v1.2.0 - Vision/Image Support

What's New

Vision/Image Support

DNFileRAG now supports PNG, JPG, JPEG, and WEBP image files. When enabled, images are processed using Ollama's llava vision model to:

• Extract visible text (OCR-like functionality)
• Generate image descriptions
• Index the content for RAG queries

New Features

• IVisionTextExtractor interface for image text extraction
• OllamaVisionTextExtractor implementation using Ollama llava model
• ImageParser for processing image files during ingestion
• VisionOptions configuration section:
  • Vision:Enabled - Enable/disable vision processing
  • Vision:Provider - Vision provider (currently: Ollama)
  • Vision:Ollama:Model - Vision model (default: llava)

Configuration

Enable in appsettings.json:

{
  "Vision": {
    "Enabled": true,
    "Provider": "Ollama",
    "Ollama": {
      "BaseUrl": "http://localhost:11434",
      "Model": "llava"
    }
  }
}

Getting Started

# Pull the llava model
ollama pull llava

# Add images to your documents folder
cp image.png src/DNFileRAG/data/documents/

Documentation

• Created CHANGELOG.md with version history
• Updated README with image format support
• Updated architecture documentation to v1.5

Full Changelog

v1.1.0...v1.2.0

v1.1.0

What's New in v1.1.0

Features

• API Key Authentication - Secure your API with configurable API keys via X-API-Key header
• Dev Mode Bypass - Set RequireApiKey: false for easy local development
• Role-Based Access Control - reader and admin roles for endpoint authorization

Improvements

• Enhanced HelpChat example with new logo and assets
• Updated documentation with correct API endpoints
• Added sample Northwind documents for testing

Bug Fixes

• Fixed API endpoint documentation inconsistencies

Full Changelog

v1.0.0...v1.1.0