feat: enable evaluators after resume by using consistent thread_id and Command API #1132
+34
−7
Conversation
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
github-actions
bot
added
test:uipath-langchain
Chibionos
force-pushed
the
feat/enable-evaluators-on-resume
branch
from
c00a487 to
540ed7a
Compare
- Use eval_item.id as runtime_id to maintain consistent thread_id across suspend/resume - Pass UiPathExecuteOptions(resume=True) when in resume mode - Add logging to track resume mode execution This ensures: 1. Checkpoints can be found when resuming (same thread_id) 2. Framework-agnostic approach (uses runtime options, not LangGraph Command) 3. Resume handling is delegated to runtime layer where framework-specific logic belongs
Chibionos
force-pushed
the
feat/enable-evaluators-on-resume
branch
from
540ed7a to
e905c9c
Compare
cristipufu
reviewed
Jan 16, 2026
| # Create a new runtime with runtime_id for this eval execution. | ||
| # Use eval_item.id to maintain consistent thread_id across suspend and resume. | ||
| # This ensures checkpoints can be found when resuming from suspended state. | ||
| runtime_id = eval_item.id |
Member
There was a problem hiding this comment.
I think someone changed this for a reason I can't remember. @akshaylive or @radu-mocanu do you guys remember why we no longer use eval_item.id as the runtime_id?
Contributor
Author
There was a problem hiding this comment.
They made it so there was no conflict in the runtime when multiple evals were started the runtime collided
Contributor
Author
There was a problem hiding this comment.
Need to fix the LangGraph aspect too. Is there a better way @cristipufu it is not resume properly for me without this change as the mapping won't work
Contributor
There was a problem hiding this comment.
as far as I remember the concern was regarding running multiple eval runtimes in parallel targeting the same evaluation set. there would be collisions since eval_item.id is static
During resume, the eval runtime should pass None as input instead of
inputs_with_overrides. This allows the UiPathResumableRuntime wrapper to:
1. Fetch triggers from storage
2. Read resume data via trigger_manager.read_trigger()
3. Build resume map: {interrupt_id: resume_data}
4. Pass this map to the delegate runtime
The wrapper's _restore_resume_input() method handles the resume data
restoration automatically when input is None. Passing the original input
would skip this mechanism and cause the resume to fail.
This completes the fix for evaluator execution after resume, working in
conjunction with the langchain runtime's Command(resume=...) wrapper.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
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
Enables evaluators to execute and produce scores after agent resumes from suspended state.
Changes
Consistent thread_id for suspend/resume
Command API integration
Enhanced logging
Testing
Tested with suspend/resume evaluation samples in uipath-langchain-python:
Built on top of
🤖 Generated with Claude Code