DeepDanbooru API Service

This is a standalone API service that uses DeepDanbooru to analyze images and return tags.

Features

• REST API endpoints for image analysis
• Uses DeepDanbooru for accurate image tagging
• Configurable confidence threshold
• Optimized for deployment on Render

API Endpoints

GET /

Root endpoint to verify the API is running.

GET /health

Health check endpoint.

POST /analyze

Analyze an image using DeepDanbooru and return the tags.

Parameters:

• file: The image file to analyze (multipart/form-data)
• threshold: (Optional) The confidence threshold for tags (default: 0.5)

Example Response:

{
  "tags": [
    {"name": "1girl", "confidence": 0.9998},
    {"name": "solo", "confidence": 0.9994},
    ...
  ],
  "tag_names": ["1girl", "solo", ...],
  "tag_count": 25,
  "processing_time_seconds": 1.23
}

Deployment

This service is designed to be deployed on Render. It requires:

• Standard instance (at least 1GB RAM)
• Docker environment

Local Development

To run the service locally:

1. Install dependencies:

   pip install -r requirements.txt
   

2. Run the FastAPI application:

   uvicorn main:app --reload
   

3. Access the API at http://localhost:8000