FastAPI PDF Invoice Extractor

I need a Python-based HTTP service built with FastAPI. This service will receive a PDF invoice file via POST request and return extracted data in JSON format.

Requirements:
- Receive one PDF per HTTP request
- Return HTTP 200 with JSON or HTTP 500 with error message
- Extract fields: invoice_number, invoice_date, total_amount, currency
- Use Tesseract for data extraction
- Synchronous response, reasonably fast (a few seconds)
- Log only errors

For testing:
- Launch locally
- Provide URL and Bearer token

Later, prepare code for production deployment in a Docker container.

Ideal skills and experience:
- Proficient in Python and FastAPI
- Experience with Tesseract and OCR tools
- Familiarity with Docker and API authentication