Python: Refactor ag-ui to clean up some patterns

[moonbox3]

Motivation and Context

This is a pure refactor of existing functionality for ag-ui. No breaking changes introduced. All integrated features still work - tested with AG-UI's Dojo app along with the getting starter script.

Description

Cleaning up some patterns in ag-ui.

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
_agent.py	50	3	94%	65, 121, 181
_endpoint.py	41	0	100%
_events.py	321	18	94%	216, 237, 266, 283, 306, 319, 351, 360–363, 424–425, 429, 490–491, 570, 583
_orchestrators.py	258	18	93%	137, 355, 387, 439–441, 444, 447–448, 452–455, 457–459, 503, 508
packages/ag-ui/agent_framework_ag_ui/_orchestration
_message_hygiene.py	109	3	97%	106, 170–171
_state_manager.py	46	1	97%	91
_tooling.py	54	13	75%	20–22, 29–31, 40–44, 55–56
packages/core/agent_framework
_agents.py	290	51	82%	329, 388–390, 436, 490, 508, 670, 849, 852–854, 983–986, 988, 991–993, 1082, 1123, 1125, 1134–1139, 1145, 1147, 1157–1158, 1165, 1167–1168, 1176–1180, 1188–1189, 1191, 1196, 1198, 1232, 1273–1274, 1276, 1289
TOTAL	15594	2362	84%

Python Unit Test Overview

Tests	Skipped	Failures	Errors	Time
2230	130 💤	0 ❌	0 🔥	53.839s ⏱️

[Copilot]

Pull Request Overview

This PR refactors the ag-ui package to improve code organization and maintainability by extracting complex orchestration logic into focused, testable modules. No breaking changes are introduced.

Key changes:

• Extracted message sanitization, deduplication, tool handling, and state management into dedicated modules under _orchestration/
• Refactored event handling into smaller, more focused methods for better readability
• Added support for Pydantic models as state schemas
• Introduced default_state parameter for FastAPI endpoints to seed initial state
• Converted f-string logging to lazy evaluation format (using % formatting) for better performance

Reviewed Changes

Copilot reviewed 13 out of 13 changed files in this pull request and generated 4 comments.

Show a summary per file

File	Description
tests/test_tooling.py	New tests for tool merging and registration utilities
tests/test_state_manager.py	New tests for state initialization, snapshots, and context injection
tests/test_message_hygiene.py	New tests for message sanitization and deduplication logic
tests/test_agent_wrapper_comprehensive.py	Added test for Pydantic model state schema support
tests/test_endpoint.py	Added test for default_state seeding in endpoints
agent_framework_ag_ui/_orchestrators.py	Refactored to use extracted orchestration modules; converted to lazy logging
agent_framework_ag_ui/_orchestration/tooling.py	New module for server tool collection, client tool registration, and tool merging
agent_framework_ag_ui/_orchestration/state_manager.py	New module managing state defaults, snapshots, and structured updates
agent_framework_ag_ui/_orchestration/message_hygiene.py	New module for message sanitization and deduplication
agent_framework_ag_ui/_orchestration/__init__.py	Module exports for orchestration helpers
agent_framework_ag_ui/_events.py	Refactored event handling into smaller methods; converted to lazy logging
agent_framework_ag_ui/_endpoint.py	Added default_state parameter with deep copy support for initial state seeding
agent_framework_ag_ui/_agent.py	Added Pydantic model support for state_schema with normalization logic

[Copilot]

Pull request overview

Copilot reviewed 26 out of 26 changed files in this pull request and generated 4 comments.

[Copilot]

Copilot encountered an error and was unable to review this pull request. You can try again by re-requesting a review.