feat: Governance runtime implementation#899
Open
viswa-uipath wants to merge 7 commits into
Open
Conversation
There was a problem hiding this comment.
Pull request overview
Adds a LangChain/LangGraph governance integration layer to uipath-langchain, wiring UiPath’s governance evaluator into LangChain callbacks and exposing registration via an entry point so uipath-runtime can discover the adapter.
Changes:
- Introduces a
LangChainAdapter+ governed proxy wrapper and a callback handler that callsEvaluatorProtocolhooks for model/tool events. - Adds governance adapter registration (
uipath.governance.adaptersentry point) with import-time idempotent self-registration. - Updates dependency/lock configuration to pull in updated
uipath-coreand a dev/test build ofuipath-runtime(currently with non-portable lockfile sources).
Reviewed changes
Copilot reviewed 3 out of 4 changed files in this pull request and generated 9 comments.
| File | Description |
|---|---|
src/uipath_langchain/governance/adapter.py |
Implements the LangChain/LangGraph adapter, wrapper, and callback handler that triggers governance evaluations. |
src/uipath_langchain/governance/__init__.py |
Registers the governance adapter on import and exposes an entry-point registration function. |
pyproject.toml |
Adds governance adapter entry point; bumps uipath-core; pins uipath-runtime to a dev build and adds a uv TestPyPI source override. |
uv.lock |
Updates locked dependencies, but currently includes a machine-local Windows editable uipath-runtime source. |
Comment on lines
+8
to
+13
| Intercepts: | ||
|
|
||
| - ``on_chain_start`` / ``on_chain_end`` → BEFORE_AGENT / AFTER_AGENT | ||
| - ``on_llm_start`` / ``on_chat_model_start`` / ``on_llm_end`` → BEFORE_MODEL / AFTER_MODEL | ||
| - ``on_tool_start`` / ``on_tool_end`` → TOOL_CALL / AFTER_TOOL | ||
|
|
Comment on lines
+110
to
+117
| if hasattr(agent, "callbacks"): | ||
| existing = getattr(agent, "callbacks", None) or [] | ||
| if isinstance(existing, list): | ||
| existing.append(callback) | ||
| agent.callbacks = existing | ||
| else: | ||
| agent.callbacks = [callback] | ||
| logger.debug("Injected governance callback via agent.callbacks") |
Comment on lines
+120
to
+126
| if hasattr(agent, "config"): | ||
| config = agent.config or {} | ||
| callbacks = config.get("callbacks", []) | ||
| callbacks.append(callback) | ||
| config["callbacks"] = callbacks | ||
| agent.config = config | ||
| logger.debug("Injected governance callback via agent.config") |
Comment on lines
+175
to
+182
| if config is None: | ||
| config = {} | ||
| if isinstance(config, dict): | ||
| callbacks = config.get("callbacks", []) | ||
| if self._callback not in callbacks: | ||
| callbacks.append(self._callback) | ||
| config["callbacks"] = callbacks | ||
| return config |
Comment on lines
+386
to
+395
| tool_name = serialized.get("name", "unknown") | ||
| tool_args = inputs or {"input": input_str} | ||
| self._evaluator.evaluate_tool_call( | ||
| tool_name=tool_name, | ||
| tool_args=tool_args, | ||
| agent_name=self._agent_name, | ||
| runtime_id=self._session_id, | ||
| trace_id=self._trace_id, | ||
| session_state=self._session_state, | ||
| ) |
Comment on lines
+401
to
+411
| def on_tool_end(self, output: Any, **kwargs: Any) -> None: | ||
| """Evaluate AFTER_TOOL rules at tool end.""" | ||
| try: | ||
| tool_result = str(output) if output is not None else "" | ||
| self._evaluator.evaluate_after_tool( | ||
| tool_name="unknown", | ||
| tool_result=tool_result, | ||
| agent_name=self._agent_name, | ||
| runtime_id=self._session_id, | ||
| trace_id=self._trace_id, | ||
| ) |
Comment on lines
+83
to
+105
| def attach( | ||
| self, | ||
| agent: Any, | ||
| agent_id: str, | ||
| session_id: str, | ||
| evaluator: EvaluatorProtocol, | ||
| ) -> "GovernedLangChainAgent": | ||
| """Attach governance to a LangChain / LangGraph agent.""" | ||
| callback = GovernanceCallbackHandler( | ||
| evaluator=evaluator, | ||
| agent_name=agent_id, | ||
| session_id=session_id, | ||
| ) | ||
| self._inject_callback(agent, callback) | ||
| return GovernedLangChainAgent( | ||
| agent=agent, | ||
| adapter=self, | ||
| agent_id=agent_id, | ||
| session_id=session_id, | ||
| evaluator=evaluator, | ||
| callback=callback, | ||
| ) | ||
|
|
Comment on lines
8
to
13
| "uipath>=2.10.79, <2.11.0", | ||
| "uipath-core>=0.5.17, <0.6.0", | ||
| "uipath-core>=0.5.18, <0.6.0", | ||
| "uipath-platform>=0.1.61, <0.2.0", | ||
| "uipath-runtime>=0.11.0, <0.12.0", | ||
| "uipath-runtime==0.11.0.dev1001180441", | ||
| "langgraph>=1.1.8, <2.0.0", | ||
| "langchain-core>=1.2.11, <2.0.0", |
Comment on lines
+161
to
+163
| [tool.uv.sources] | ||
| uipath-runtime = { index = "testpypi" } | ||
|
|
Registers a LangChain/LangGraph adapter with the uipath-core adapter registry so GovernanceRuntime can attach BEFORE_MODEL / AFTER_MODEL / TOOL_CALL / AFTER_TOOL hooks via LangChain's callback system. Exposed as a uipath.governance.adapters entry point and self-registers on import. Bumps uipath-core to 0.6.x and uipath-runtime to 0.11.x to pick up the new adapter contracts. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
viswa-uipath
force-pushed
the
feat/governance
branch
from
7e335b2 to
8675ab9
Compare
uipath-runtime is pinned to a dev pre-release served from testpypi. uipath==2.10.79 transitively pins uipath-runtime>=0.11.0,<0.12.0, which excludes pre-releases per PEP 440 (0.11.0.dev* sorts below 0.11.0). uv sync was failing on every CI runner, so no tests ran and the coverage report rendered as 0%. Add prerelease = "allow" plus an override-dependencies entry for uipath-runtime under [tool.uv] so the dev pin can satisfy the umbrella's stable-only constraint. Re-lock so uipath-runtime resolves from testpypi instead of the local editable Windows path that wasn't portable to the Linux/macOS runners. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
caplog.at_level(level, logger=NAME) only sets the level on the named logger; pytest's caplog handler is attached to root, so any message that exceeds the root logger's effective level still gets filtered out before reaching the handler. CI's root logger config blocked the adapter logger's WARNING/DEBUG records — the messages emitted (they appear in the test output), caplog.records was empty, the ``any(...)`` assertions failed. Drop the logger= argument so at_level() adjusts the root logger and caplog reliably captures every record at that severity regardless of runner config. The tests only assert on the message content, not the emitting logger, so the change is behaviour-preserving. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
caplog wasn't catching adapter logger records on the Linux CI runners even though the message reached stderr via lastResort. Some interaction between caplog's root-attached LogCaptureHandler and the package's propagation chain on those runners — fighting it isn't worth it. Switch the warning/debug assertions to mock.patch the adapter module's logger and verify the calls directly. The behaviour under test (exception swallowed, warning emitted, debug breadcrumb on the attach path) is the same; the assertion is just more direct and doesn't depend on logging handler topology. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
CI's mypy config has disallow_any_generics=true and warn_unused_ignores=true, which the test scaffolding tripped on every inline helper class: - bare ``list`` / ``dict`` annotations needed parameterisation - ``# type: ignore[no-untyped-def]`` on the duck-typed agent methods weren't actually needed (mypy lets untyped functions slide outside the package boundary) Annotate ``list[Any]`` / ``dict[str, Any]`` and drop the noisy type-ignore comments. Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
|
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
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.
TEST IN PROGRESS: Changes for lang-chain adapter to support governance
Development Package
uipath pack --nolockto get the latest dev build from this PR (requires version range).