Skip to content

Reopening: Evals does not evaluate all metrics if one fails #1256

Open
@vinod-seshadri-TA

Description

@vinod-seshadri-TA

Describe the bug
This issue is opened as a follow-up to issue #1126 (#1126), which was closed without addressing the concern raised. My original issue described that AgentEvaluator only reports the first failing metric and does not evaluate subsequent metrics, which prevents a complete report of all metric failures. The previous issue was closed without a response to the concerns, and I do not have permissions to reopen it.

To Reproduce

  • Create a test case with multiple metrics in the criteria dict
  • Ensure that the first metric will fail and the second will also fail if evaluated
  • Run the test with pytest
  • Only the first metric failure is reported; subsequent metrics are not evaluated

Expected behavior
All metrics should be evaluated and failures for each should be reported, not just the first one. This ensures a complete overview for debugging and improvement.

Screenshots
N/A

Desktop (please complete the following information):

  • OS: [Please specify]
  • Python version(python -V): [Please specify]
  • ADK version(pip show google-adk): [Please specify]

Additional context
This issue is a direct follow-up to #1126, which was closed without a proper response. Please address the original concern in the follow up comments . Thank you

Metadata

Metadata

Assignees

No one assigned

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions