[AAASM-990] ✨ (adapters): Add edge emission hooks for LangGraph and OpenAI Agents#42
Merged
Merged
Conversation
Posts to POST /topology/edges with source/target agent IDs, edge type, and optional metadata JSON. Closes #AAASM-990
Wraps GatewayClient.report_edge() in a daemon thread so callers are never blocked by network I/O.
Stores the completed node name in threading.local after each node exit; on the next node enter, fires a fire-and-forget Messages edge via _EDGE_EMITTER if one is registered. set_edge_emitter() and LangGraphPatch.edge_emitter field expose the wiring point.
After each Handoff.__call__ completes, fire a delegates_to edge from process_agent_id to the handoff target via _EDGE_EMITTER. set_edge_emitter() and OpenAIAgentsPatch.edge_emitter expose the wiring point.
…pters Four tests for LangGraph messages edges (sequential 2-node, first-node has no predecessor, 3-node chain, no-emitter guard). Four tests for OpenAI Agents delegates_to edges (basic handoff, None emitter, None process_agent_id, name-attribute fallback). Also applies ruff UP045 auto-fixes to gateway.py and emitter.py.
Codecov Report❌ Patch coverage is 📢 Thoughts on this report? Let us know! |
LangGraph: include transition_input_keys (state keys at node exit time) in Messages edge metadata. OpenAI Agents: include reason (from handoff tool_description/description) in DelegatesTo edge metadata. Tests updated to assert metadata fields are present and correct.
|
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
GatewayClient.report_edge()— new sync method posting toPOST /topology/edgeswith source/target agent IDs, edge type, and optional metadata JSON (REST endpoint landing in AAASM-995).EdgeEmitter— fire-and-forget wrapper (agent_assembly/client/emitter.py) that dispatchesreport_edge()on a daemon thread so adapters are never blocked by network I/O._NODE_TRANSITIONthread-local +_EDGE_EMITTERmodule global: after each node exits, the next node enter emits amessagesedge (source = previous node name, target = current node name).Handoff.__call__completes, emits adelegates_toedge (source =process_agent_id, target = handoffagent_name/name).Why
AAASM-990 AC requires Python SDK adapter emission hooks so the topology store records real node-to-node and delegation edges at runtime. The gateway-side
ReportEdgegRPC RPC was merged in PR #254; this PR adds the Python SDK side.How to verify
uv sync .venv/bin/python -m pytest test/unit/adapters/langgraph/test_edge_emission.py \ test/unit/adapters/openai_agents/test_edge_emission.py -v # All 8 tests pass .venv/bin/python -m pytest test/unit/ -q # 258 passedCloses #AAASM-990