[Feature]: Sample for direct PDF ingestion & structured extraction with GPT-4.1 in Azure OpenAI Service

[jamesmcroft]

Feature Description

Demonstrate how to use Azure OpenAI GPT-4.1 models to accept PDF files as the content input (based on the OpenAI file inputs feature https://platform.openai.com/docs/guides/pdf-files?api-mode=chat) and return a JSON structured output.

Purpose is to evaluate both the effectiveness of GPT-4.1, as well as the effectiveness of processing an entire document using the file input type.

Use Case

With GPT-4.1 and direct file input:

• Engineering effort should drop to a single GPT call inside existing solutions.
• Newer GPT-4.1 models should enhance the accuracy and effectiveness of document data extraction solutions.

Motivation

• Simplified solution – reduces the effort required to build document ingestion pipelines with GPT models

[jamesmcroft]

Updated default deployment to use GPT-4.1, and provided notebook outputs for document processing with the model.

[jamesmcroft]

Still need to provide a sample for OpenAI input files