feat(evaluation): unify validators with azureml-assets#47526
Draft
m7md7sien wants to merge 4 commits into
Draft
Conversation
- add DEVELOPER role, EvaluationLevel, MessagesOrQueryResponseInputValidator + level utils - support actions/expected_actions aliases in TaskNavigationEfficiencyValidator - align check_for_unsupported_tools flags in tool_call/input/output evaluators
github-actions
Bot
added
the
Evaluation
Contributor
There was a problem hiding this comment.
Pull request overview
This PR updates azure-ai-evaluation’s internal evaluator input validation layer to better align with azureml-assets naming and behavior, while expanding supported conversation roles and adding utilities for evaluation-level handling.
Changes:
- Added
DEVELOPERmessage role support and introducedEvaluationLevelplus evaluation-level utility helpers. - Added
MessagesOrQueryResponseInputValidatorto support both multi-turn (messages) and single-turn (query/response) input shapes. - Added
actions/expected_actionsaliases for task navigation efficiency inputs, and alignedcheck_for_unsupported_toolsbehavior across tool-related evaluators/validators.
Reviewed changes
Copilot reviewed 9 out of 9 changed files in this pull request and generated 3 comments.
Show a summary per file
| File | Description |
|---|---|
| sdk/evaluation/azure-ai-evaluation/azure/ai/evaluation/_evaluators/_tool_output_utilization/_tool_output_utilization.py | Enables unsupported-tool checking for tool output utilization inputs. |
| sdk/evaluation/azure-ai-evaluation/azure/ai/evaluation/_evaluators/_tool_input_accuracy/_tool_input_accuracy.py | Adjusts unsupported-tool checking behavior for tool input accuracy validation. |
| sdk/evaluation/azure-ai-evaluation/azure/ai/evaluation/_evaluators/_tool_call_accuracy/_tool_call_accuracy.py | Adjusts unsupported-tool checking behavior for tool call accuracy validation. |
| sdk/evaluation/azure-ai-evaluation/azure/ai/evaluation/_evaluators/_common/_validators/_validation_constants.py | Adds DEVELOPER role and introduces the EvaluationLevel enum. |
| sdk/evaluation/azure-ai-evaluation/azure/ai/evaluation/_evaluators/_common/_validators/_task_navigation_efficiency_validator.py | Adds normalization to accept actions/expected_actions aliases. |
| sdk/evaluation/azure-ai-evaluation/azure/ai/evaluation/_evaluators/_common/_validators/_messages_or_query_response_validator.py | New validator supporting either messages or query/response input formats. |
| sdk/evaluation/azure-ai-evaluation/azure/ai/evaluation/_evaluators/_common/_validators/_evaluation_level_utils.py | New helper utilities for resolving evaluation levels and reshaping message inputs. |
| sdk/evaluation/azure-ai-evaluation/azure/ai/evaluation/_evaluators/_common/_validators/_conversation_validator.py | Adds developer-role validation handling and minor error-message cleanup. |
| sdk/evaluation/azure-ai-evaluation/azure/ai/evaluation/_evaluators/_common/_validators/init.py | Exposes new enums/validators/utilities from the validators package. |
Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>
Co-authored-by: m7md7sien <16615690+m7md7sien@users.noreply.github.com>
…luated Co-authored-by: m7md7sien <16615690+m7md7sien@users.noreply.github.com>
mmkawale
reviewed
Jun 18, 2026
| target=self.error_target, | ||
| ) | ||
| # The final assistant message must contain text | ||
| last_content = messages[-1].get("content", "") |
Contributor
There was a problem hiding this comment.
Here we assume that the last message will have a role as assistant, but that may not be the case. Can we explicitly check that the last message's role is assistant before moving on to content check?
| self._validator = ToolCallsValidator( | ||
| error_target=ErrorTarget.TOOL_CALL_ACCURACY_EVALUATOR, | ||
| check_for_unsupported_tools=True, | ||
| check_for_unsupported_tools=False, |
Contributor
There was a problem hiding this comment.
This is the same change I am making in my sdk pr: https://github.com/Azure/azure-sdk-for-python/pull/47462/changes#diff-f0cd98f94f077616907714246b399d03dcc97bde3cde5dbe0ff1dac8c5253869
| error_target=ErrorTarget.TOOL_OUTPUT_UTILIZATION_EVALUATOR, optional_tool_definitions=False | ||
| error_target=ErrorTarget.TOOL_OUTPUT_UTILIZATION_EVALUATOR, | ||
| optional_tool_definitions=False, | ||
| check_for_unsupported_tools=True, |
Contributor
There was a problem hiding this comment.
In assets we pass this flag check_for_unsupported_tools correctly. It would be great to create a matrix for all the built in evals with the expected inputs and outputs along with the values for these flags.
| from ._tool_definitions_validator import ToolDefinitionsValidator | ||
|
|
||
|
|
||
| class MessagesOrQueryResponseInputValidator(ToolDefinitionsValidator): |
Contributor
There was a problem hiding this comment.
Let's add unit tests for these new validators.
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.
Description
All SDK Contribution checklist:
General Guidelines and Best Practices
Testing Guidelines