Python: Azure AI mapping HostedImageGenerationTool to ImageGenTool

[giles17]

Motivation and Context

This PR fixes the Azure AI image generation sample by adding the missing handler for HostedImageGenerationTool in the Azure AI integration layer and updating the sample to use the correct content types for image generation results.

Changes:

• Added HostedImageGenerationTool to ImageGenTool mapping in Azure AI integration
• Updated sample to use ImageGenerationToolResultContent instead of DataContent for extracting image data
• Changed image save location from script directory to OS temporary directory
• Updated model name from gpt-image-1-mini to gpt-image-1 in tool options (the Image Generation tool is empowered by the gpt-image-1 model.)

Description

Contribution Checklist

• The code builds clean without any errors or warnings
• The PR follows the Contribution Guidelines
• All unit tests pass, and I have added new tests where possible
• Is this a breaking change? If yes, add "[BREAKING]" prefix to the title of the PR.

[markwallace-microsoft]

Python Test Coverage Report •

File	Stmts	Miss	Cover	Missing
packages/azure-ai/agent_framework_azure_ai
_shared.py	280	105	62%	120, 122, 124, 126, 171, 259–262, 265–267, 270–271, 281, 288–290, 316–325, 372, 388–390, 392–409, 417–420, 423–424, 427–434, 439–440, 446–449, 457–458, 461–462, 464, 476, 478–479, 481–482, 484–485, 487–495, 497, 533, 535–538, 540, 546, 552, 556, 571, 574–575, 577
TOTAL	17432	2677	84%

Python Unit Test Overview

Tests	Skipped	Failures	Errors	Time
3148	213 💤	0 ❌	0 🔥	1m 2s ⏱️

[Copilot]

Pull request overview

This PR fixes the Azure AI image generation sample by adding the missing handler for HostedImageGenerationTool in the Azure AI integration layer and updating the sample to use the correct content types for image generation results.

Changes:

• Added HostedImageGenerationTool to ImageGenTool mapping in Azure AI integration
• Updated sample to use ImageGenerationToolResultContent instead of DataContent for extracting image data
• Changed image save location from script directory to OS temporary directory
• Updated model name from gpt-image-1-mini to gpt-image-1 in tool options

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 2 comments.

File	Description
python/samples/getting_started/agents/azure_ai/azure_ai_with_image_generation.py	Updated sample to use correct content types and save location; however, uses incorrect option keys
python/packages/azure-ai/agent_framework_azure_ai/_shared.py	Added handler to map HostedImageGenerationTool to Azure AI's ImageGenTool

[Copilot]

The option keys used here don't match the HostedImageGenerationToolOptions type definition. According to the HostedImageGenerationToolOptions, the correct keys should be:

• "model_id" instead of "model"
• "image_size" instead of "size"
• "quality" is not a valid option key in HostedImageGenerationToolOptions

The valid keys defined in HostedImageGenerationToolOptions are: count, image_size, media_type, model_id, response_format, and streaming_count.

Suggested change

	"model": "gpt-image-1",
	"quality": "low",
	"size": "1024x1024",
	"model_id": "gpt-image-1",
	"image_size": "1024x1024",

[Copilot]

The code assumes that image_data[0] (which is content.outputs) will always be a DataContent object with a get_data_bytes() method. However, according to the ImageGenerationToolResultContent type definition, outputs can be either DataContent | UriContent | None. The UriContent class does not have a get_data_bytes() method, which would cause an AttributeError at runtime if a UriContent is returned.

Consider adding a type check or handling for both DataContent (with get_data_bytes()) and UriContent (which would need a different approach, such as downloading from the URI).