Product Image Pipeline

Portfolio context: Extracted from founder-led production systems — multi-marketplace inventory, orders, and warehouse execution. Full portfolio · aspiranted.github.io

Image management and label generation library for e-commerce operations. Handles the full lifecycle: photo upload validation, URL management, shipping label generation with dynamic layouts, barcode labels, and carrier label processing.

Extracted from a production system processing 500+ labels/day across multiple carriers.

Architecture

src/
+-- validation/
¦   +-- image_validator.py      # URL validation, content type, size limits
+-- label/
¦   +-- layout.py               # Dynamic label layout engine (PIL)
¦   +-- barcode_label.py        # Code128 barcode labels with rotation
¦   +-- downloader.py           # Carrier label download (PDF/image)
+-- photo/
    +-- photo_store.py          # Photo upload/retrieval abstraction

Key Features

Image Validation (image_validator.py)

• URL format validation (HTTP/HTTPS)
• Content type enforcement (image/*)
• Configurable file size limits
• URL list normalization (strip, filter, deduplicate)
• CSV conversion for comma-separated URL storage

Label Layout Engine (layout.py)

• Dynamic vertical layout — content drawn top-to-bottom with cursor tracking
• Multi-carrier support — postal carriers (Correos) get QR codes, others get text codes
• Carrier label embedding — downloads and embeds carrier-provided labels (PDF/image)
• Multi-item orders — shows item list for bundle shipments
• Configurable — font sizes, margins, dimensions via LayoutConfig
• Auto-crop — starts with max height, crops to actual content

Barcode Labels (barcode_label.py)

• Landscape-to-portrait rotation — built in 1180x720 (readable), rotated 90° for 62mm paper
• Code128 barcodes — scanner-friendly with configurable module dimensions
• Large centered text — identifier visible from distance

Carrier Label Downloader (downloader.py)

• PDF support — converts first page to image via PyMuPDF at print DPI
• Image support — PNG, JPG, etc.
• Auto-resize — scales to target width maintaining aspect ratio
• Injectable HTTP — http_get parameter for testing without network

Photo Store (photo_store.py)

• Callback-driven — database-agnostic via save_photo_fn, list_photos_fn, etc.
• Upload validation — content type + size checks before persistence
• URL management — normalize, validate, and persist image URL lists

Design Decisions

Decision	Rationale
PIL for label generation	Production-proven, handles fonts/QR/barcodes natively
Landscape build + 90° rotate	Content readable on label, fits 62mm paper physically
Postal QR vs carrier label	Correos machines scan QR reliably; GLS/InPost have their own labels
Dynamic height with crop	Different content amounts produce different label heights
Injectable HTTP/storage	Testing without network or database dependencies

Installation

pip install -r requirements.txt

Running Tests

pytest tests/ -v

108 tests covering validation, layout generation, barcode labels, downloading, and photo storage.

License

MIT