Python: Fix AzureAIClient tool call bug for AG-UI use

[moonbox3]

Motivation and Context

The Responses API expects tool calls to start with fc_*. AG-UI is passing tool call IDs that need this prefix. We also need to make sure we're only emitting ToolCallStartEvent once per tool call (when it's a new tool call).

Description

• Fixes Python - AGUI - CopilotKit - Errors when using AzureAIClient instead of samples using AzureOpenAIChatClient #3064

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/ag-ui/agent_framework_ag_ui
_events.py	261	11	95%	197, 216, 245, 259, 282, 362–363, 392, 403, 426, 482
packages/core/agent_framework/openai
_responses_client.py	517	102	80%	153, 158, 160–161, 172, 190, 198, 201–208, 210, 215, 221, 225, 240, 243–244, 246, 283, 313, 357–358, 380, 385, 419, 437, 440, 476, 530, 609, 614, 618–620, 624–625, 648, 663–664, 668–670, 717, 737–738, 751–752, 768–769, 858–859, 868–869, 900–901, 917, 919, 993–1001, 1024–1030, 1049, 1067–1068, 1077, 1082–1084, 1090–1091, 1104, 1119, 1155–1156, 1158–1160, 1174–1175, 1178, 1180, 1190–1191, 1197, 1212
TOTAL	17626	2888	83%

Python Unit Test Overview

Tests	Skipped	Failures	Errors	Time
2614	154 💤	0 ❌	0 🔥	59.789s ⏱️

[Copilot]

Pull request overview

This PR fixes a bug in the AzureAIClient tool call handling for AG-UI integration. The OpenAI Responses API requires tool call IDs to start with the fc_ prefix, and AG-UI was passing IDs without this prefix, causing failures. Additionally, the PR ensures that ToolCallStartEvent is only emitted once per tool call.

Key changes:

• Added logic to prepend fc_ prefix to tool call IDs when missing
• Updated event emission logic to prevent duplicate ToolCallStartEvent emissions
• Improved code formatting across multiple files for better readability

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 1 comment.

File	Description
python/packages/core/agent_framework/openai/_responses_client.py	Added fc_ prefix validation and prepending logic for tool call IDs; improved code formatting across import statements and function calls
python/packages/ag-ui/agent_framework_ag_ui/_events.py	Modified ToolCallStartEvent emission logic to only emit once per new tool call, preventing duplicate events

Comments suppressed due to low confidence (1)

python/packages/core/agent_framework/openai/_responses_client.py:635

• The new logic to add the 'fc_' prefix to function call IDs lacks test coverage. Consider adding a unit test that verifies:

1. When a FunctionCallContent has a call_id without 'fc_' prefix, it gets added
2. When a FunctionCallContent has a call_id that already starts with 'fc_', it's not duplicated
3. When fc_id is provided in additional_properties, it's used correctly
   This is critical functionality for AG-UI integration and should have explicit test coverage.

            case FunctionCallContent():
                # Use fc_id from additional_properties if available, otherwise fallback to call_id
                fc_id = call_id_to_id.get(content.call_id, content.call_id)
                # OpenAI Responses API requires IDs to start with `fc_`
                if not fc_id.startswith("fc_"):
                    fc_id = f"fc_{fc_id}"
                return {
                    "call_id": content.call_id,
                    "id": fc_id,
                    "type": "function_call",
                    "name": content.name,
                    "arguments": content.arguments,
                    "status": None,
                }

[eavanvalkenburg]

all those syntax changes are a bit strange...

[moonbox3]

all those syntax changes are a bit strange...

I agree. Something is going on with ruff, even though it says it's using 120 line length.