Label Parser Monitor Service

A containerized service that monitors a directory for new product images, processes them through OpenAI's vision API, and saves structured data to CSV. Now includes a web interface for interactive image processing.

Features

• Directory Monitoring: Automatically processes images in d_warehouse directory
• Web Interface: Upload and process images through a modern web UI
• AI Processing: Uses OpenAI GPT-4 Vision to extract product information
• Complete Text Extraction: Captures ALL visible text from product labels
• CSV Export: Download results as CSV files
• Retry Logic: Exponential backoff for API failures
• Containerized: Docker deployment with volume persistence
• Henkel Branding: Professional UI with company branding

Setup

1. Copy environment variables:

cp .env.example .env

2. Add your OpenAI API key to .env:

OPENAI_API_KEY=your_actual_api_key_here
OPENAI_TIMEOUT=60

3. Build and run with Docker Compose:

docker-compose up --build

Usage

Web Interface

1. Open your browser to http://localhost:8000
2. Upload a product image using drag-and-drop or file browser
3. Click "Process Image" and wait for AI analysis
4. View results and download CSV file

Directory Monitoring

1. Place product images in data/d_warehouse/
2. The monitor service automatically processes new images
3. Results are saved to data/d_mart/processed_images.csv

Local Development

Run the web interface locally:

python run_web.py

Run the monitor service locally:

python run_local.py

Testing

Run the test suite to verify functionality:

# Test without OpenAI (uses mock data)
python test_monitor.py

# Test with real OpenAI API (requires API key)
python test_monitor.py
# Choose 'y' when prompted

Configuration

Environment variables:

Variable	Default	Description
OPENAI_API_KEY	-	Your OpenAI API key
OPENAI_MODEL	gpt-4o	OpenAI model to use
OPENAI_TEMPERATURE	0.1	Response randomness
OPENAI_MAX_TOKENS	100	Max response length
OPENAI_TIMEOUT	60	API timeout in seconds

CSV Output Format

Column	Description
image_name	Original image filename
item	Complete product description with ALL label text
price	Product price
brand	Brand name
size	Product size
product_type	Product category

Supported Image Formats

• JPG/JPEG
• PNG
• BMP
• TIFF

Services

The application runs two services:

1. Monitor Service (label-parser-monitor): Watches directory for new images
2. Web Service (label-parser-web): Provides web interface on port 8000

Troubleshooting

API Timeout Errors:

• Increase OPENAI_TIMEOUT value
• Check network connectivity
• Verify API key is valid

Processing Failures:

• Service retries failed requests 3 times with exponential backoff
• Check logs in logs/image_parser.log
• Ensure images are valid and not corrupted

Web Interface Issues:

• Ensure port 8000 is not in use
• Check browser console for JavaScript errors
• Verify file upload size limits