Skip to content

When evals fail, exit with status code 1 #61

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
maraisr merged 1 commit into from
Jun 10, 2025
Merged

When evals fail, exit with status code 1 #61

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
17 changes: 14 additions & 3 deletions cmd/eval/eval.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -48,6 +48,8 @@ type EvaluationResult struct {
Details string `json:"details,omitempty"`
}

var FailedTests = errors.New("❌ Some tests failed.")

// NewEvalCommand returns a new command to evaluate prompts against models
func NewEvalCommand(cfg *command.Config) *cobra.Command {
cmd := &cobra.Command{
Expand Down Expand Up @@ -106,7 +108,14 @@ func NewEvalCommand(cfg *command.Config) *cobra.Command {
jsonOutput: jsonOutput,
}

return handler.runEvaluation(cmd.Context())
err = handler.runEvaluation(cmd.Context())
if err == FailedTests {
// Cobra by default will show the help message when an error occurs,
// which is not what we want for failed evaluations.
// Instead, we just want to exit with a non-zero code.
cmd.SilenceUsage = true
}
return err
},
}

Expand Down Expand Up @@ -206,6 +215,10 @@ func (h *evalCommandHandler) runEvaluation(ctx context.Context) error {
h.printSummary(passedTests, totalTests, passRate)
}

if totalTests-passedTests > 0 {
return FailedTests
}

return nil
}

Expand Down Expand Up @@ -249,8 +262,6 @@ func (h *evalCommandHandler) printSummary(passedTests, totalTests int, passRate

if passedTests == totalTests {
h.cfg.WriteToOut("🎉 All tests passed!\n")
} else {
h.cfg.WriteToOut("❌ Some tests failed.\n")
}
}

Expand Down
6 changes: 3 additions & 3 deletions cmd/eval/eval_test.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -291,7 +291,7 @@ evaluators:
cmd.SetArgs([]string{promptFile})

err = cmd.Execute()
require.NoError(t, err)
require.ErrorIs(t, err, FailedTests)

output := out.String()
require.Contains(t, output, "Failing Test")
Expand Down Expand Up @@ -361,7 +361,7 @@ evaluators:
cmd.SetArgs([]string{"--json", promptFile})

err = cmd.Execute()
require.NoError(t, err)
require.ErrorIs(t, err, FailedTests)

output := out.String()

Expand Down Expand Up @@ -534,7 +534,7 @@ evaluators:
cmd.SetArgs([]string{"--json", promptFile})

err = cmd.Execute()
require.NoError(t, err)
require.ErrorIs(t, err, FailedTests)

output := out.String()

Expand Down
Loading