eric/audita
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases 6 Wiki Activity

Simplify references and output instructions in LLM prompts

Browse Source
This commit is contained in:
Eric Rakestraw
2026-09-25 12:38:23 +00:00
parent 325c4ea7d3
commit 011b021d47
31 changed files with 248 additions and 129 deletions
Download Patch File Download Diff File Expand all files Collapse all files

26
internal/framework/runner/runner_test.go
View File

@@ -621,7 +621,7 @@ func (c *captureContractStructuredClient) CompleteStructured(ctx context.Context
if target, ok := out.(*validators.LLMValidationResponse); ok {
*target = validators.LLMValidationResponse{
Validations: []validators.LLMValidationDecision{
{CorrectionIndex: 0, Approved: true, Confidence: 0.9, Reason: "ok"},
{CorrectionIndex: 1, Approved: true, Confidence: 0.9, Reason: "ok"},
},
}
}
@@ -653,7 +653,7 @@ func (c *stageAwareStructuredClient) CompleteStructured(ctx context.Context, req
target := out.(*validators.LLMValidationResponse)
*target = validators.LLMValidationResponse{
Validations: []validators.LLMValidationDecision{
{CorrectionIndex: section, Approved: true, Confidence: 0.99, Reason: "ok"},
{CorrectionIndex: 1, Approved: true, Confidence: 0.99, Reason: "ok"},
},
}
return contracts.StructuredCompletionResponse{}, nil
@@ -942,7 +942,7 @@ func (f *fakeStructuredClient) CompleteStructured(ctx context.Context, req contr
}
func TestRunnerLLMValidatorApprovalApplied(t *testing.T) {
client := &fakeStructuredClient{responses: []validators.LLMValidationResponse{{Validations: []validators.LLMValidationDecision{{CorrectionIndex: 0, Approved: true, Confidence: 0.9, Reason: "ok"}}}}}
client := &fakeStructuredClient{responses: []validators.LLMValidationResponse{{Validations: []validators.LLMValidationDecision{{CorrectionIndex: 1, Approved: true, Confidence: 0.9, Reason: "ok"}}}}}
llmValidator, err := validators.NewLLMBackedValidator("spoken_form_plausibility_review", validators.LLMValidatorTypeSpokenFormPlausibility)
if err != nil {
t.Fatalf("NewLLMBackedValidator: %v", err)
@@ -969,7 +969,7 @@ func TestRunnerLLMValidatorApprovalApplied(t *testing.T) {
}
func TestRunnerLLMValidatorRejectionPreventsApplication(t *testing.T) {
client := &fakeStructuredClient{responses: []validators.LLMValidationResponse{{Validations: []validators.LLMValidationDecision{{CorrectionIndex: 0, Approved: false, Confidence: 0.95, Reason: "reject"}}}}}
client := &fakeStructuredClient{responses: []validators.LLMValidationResponse{{Validations: []validators.LLMValidationDecision{{CorrectionIndex: 1, Approved: false, Confidence: 0.95, Reason: "reject"}}}}}
llmValidator, _ := validators.NewLLMBackedValidator("spoken_form_plausibility_review", validators.LLMValidatorTypeSpokenFormPlausibility)
r := New(fakeFactory{modules: map[string]contracts.TranscriptModule{
"m": fakeModule{key: "m", policy: proposals.ReplacementPolicyRequireUnique, validators: []contracts.Validator{llmValidator}, proposeF: func(req contracts.ProposalRequest) ([]proposals.CorrectionProposal, error) {
@@ -1028,8 +1028,8 @@ func TestRunnerLLMValidatorMalformedResponseRejectsBatchAndKeepsPartialProgress(
func TestRunnerReconciledSurvivorsCompleteLaterValidation(t *testing.T) {
client := &fakeStructuredClient{responses: []validators.LLMValidationResponse{{Validations: []validators.LLMValidationDecision{
{CorrectionIndex: 0, Approved: true, Confidence: .9, Reason: "ok"},
{CorrectionIndex: 2, Approved: true, Confidence: .9, Reason: "ok"},
{CorrectionIndex: 1, Approved: true, Confidence: .9, Reason: "ok"},
{CorrectionIndex: 3, Approved: true, Confidence: .9, Reason: "ok"},
{CorrectionIndex: 99, Approved: true, Confidence: .9, Reason: "unexpected"},
}}}}
first, _ := validators.NewLLMBackedValidator("editorial_review", validators.LLMValidatorTypeEditorialReview)
@@ -1095,8 +1095,8 @@ func TestRunnerLLMValidatorMissingDecisionRejectsBatch(t *testing.T) {
func TestRunnerLLMValidatorDuplicateDecisionRejectsBatch(t *testing.T) {
client := &fakeStructuredClient{responses: []validators.LLMValidationResponse{{Validations: []validators.LLMValidationDecision{
{CorrectionIndex: 0, Approved: true, Confidence: 0.9, Reason: "ok"},
{CorrectionIndex: 0, Approved: false, Confidence: 0.9, Reason: "dup"},
{CorrectionIndex: 1, Approved: true, Confidence: 0.9, Reason: "ok"},
{CorrectionIndex: 1, Approved: false, Confidence: 0.9, Reason: "dup"},
}}}}
llmValidator, _ := validators.NewLLMBackedValidator("spoken_form_plausibility_review", validators.LLMValidatorTypeSpokenFormPlausibility)
r := New(fakeFactory{modules: map[string]contracts.TranscriptModule{
@@ -1122,7 +1122,7 @@ func (lenEstimator) EstimateTokens(text string) int { return len(text) }
func TestRunnerLLMValidatorBatching(t *testing.T) {
client := &fakeStructuredClient{responses: []validators.LLMValidationResponse{
{Validations: []validators.LLMValidationDecision{{CorrectionIndex: 0, Approved: true, Confidence: 0.9, Reason: "ok"}}},
{Validations: []validators.LLMValidationDecision{{CorrectionIndex: 1, Approved: true, Confidence: 0.9, Reason: "ok"}}},
{Validations: []validators.LLMValidationDecision{{CorrectionIndex: 1, Approved: true, Confidence: 0.9, Reason: "ok"}}},
}}
llmValidator, _ := validators.NewLLMBackedValidator("spoken_form_plausibility_review", validators.LLMValidatorTypeSpokenFormPlausibility)
@@ -1159,7 +1159,7 @@ func TestRunnerLLMValidatorBatching(t *testing.T) {
func TestRunnerLLMValidatorDiagnosticsWrittenAndRedacted(t *testing.T) {
primarySecret := "runner-primary-secret"
validationSecret := "runner-validation-secret"
client := &fakeStructuredClient{responses: []validators.LLMValidationResponse{{Validations: []validators.LLMValidationDecision{{CorrectionIndex: 0, Approved: true, Confidence: 0.9, Reason: validationSecret}}}}}
client := &fakeStructuredClient{responses: []validators.LLMValidationResponse{{Validations: []validators.LLMValidationDecision{{CorrectionIndex: 1, Approved: true, Confidence: 0.9, Reason: validationSecret}}}}}
llmValidator, _ := validators.NewLLMBackedValidator("spoken_form_plausibility_review", validators.LLMValidatorTypeSpokenFormPlausibility)
t.Setenv("RUNNER_PRIMARY_KEY", primarySecret)
t.Setenv("RUNNER_VALIDATION_KEY", validationSecret)
@@ -1265,7 +1265,7 @@ func TestRunnerProposalGenerationHelperFlowsThroughPipeline(t *testing.T) {
responses: []proposal_generation.StructuredCorrectionSet{
{
Corrections: []proposal_generation.StructuredCorrectionProposal{
{TargetSegmentID: 1, OriginalText: "teh", CorrectedText: "the", Confidence: 0.9},
{SegmentReference: 1, OriginalText: "teh", CorrectedText: "the", Confidence: 0.9},
},
},
},
@@ -1308,7 +1308,7 @@ func TestRunnerGrammarModuleUsesConfidenceThreshold(t *testing.T) {
responses: []proposal_generation.StructuredCorrectionSet{
{
Corrections: []proposal_generation.StructuredCorrectionProposal{
{TargetSegmentID: 1, OriginalText: "hello ,world", CorrectedText: "Hello, world", Confidence: 0.5},
{SegmentReference: 1, OriginalText: "hello ,world", CorrectedText: "Hello, world", Confidence: 0.5},
},
},
},
@@ -1342,7 +1342,7 @@ func TestRunnerGlossaryModuleUsesConfidenceThreshold(t *testing.T) {
responses: []proposal_generation.StructuredCorrectionSet{
{
Corrections: []proposal_generation.StructuredCorrectionProposal{
{TargetSegmentID: 1, OriginalText: "gestures", CorrectedText: "Jesters", Confidence: 0.5},
{SegmentReference: 1, OriginalText: "gestures", CorrectedText: "Jesters", Confidence: 0.5},
},
},
},