fix: enhance message handling and context management in orchestrators

[Dhanushree-Microsoft]

Purpose

This pull request introduces several improvements and refactorings to the agent framework, focusing on cleaner option handling, improved tool type definitions, robust context trimming for Azure OpenAI responses, and more consistent context provider interfaces. The changes also address minor bugs and streamline the construction and orchestration of group chat workflows.

Does this introduce a breaking change?

• Yes
• No

[github-actions]

Processor Coverage Report •

File	Stmts	Miss	Cover	Missing
src/processor/src/libs/agent_framework
agent_builder.py	199	29	85%	448, 450, 454, 456, 458, 460, 462, 464, 466, 472, 474, 476, 483, 488, 791, 793, 795, 797, 799, 801, 803, 805, 807, 809, 815, 817, 819, 826, 831
agent_info.py	25	0	100%
azure_openai_response_retry.py	455	208	54%	91, 201, 209, 231, 257, 267–269, 334, 336, 353–357, 359–366, 375, 382–387, 389–396, 398–403, 405, 407, 409–410, 413, 415–429, 432, 434, 436, 504, 511–514, 516–518, 524–526, 548, 569–570, 575–578, 581–584, 586–589, 591, 601–603, 605–608, 610–613, 615–619, 621, 629–631, 633, 656–661, 664–666, 668, 672–677, 679–680, 686–687, 691–694, 698, 701, 709–710, 716–717, 722, 724, 745, 751–753, 757–758, 774–775, 777–780, 784, 787, 795–796, 802–804, 806–809, 811–822, 825, 832–833, 858–859, 866–867, 869, 873, 876–877, 881–882, 884–886, 894, 896–898, 900–902, 904–905, 915
groupchat_orchestrator.py	544	90	83%	129–130, 137–138, 143–144, 471, 474–478, 480, 482–483, 486, 489, 492, 494, 496–499, 506–507, 514–515, 522–523, 526–528, 530–531, 536–538, 541–542, 547, 550–552, 555–556, 561, 564, 570, 573, 577, 579–583, 587–588, 594, 597, 601, 606, 609, 620–621, 623, 625–626, 628, 638–639, 641, 734, 908, 927, 929, 931, 960, 1089, 1091, 1207, 1216, 1218, 1263–1264, 1267, 1288–1291, 1309
shared_memory_context_provider.py	122	21	82%	26, 115, 143, 218, 229, 233, 238, 270, 281, 286, 309–311, 318–319, 321–326
src/processor/src/libs/base
orchestrator_base.py	169	52	69%	65, 71, 74–77, 83–84, 86–87, 130, 141, 146, 151–152, 156, 164, 166, 168, 176–177, 184, 190, 192, 199, 203, 212–215, 222, 224–225, 227, 326–328, 331, 337, 344–346, 373–375, 378, 385, 392–394, 445–446
src/processor/src/steps
migration_processor.py	268	44	83%	88–89, 145–146, 160–163, 167, 175, 210–211, 217–218, 222, 226, 297–300, 303–306, 310–311, 324–325, 383–384, 484–485, 535–536, 543–544, 582–583, 601–602, 666–667, 691, 720
TOTAL	5920	849	85%

Tests	Skipped	Failures	Errors	Time
812	0 💤	0 ❌	0 🔥	19.841s ⏱️

[github-actions]

Coverage Report •

File	Stmts	Miss	Cover	Missing
TOTAL	3097	208	93%

report-only-changed-files is enabled. No files were changed during this commit :)

Tests	Skipped	Failures	Errors	Time
588	0 💤	0 ❌	0 🔥	30.228s ⏱️

[Copilot]

Pull request overview

This PR updates the processor’s agent/orchestration integration to align with newer agent_framework APIs, improving how workflows stream events, how agent options/context providers are wired, and adding more robust Azure OpenAI Responses context-trimming behavior.

Changes:

• Refactors workflow construction and streaming execution to use updated WorkflowBuilder/Workflow.run(..., stream=True) patterns.
• Standardizes agent construction by moving model/options into ChatOptions (default_options) and attaching context providers consistently (including in-memory history).
• Enhances Azure OpenAI Responses context trimming by adding a post-trim cleanup step intended to prevent orphaned tool call/result pairs.

Reviewed changes

Copilot reviewed 7 out of 7 changed files in this pull request and generated 2 comments.

Show a summary per file

File	Description
src/processor/src/steps/migration_processor.py	Updates workflow graph construction and switches to the newer streaming run API.
src/processor/src/libs/base/orchestrator_base.py	Adjusts agent creation: sets store=False and adds InMemoryHistoryProvider alongside shared memory when enabled.
src/processor/src/libs/agent_framework/shared_memory_context_provider.py	Migrates context provider lifecycle hooks to the newer before_run/after_run interface and injects via context.
src/processor/src/libs/agent_framework/groupchat_orchestrator.py	Switches to agent_framework_orchestrations.GroupChatBuilder and updates streaming + response_format passing.
src/processor/src/libs/agent_framework/azure_openai_response_retry.py	Adds orphaned tool-message cleanup after trimming; improves endpoint/kwargs handling and adjusts Foundry base URL behavior.
src/processor/src/libs/agent_framework/agent_info.py	Broadens accepted tool types (FunctionTool + MCP tool types) via a shared ToolType alias.
src/processor/src/libs/agent_framework/agent_builder.py	Builds ChatOptions for default_options and normalizes middleware/context providers into sequences for Agent construction.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[github-actions]

🎉 This PR is included in version 2.1.1 🎉

The release is available on GitHub release

Your semantic-release bot 📦🚀