Receipt Text Extractor (Ollama + Vision Model)

This Python script extracts structured text data from receipt images using Ollama and a multimodal LLM (default: llama3.2-vision).
It processes the image, sends it to the model, and returns strictly valid JSON with merchant, items, date, and totals.

Features

• Image preprocessing: resizes large images to a max side of 1280px for efficiency.
• Structured JSON output: always follows the same schema (merchant, items, totals, VAT, etc.).
• Strict format: model is instructed to return only JSON (no explanations or extra text).
• CLI usage: specify receipt image path and optional model.