Feature: Add image file support for invoice uploads

Feature Request: Image File Support for Invoice Processing

Description

Currently, the invoice processing system only supports PDF and DOCX file formats. We need to add support for image file formats to handle scanned invoices and photos taken of paper invoices.

Supported Formats to Add

• JPG/JPEG - Most common format for photos
• PNG - High quality scanned documents
• TIFF - Professional scanning format
• BMP - Basic bitmap support

Current Implementation

• The system uses PaddleOCR which already supports image formats
• File validation is currently restricted to .pdf and .docx in services/invoice_processing_service.py
• Pattern extraction and job processing should work identically for images

Required Changes

1. Update File Validation

• Modify ALLOWED_FILE_TYPES in InvoiceProcessingService
• Add image MIME type validation
• Consider file size limits for high-resolution images

2. OCR Service Updates

• Verify PaddleOCR handles all image formats correctly
• Test quality/confidence scores for different image types
• Add image preprocessing if needed (rotation, contrast enhancement)

3. File Organization

• Update file storage paths to handle image files appropriately
• Ensure proper thumbnail generation for web interface (future)

4. Testing Requirements

• Test with real invoice images from various sources
• Validate OCR quality across different image types
• Ensure pattern extraction works with image-based OCR text

Priority

Medium - Not blocking current functionality, but important for user convenience and mobile upload capabilities.

Acceptance Criteria

• [ ] JPG, PNG, TIFF, and BMP files can be uploaded via /invoices/upload endpoint
• [ ] OCR processing works correctly for all image formats
• [ ] Pattern extraction maintains same accuracy as PDF processing
• [ ] File organization follows existing truck/vendor categorization
• [ ] Error handling provides clear feedback for unsupported/corrupted images
• [ ] Integration tests pass with real invoice images

Technical Notes

• PaddleOCR documentation: https://github.com/PaddlePaddle/PaddleOCR
• Consider image preprocessing for better OCR results
• Mobile devices often capture images at various orientations

Related Files

• services/invoice_processing_service.py - File validation
• services/ocr_service.py - OCR processing
• api/invoices.py - Upload endpoint
• schemas/invoice.py - Response schemas