Port to OpenAI vs Azure OpenAI, port infra

[pamelafox]

Purpose

• Ported from AzureOpenAI to OpenAI, as we no longer need to specify an API version, and we can pass in the token provider in the api_key parameter.
• Enhance the README to describe every file (we were missing function calling)
• Remove Langchain/Pydantic/Llamaindex as those are covered well in python-ai-agents-demos, so now this repo focuses on the lower-level openai package
• Adds infrastructure-as-code for provisioning Azure OpenAI models (chat AND embedding)
• Adds more detailed README sections about GitHub Models, Azure OpenAI, OpenAI.com, Ollama

Breakdown from Copilot:

Azure OpenAI integration and environment variable standardization:

• Refactored all scripts to use the generic openai.OpenAI/openai.AsyncOpenAI clients for Azure, replacing AzureOpenAI/AsyncAzureOpenAI, and updated the environment variable from AZURE_OPENAI_DEPLOYMENT to AZURE_OPENAI_CHAT_DEPLOYMENT for clarity and consistency. Also updated endpoint variables to include /openai/v1 in the URL. (chained_calls.py, chat.py, chat_async.py, chat_history.py, chat_history_stream.py, .env.sample, .env.sample.azure) [1] [2] [3] [4] [5] [6] [7] [8] [9]

Async resource management improvements:

• Improved chat_async.py to properly manage and close async clients and credentials, preventing resource leaks and ensuring clean shutdowns. [1] [2]

Documentation and onboarding enhancements:

• Significantly expanded the README.md to provide detailed, step-by-step instructions for configuring and running scripts with GitHub Models, Azure OpenAI, OpenAI.com, and Ollama. Clarified environment variable usage and added sections for function calling, RAG, structured outputs, and infrastructure setup. [1] [2] [3] [4] [5]

Infrastructure as Code (IaC) support:

• Added azure.yaml to define Azure Developer CLI hooks for provisioning and environment variable setup via scripts, enabling one-command resource provisioning for Azure OpenAI demos.

Code cleanup:

• Removed the legacy Azure-specific logic and the entire chat_langchain.py script, which previously handled Azure integration differently, to streamline and unify model usage patterns.

Does this introduce a breaking change?

[ ] Yes
[X] No

Pull Request Type

What kind of change does this Pull Request introduce?

[ ] Bugfix
[X] Feature
[ ] Code style update (formatting, local variables)
[ ] Refactoring (no functional changes, no api changes)
[X] Documentation content changes
[ ] Other... Please describe:

How to Test

• Run "azd up", run every sample.

[Copilot]

Pull Request Overview

This PR migrates the codebase from Azure-specific OpenAI clients to the generic OpenAI client for better consistency and simplicity, while adding comprehensive infrastructure provisioning capabilities.

• Replaces AzureOpenAI/AsyncAzureOpenAI with generic openai.OpenAI/openai.AsyncOpenAI clients across all scripts
• Standardizes environment variable naming from AZURE_OPENAI_DEPLOYMENT to AZURE_OPENAI_CHAT_DEPLOYMENT
• Adds Infrastructure as Code (Bicep templates) with Azure Developer CLI integration for automated resource provisioning
• Removes higher-level framework dependencies (Langchain, LlamaIndex, PydanticAI) to focus on core OpenAI package usage
• Enhances documentation with detailed setup instructions for all supported platforms

Reviewed Changes

Copilot reviewed 66 out of 68 changed files in this pull request and generated 3 comments.

Show a summary per file

File	Description
Multiple .py files	Updated Azure OpenAI client initialization to use generic OpenAI client with base_url and api_key parameters
spanish/chat_async.py	Enhanced async resource management with proper credential cleanup
chat_async.py	Enhanced async resource management with proper credential cleanup
Infrastructure files	Added Bicep templates and deployment scripts for Azure resource provisioning
README.md	Significantly expanded documentation with platform-specific setup instructions
Framework files	Removed Langchain, LlamaIndex, and PydanticAI example scripts

---

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}

[Copilot]

The import has been changed from azure.identity to azure.identity.aio, but this change is inconsistent with other files in the PR that still use the synchronous azure.identity module. This could cause import errors if azure.identity.aio is not available in the installed version of the azure-identity package.

Suggested change

	import azure.identity.aio
	import azure.identity

[pamelafox]

This one uses async as its the async example specifically

[Copilot]

The import has been changed from azure.identity to azure.identity.aio, but this change is inconsistent with other files in the PR that still use the synchronous azure.identity module. This could cause import errors if azure.identity.aio is not available in the installed version of the azure-identity package.

[pamelafox]

This one uses async as its the async example specifically

[Copilot]

The comment mentions 'gpt-oss' which appears to be a typo. It should likely be 'o1' or another valid reasoning model name, as 'gpt-oss' is not a known OpenAI model.

Suggested change

	model=MODEL_NAME, # Must be a reasoning model like gpt-5 or gpt-oss
	model=MODEL_NAME, # Must be a reasoning model like gpt-5 or o1

[pamelafox]

Lol, gpt-oss is a new open source model

[madebygps]

I pushed some changes, rest lgtm