed video resolution (640x480)
• Fixed audio sample rate (16kHz)
• Requires port 7860 to be available

License

dora-gradio's code are released under the MIT License