Skip to content

fix: make model field optional in LegacyLlmAsAJudgeEvaluator [AE-1306]#1488

Open
ameyjain wants to merge 5 commits intomainfrom
fix/legacy-llm-evaluator-model-field-required
Open

fix: make model field optional in LegacyLlmAsAJudgeEvaluator [AE-1306]#1488
ameyjain wants to merge 5 commits intomainfrom
fix/legacy-llm-evaluator-model-field-required

Conversation

@ameyjain
Copy link
Contributor

@ameyjain ameyjain commented Mar 25, 2026

Summary

  • model field in LegacyLlmAsAJudgeEvaluator was required (model: str) but evaluator JSON files stored in StudioWeb do not include it
  • This caused a ValidationError: model Field required when running evals on the serverless path
  • Added Field(default="same-as-agent") and imported Field from pydantic

Root Cause

The non-serverless (Temporal) path injects the model inline via C# EvalDefinitionBuilder, so the required field was never an issue there. The serverless path reads evaluator config directly from JSON files which don't have a model property, triggering the Pydantic validation error.

Test plan

  • Run an eval with LLM-as-a-judge evaluator on the serverless path and verify it completes without model Field required error
  • Verify existing non-serverless eval runs are unaffected

JIRA

https://uipath.atlassian.net/browse/AE-1306

Screenshot 2026-03-25 at 1 40 57 PM

@github-actions github-actions bot added test:uipath-langchain Triggers tests in the uipath-langchain-python repository test:uipath-llamaindex Triggers tests in the uipath-llamaindex-python repository labels Mar 25, 2026
The `model` field was required but evaluator JSON files stored in StudioWeb
do not include it. This caused a `ValidationError: model Field required`
when running evals on the serverless path.

Fix adds a default value of `"same-as-agent"` and imports `Field` from pydantic.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
@ameyjain ameyjain force-pushed the fix/legacy-llm-evaluator-model-field-required branch from aa28702 to a2e24db Compare March 25, 2026 17:44
@ameyjain ameyjain changed the title fix: make model field optional in LegacyLlmAsAJudgeEvaluator fix: make model field optional in LegacyLlmAsAJudgeEvaluator [AE-1306] Mar 25, 2026
ameyjain and others added 3 commits March 25, 2026 13:56
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Same fix as for LegacyLlmAsAJudgeEvaluator: the `model` field was required
but evaluator JSON files stored in StudioWeb do not include it, causing
`ValidationError: model Field required` on the serverless eval path.

Affected evaluators:
- LegacyTrajectoryEvaluator
- LegacyFaithfulnessEvaluator
- LegacyContextPrecisionEvaluator

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
@ameyjain ameyjain closed this Mar 26, 2026
@ameyjain ameyjain reopened this Mar 26, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

test:uipath-langchain Triggers tests in the uipath-langchain-python repository test:uipath-llamaindex Triggers tests in the uipath-llamaindex-python repository

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants