634 lines
29 KiB
Go
634 lines
29 KiB
Go
package pipeline
|
|
|
|
import (
|
|
"context"
|
|
"encoding/json"
|
|
"errors"
|
|
"reflect"
|
|
"strings"
|
|
"testing"
|
|
"time"
|
|
|
|
"gitea.maximumdirect.net/eric/notarius/internal/core/source"
|
|
"gitea.maximumdirect.net/eric/notarius/internal/framework/chunkmap"
|
|
"gitea.maximumdirect.net/eric/notarius/internal/framework/contracts"
|
|
)
|
|
|
|
type recordingChunkPlanStore struct {
|
|
record ChunkPlanRecord
|
|
decision ChunkPlanDecision
|
|
loadErr error
|
|
saveErr error
|
|
loads int
|
|
saves int
|
|
saved ChunkPlanRecord
|
|
}
|
|
|
|
func (s *recordingChunkPlanStore) Load(string) (ChunkPlanRecord, ChunkPlanDecision, error) {
|
|
s.loads++
|
|
return s.record, s.decision, s.loadErr
|
|
}
|
|
|
|
func (s *recordingChunkPlanStore) Save(record ChunkPlanRecord) error {
|
|
s.saves++
|
|
s.saved = record
|
|
return s.saveErr
|
|
}
|
|
|
|
type countingChunkValidator struct {
|
|
calls int
|
|
result contracts.ValidationResult
|
|
err error
|
|
}
|
|
|
|
type capturingChunkMapOutput struct {
|
|
requests []contracts.OutputRequest
|
|
}
|
|
|
|
func (*capturingChunkMapOutput) Key() string { return "capture/chunk-map" }
|
|
|
|
func (output *capturingChunkMapOutput) Encode(_ context.Context, request contracts.OutputRequest) (contracts.OutputResult, error) {
|
|
output.requests = append(output.requests, request)
|
|
return contracts.OutputResult{}, nil
|
|
}
|
|
|
|
func (*countingChunkValidator) Name() string { return "test/counting-chunks" }
|
|
func (*countingChunkValidator) ExecutionClass() contracts.ExecutionClass {
|
|
return contracts.ExecutionClassDeterministic
|
|
}
|
|
func (v *countingChunkValidator) Validate(context.Context, contracts.ChunkValidationRequest) (contracts.ValidationResult, error) {
|
|
v.calls++
|
|
return v.result, v.err
|
|
}
|
|
|
|
type manifestChunker struct {
|
|
terminalChunker
|
|
metadata map[string]any
|
|
}
|
|
|
|
func (c manifestChunker) ManifestMetadata() map[string]any { return c.metadata }
|
|
|
|
type llmCountingChunker struct {
|
|
terminalChunker
|
|
llmCalls *int
|
|
}
|
|
|
|
func (c llmCountingChunker) Plan(ctx context.Context, request contracts.ChunkRequest) (contracts.ChunkPlanResult, error) {
|
|
(*c.llmCalls)++
|
|
return c.terminalChunker.Plan(ctx, request)
|
|
}
|
|
|
|
type retryingChunker struct {
|
|
key string
|
|
plan source.ChunkPlan
|
|
calls int
|
|
}
|
|
|
|
func (c *retryingChunker) Key() string { return c.key }
|
|
func (*retryingChunker) ReferenceSlots() []contracts.ReferenceSlot { return nil }
|
|
func (c *retryingChunker) Plan(context.Context, contracts.ChunkRequest) (contracts.ChunkPlanResult, error) {
|
|
c.calls++
|
|
if c.calls == 1 {
|
|
return contracts.ChunkPlanResult{Warnings: []contracts.Warning{{Scope: "discarded", ReasonCode: "retry", Message: "discarded warning"}}}, errors.New("retry generation")
|
|
}
|
|
return contracts.ChunkPlanResult{Plan: source.CloneChunkPlan(c.plan), Warnings: []contracts.Warning{{Scope: "accepted", ReasonCode: "observed", Message: "accepted warning"}}}, nil
|
|
}
|
|
|
|
type dependencyLoader struct {
|
|
CheckpointLoader
|
|
extractDependencies []CheckpointFingerprint
|
|
}
|
|
|
|
func (l *dependencyLoader) Extract(_ string, _ string, dependencies []CheckpointFingerprint) (ExtractCheckpoint, CheckpointDecision) {
|
|
l.extractDependencies = append([]CheckpointFingerprint(nil), dependencies...)
|
|
return ExtractCheckpoint{}, CheckpointDecision{Reason: "not found"}
|
|
}
|
|
|
|
func TestRunnerChunkPlanModeMatrix(t *testing.T) {
|
|
tests := []struct {
|
|
name string
|
|
mode ChunkCacheMode
|
|
decision ChunkPlanDecision
|
|
wantLoads int
|
|
wantSaves int
|
|
wantCalls int
|
|
wantAction string
|
|
}{
|
|
{name: "auto hit", mode: ChunkCacheAuto, decision: ChunkPlanDecision{Status: ChunkPlanHit}, wantLoads: 1, wantAction: "reused"},
|
|
{name: "auto missing", mode: ChunkCacheAuto, decision: ChunkPlanDecision{Status: ChunkPlanMissing}, wantLoads: 1, wantSaves: 1, wantCalls: 1, wantAction: "generated"},
|
|
{name: "auto invalid", mode: ChunkCacheAuto, decision: ChunkPlanDecision{Status: ChunkPlanInvalid}, wantLoads: 1, wantSaves: 1, wantCalls: 1, wantAction: "generated"},
|
|
{name: "bypass", mode: ChunkCacheBypass, wantCalls: 1, wantAction: "bypassed"},
|
|
{name: "empty bypass", wantCalls: 1, wantAction: "bypassed"},
|
|
{name: "refresh", mode: ChunkCacheRefresh, wantSaves: 1, wantCalls: 1, wantAction: "refreshed"},
|
|
}
|
|
for _, tc := range tests {
|
|
t.Run(tc.name, func(t *testing.T) {
|
|
prepared, plan := preparedTerminalDebugPipeline(t)
|
|
calls := 0
|
|
llmCalls := 0
|
|
prepared.chunker = llmCountingChunker{terminalChunker: terminalChunker{key: prepared.resolved.Chunk.Module, plan: plan, calls: &calls}, llmCalls: &llmCalls}
|
|
store := &recordingChunkPlanStore{record: chunkPlanRecord(t, prepared, plan), decision: tc.decision}
|
|
output, err := New().Run(context.Background(), RunInput{Prepared: prepared, RawInput: []byte("input"), ChunkCacheMode: tc.mode, ChunkPlans: store})
|
|
if err != nil {
|
|
t.Fatalf("Run() error = %v", err)
|
|
}
|
|
if store.loads != tc.wantLoads || store.saves != tc.wantSaves || calls != tc.wantCalls {
|
|
t.Fatalf("calls = load %d save %d module %d, want %d %d %d", store.loads, store.saves, calls, tc.wantLoads, tc.wantSaves, tc.wantCalls)
|
|
}
|
|
if output.Manifest.ChunkPlan == nil || output.Manifest.ChunkPlan.Action != tc.wantAction || output.ChunkPlan == nil || output.ChunkPlan.Action != tc.wantAction {
|
|
t.Fatalf("chunk plan manifest = %#v summary = %#v, want action %q", output.Manifest.ChunkPlan, output.ChunkPlan, tc.wantAction)
|
|
}
|
|
})
|
|
}
|
|
}
|
|
|
|
func TestRunnerUsesStatusDerivedChunkPlanSummaryReasons(t *testing.T) {
|
|
for _, tc := range []struct {
|
|
name string
|
|
status ChunkPlanStatus
|
|
wantStatus string
|
|
wantReason string
|
|
}{
|
|
{name: "hit", status: ChunkPlanHit, wantStatus: "hit", wantReason: "stored chunk plan is valid"},
|
|
{name: "missing", status: ChunkPlanMissing, wantStatus: "missing", wantReason: "chunk plan not found"},
|
|
{name: "invalid", status: ChunkPlanInvalid, wantStatus: "invalid", wantReason: "stored chunk plan is invalid"},
|
|
} {
|
|
t.Run(tc.name, func(t *testing.T) {
|
|
prepared, plan := preparedTerminalDebugPipeline(t)
|
|
store := &recordingChunkPlanStore{
|
|
record: chunkPlanRecord(t, prepared, plan),
|
|
decision: ChunkPlanDecision{Status: tc.status, Reason: "SENTINEL_STORE_REASON"},
|
|
}
|
|
output, err := New().Run(context.Background(), RunInput{Prepared: prepared, RawInput: []byte("input"), ChunkCacheMode: ChunkCacheAuto, ChunkPlans: store})
|
|
if err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
if output.ChunkPlan.LookupStatus != tc.wantStatus || output.ChunkPlan.LookupReason != tc.wantReason || strings.Contains(output.ChunkPlan.LookupReason, "SENTINEL_STORE_REASON") {
|
|
t.Fatalf("summary = %#v", output.ChunkPlan)
|
|
}
|
|
})
|
|
}
|
|
|
|
prepared, _ := preparedTerminalDebugPipeline(t)
|
|
output, err := New().Run(context.Background(), RunInput{
|
|
Prepared: prepared, RawInput: []byte("input"), ChunkCacheMode: ChunkCacheAuto,
|
|
ChunkPlans: &recordingChunkPlanStore{loadErr: errors.New("read failed")},
|
|
})
|
|
if err == nil || output.ChunkPlan == nil || output.ChunkPlan.LookupStatus != "skipped" || output.ChunkPlan.LookupReason != "chunk plan lookup skipped" {
|
|
t.Fatalf("operational lookup output=%#v error=%v", output.ChunkPlan, err)
|
|
}
|
|
}
|
|
|
|
func TestRunnerChunkPlanHitUsesStoredProducerProvenance(t *testing.T) {
|
|
prepared, plan := preparedTerminalDebugPipeline(t)
|
|
prepared.resolved.Chunk.Module = "chunk/requested"
|
|
prepared.resolved.Chunk.LLMProfile = "current-profile"
|
|
record := chunkPlanRecord(t, prepared, plan)
|
|
record.Producer.InputModule = "input/original"
|
|
record.Producer.ChunkModule = "chunk/original"
|
|
record.Producer.LLMProfile = "original-profile"
|
|
record.Producer.Metadata = map[string]any{"prompt": map[string]any{"id": "original"}}
|
|
store := &recordingChunkPlanStore{record: record, decision: ChunkPlanDecision{Status: ChunkPlanHit}}
|
|
output, err := New().Run(context.Background(), RunInput{Prepared: prepared, RawInput: []byte("input"), ChunkCacheMode: ChunkCacheAuto, ChunkPlans: store})
|
|
if err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
manifest := output.Manifest.ChunkPlan
|
|
if manifest.RequestedModule != "chunk/requested" || manifest.ProducerModule != "chunk/original" || manifest.ProducerLLMProfile != "original-profile" {
|
|
t.Fatalf("chunk plan manifest = %#v", manifest)
|
|
}
|
|
storedNested := record.Producer.Metadata["prompt"].(map[string]any)
|
|
storedNested["id"] = "mutated"
|
|
if got := manifest.ProducerMetadata["prompt"].(map[string]any)["id"]; got != "original" {
|
|
t.Fatalf("producer metadata changed through stored alias: %v", got)
|
|
}
|
|
}
|
|
|
|
func TestRunnerProvidesAcceptedChunkMapToOutput(t *testing.T) {
|
|
for _, test := range []struct {
|
|
name string
|
|
configure func(*PreparedPipeline, source.ChunkPlan) (RunInput, chunkmap.Producer)
|
|
}{
|
|
{
|
|
name: "bypassed plan",
|
|
configure: func(prepared *PreparedPipeline, _ source.ChunkPlan) (RunInput, chunkmap.Producer) {
|
|
return RunInput{Prepared: prepared, RawInput: []byte("input"), ChunkCacheMode: ChunkCacheBypass}, chunkmap.Producer{
|
|
InputModule: prepared.input.Key(), ChunkModule: prepared.chunker.Key(),
|
|
}
|
|
},
|
|
},
|
|
{
|
|
name: "cached plan",
|
|
configure: func(prepared *PreparedPipeline, plan source.ChunkPlan) (RunInput, chunkmap.Producer) {
|
|
prepared.resolved.Chunk.Module = "chunk/requested"
|
|
record := chunkPlanRecord(t, prepared, plan)
|
|
record.Producer.InputModule = "input/original"
|
|
record.Producer.ChunkModule = "chunk/original"
|
|
record.Producer.LLMProfile = "original-profile"
|
|
return RunInput{
|
|
Prepared: prepared, RawInput: []byte("input"), ChunkCacheMode: ChunkCacheAuto,
|
|
ChunkPlans: &recordingChunkPlanStore{record: record, decision: ChunkPlanDecision{Status: ChunkPlanHit}},
|
|
}, chunkmap.Producer{InputModule: "input/original", ChunkModule: "chunk/original", LLMProfile: "original-profile"}
|
|
},
|
|
},
|
|
} {
|
|
t.Run(test.name, func(t *testing.T) {
|
|
prepared, plan := preparedTerminalDebugPipeline(t)
|
|
encoder := &capturingChunkMapOutput{}
|
|
prepared.output = encoder
|
|
input, wantProducer := test.configure(prepared, plan)
|
|
if _, err := New().Run(context.Background(), input); err != nil {
|
|
t.Fatalf("Run() error = %v", err)
|
|
}
|
|
if len(encoder.requests) != 1 || encoder.requests[0].ChunkMap == nil {
|
|
t.Fatalf("output requests = %#v, want one accepted chunk map", encoder.requests)
|
|
}
|
|
artifact := encoder.requests[0].ChunkMap
|
|
if artifact.Kind != chunkmap.ArtifactKind || artifact.Schema.ID != chunkmap.SchemaID || artifact.MediaType != chunkmap.MediaType {
|
|
t.Fatalf("chunk map artifact = %#v, want fixed serialized identity", artifact)
|
|
}
|
|
value, err := chunkmap.New().Decode(artifact.Content)
|
|
if err != nil {
|
|
t.Fatalf("decode output chunk map: %v", err)
|
|
}
|
|
if value.RequestedChunker != prepared.resolved.Chunk.Module || value.Producer != wantProducer {
|
|
t.Fatalf("chunk map = %#v, want current request and producer %#v", value, wantProducer)
|
|
}
|
|
})
|
|
}
|
|
}
|
|
|
|
func TestRunnerOmitsChunkMapForRejectedPlan(t *testing.T) {
|
|
prepared, _ := preparedTerminalDebugPipeline(t)
|
|
encoder := &capturingChunkMapOutput{}
|
|
prepared.output = encoder
|
|
prepared.chunkValidators.validators = []preparedValidator{{
|
|
resolved: ResolvedValidator{Binding: Binding("reject/chunk"), Target: ValidatorTargetChunk},
|
|
chunk: &countingChunkValidator{result: contracts.ValidationResult{Approved: false, ReasonCode: "rejected", Message: "not accepted"}},
|
|
}}
|
|
output, err := New().Run(context.Background(), RunInput{Prepared: prepared, RawInput: []byte("input")})
|
|
if err != nil {
|
|
t.Fatalf("Run() error = %v", err)
|
|
}
|
|
if len(output.Rejected) != 1 || len(encoder.requests) != 1 || encoder.requests[0].ChunkMap != nil {
|
|
t.Fatalf("output = %#v requests = %#v, want rejected plan without chunk map", output.Rejected, encoder.requests)
|
|
}
|
|
}
|
|
|
|
func TestRunnerRetainsChunkMapAfterLaneRejection(t *testing.T) {
|
|
prepared, _ := preparedTerminalDebugPipeline(t)
|
|
encoder := &capturingChunkMapOutput{}
|
|
prepared.output = encoder
|
|
prepared.Steps[0].lanes[0].mergeValidators.validators = []preparedValidator{{
|
|
resolved: ResolvedValidator{Binding: Binding("reject/lane"), Target: ValidatorTargetTyped, ArtifactKind: "test/notes"},
|
|
typedValidate: func(context.Context, any, typedValidationTarget) (contracts.ValidationResult, error) {
|
|
return contracts.ValidationResult{Approved: false, ReasonCode: "rejected", Message: "not accepted"}, nil
|
|
},
|
|
}}
|
|
output, err := New().Run(context.Background(), RunInput{Prepared: prepared, RawInput: []byte("input")})
|
|
if err != nil {
|
|
t.Fatalf("Run() error = %v", err)
|
|
}
|
|
if len(output.Rejected) != 1 || len(encoder.requests) != 1 || encoder.requests[0].ChunkMap == nil {
|
|
t.Fatalf("output = %#v requests = %#v, want lane rejection with accepted chunk map", output.Rejected, encoder.requests)
|
|
}
|
|
}
|
|
|
|
func TestRunnerChunkMapRequestDoesNotAliasStoredPlan(t *testing.T) {
|
|
prepared, plan := preparedTerminalDebugPipeline(t)
|
|
encoder := &capturingChunkMapOutput{}
|
|
prepared.output = encoder
|
|
record := chunkPlanRecord(t, prepared, plan)
|
|
record.Plan.Annotations = source.ChunkAnnotations{"dnd/scenes": json.RawMessage(`{"title":"Opening"}`)}
|
|
store := &recordingChunkPlanStore{record: record, decision: ChunkPlanDecision{Status: ChunkPlanHit}}
|
|
if _, err := New().Run(context.Background(), RunInput{Prepared: prepared, RawInput: []byte("input"), ChunkCacheMode: ChunkCacheAuto, ChunkPlans: store}); err != nil {
|
|
t.Fatalf("Run() error = %v", err)
|
|
}
|
|
store.record.Plan.Annotations["dnd/scenes"][0] = '['
|
|
value, err := chunkmap.New().Decode(encoder.requests[0].ChunkMap.Content)
|
|
if err != nil {
|
|
t.Fatalf("decode captured chunk map after store mutation: %v", err)
|
|
}
|
|
if string(value.PlanAnnotations["dnd/scenes"]) != `{"title":"Opening"}` {
|
|
t.Fatalf("captured chunk map aliases stored plan: %s", value.PlanAnnotations["dnd/scenes"])
|
|
}
|
|
}
|
|
|
|
func TestRunnerChunkPlanManifestRetainsCandidateOnRejection(t *testing.T) {
|
|
prepared, _ := preparedTerminalDebugPipeline(t)
|
|
validator := &countingChunkValidator{result: contracts.ValidationResult{Approved: false, ReasonCode: "policy", Message: "no"}}
|
|
prepared.chunkValidators.validators = []preparedValidator{{resolved: ResolvedValidator{Binding: Binding(validator.Name()), Target: ValidatorTargetChunk}, chunk: validator}}
|
|
output, err := New().Run(context.Background(), RunInput{Prepared: prepared, RawInput: []byte("input"), ChunkCacheMode: ChunkCacheRefresh, ChunkPlans: &recordingChunkPlanStore{}})
|
|
if err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
if output.Manifest.ChunkPlan.Action != "refreshed" || output.Manifest.ChunkPlan.PlanDigest == "" || output.ChunkPlan.ValidationStatus != "rejected" || output.ChunkPlan.PublicationStatus != "not_published" {
|
|
t.Fatalf("manifest = %#v summary = %#v", output.Manifest.ChunkPlan, output.ChunkPlan)
|
|
}
|
|
}
|
|
|
|
func TestRunnerValidatesChunkPlanPolicyInputs(t *testing.T) {
|
|
for _, tc := range []struct {
|
|
name string
|
|
mode ChunkCacheMode
|
|
want string
|
|
}{
|
|
{name: "auto requires store", mode: ChunkCacheAuto, want: "store is required"},
|
|
{name: "refresh requires store", mode: ChunkCacheRefresh, want: "store is required"},
|
|
{name: "invalid mode", mode: "sometimes", want: "not supported"},
|
|
} {
|
|
t.Run(tc.name, func(t *testing.T) {
|
|
prepared, _ := preparedTerminalDebugPipeline(t)
|
|
_, err := New().Run(context.Background(), RunInput{Prepared: prepared, ChunkCacheMode: tc.mode})
|
|
if err == nil || !strings.Contains(err.Error(), tc.want) {
|
|
t.Fatalf("Run() error = %v, want %q", err, tc.want)
|
|
}
|
|
})
|
|
}
|
|
}
|
|
|
|
func TestRunnerChunkPlanStoreErrorsAreFrameworkErrors(t *testing.T) {
|
|
for _, tc := range []struct {
|
|
name string
|
|
mode ChunkCacheMode
|
|
decision ChunkPlanDecision
|
|
loadErr error
|
|
saveErr error
|
|
wantError string
|
|
wantCalls int
|
|
}{
|
|
{name: "load", mode: ChunkCacheAuto, loadErr: errors.New("read failed"), wantError: "load chunk plan", wantCalls: 0},
|
|
{name: "save", mode: ChunkCacheRefresh, saveErr: errors.New("write failed"), wantError: "save chunk plan", wantCalls: 1},
|
|
} {
|
|
t.Run(tc.name, func(t *testing.T) {
|
|
prepared, plan := preparedTerminalDebugPipeline(t)
|
|
calls := 0
|
|
llmCalls := 0
|
|
prepared.chunker = llmCountingChunker{terminalChunker: terminalChunker{key: prepared.resolved.Chunk.Module, plan: plan, calls: &calls}, llmCalls: &llmCalls}
|
|
store := &recordingChunkPlanStore{decision: tc.decision, loadErr: tc.loadErr, saveErr: tc.saveErr}
|
|
_, err := New().Run(context.Background(), RunInput{Prepared: prepared, RawInput: []byte("input"), ChunkCacheMode: tc.mode, ChunkPlans: store})
|
|
if err == nil || !strings.Contains(err.Error(), tc.wantError) || calls != tc.wantCalls {
|
|
t.Fatalf("Run() error = %v, module calls = %d, want %q and %d", err, calls, tc.wantError, tc.wantCalls)
|
|
}
|
|
})
|
|
}
|
|
}
|
|
|
|
func TestRunnerRegeneratesStructurallyInvalidHit(t *testing.T) {
|
|
prepared, plan := preparedTerminalDebugPipeline(t)
|
|
calls := 0
|
|
prepared.chunker = terminalChunker{key: prepared.resolved.Chunk.Module, plan: plan, calls: &calls}
|
|
invalid := chunkPlanRecord(t, prepared, plan)
|
|
invalid.Plan.Ranges[0].StartUnitID = 999
|
|
store := &recordingChunkPlanStore{record: invalid, decision: ChunkPlanDecision{Status: ChunkPlanHit}}
|
|
if _, err := New().Run(context.Background(), RunInput{Prepared: prepared, RawInput: []byte("input"), ChunkCacheMode: ChunkCacheAuto, ChunkPlans: store}); err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
if calls != 1 || store.loads != 1 || store.saves != 1 {
|
|
t.Fatalf("calls = module %d load %d save %d, want 1 1 1", calls, store.loads, store.saves)
|
|
}
|
|
}
|
|
|
|
func TestRunnerReusesAnnotationsFromDifferentChunkerAsOptionalData(t *testing.T) {
|
|
prepared, plan := preparedTerminalDebugPipeline(t)
|
|
plan.Annotations = source.ChunkAnnotations{"test/chunker": json.RawMessage(`{"label":"fixture"}`)}
|
|
plan.Ranges[0].Annotations = source.ChunkAnnotations{"test/chunker": json.RawMessage(`{"category":"sample"}`)}
|
|
record := chunkPlanRecord(t, prepared, plan)
|
|
record.Producer.ChunkModule = "chunk/producer"
|
|
store := &recordingChunkPlanStore{record: record, decision: ChunkPlanDecision{Status: ChunkPlanHit}}
|
|
output, err := New().Run(context.Background(), RunInput{Prepared: prepared, RawInput: []byte("input"), ChunkCacheMode: ChunkCacheAuto, ChunkPlans: store})
|
|
if err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
if output.Manifest.ChunkPlan.ProducerModule != "chunk/producer" || output.Manifest.ChunkPlan.Action != "reused" {
|
|
t.Fatalf("different-chunker annotation plan was not reused: %#v", output.Manifest.ChunkPlan)
|
|
}
|
|
}
|
|
|
|
func TestRunnerRetriesBeforePublishingAcceptedPlan(t *testing.T) {
|
|
prepared, plan := preparedTerminalDebugPipeline(t)
|
|
prepared.resolved.Chunk.Retries = 1
|
|
chunker := &retryingChunker{key: prepared.resolved.Chunk.Module, plan: plan}
|
|
prepared.chunker = chunker
|
|
store := &recordingChunkPlanStore{decision: ChunkPlanDecision{Status: ChunkPlanMissing}}
|
|
output, err := New().Run(context.Background(), RunInput{Prepared: prepared, RawInput: []byte("input"), ChunkCacheMode: ChunkCacheAuto, ChunkPlans: store})
|
|
if err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
if chunker.calls != 2 || store.saves != 1 {
|
|
t.Fatalf("module calls = %d saves = %d, want 2 and 1", chunker.calls, store.saves)
|
|
}
|
|
if len(output.Warnings) != 1 || output.Warnings[0].Scope != "accepted" || len(store.saved.Warnings) != 1 || store.saved.Warnings[0].Scope != "accepted" {
|
|
t.Fatalf("output warnings = %#v stored warnings = %#v", output.Warnings, store.saved.Warnings)
|
|
}
|
|
}
|
|
|
|
func TestRunnerAutoHitValidatesOnceWithoutRegenerationOrMutation(t *testing.T) {
|
|
tests := []struct {
|
|
name string
|
|
result contracts.ValidationResult
|
|
validatorErr error
|
|
wantError string
|
|
wantReject bool
|
|
}{
|
|
{name: "warning", result: contracts.ValidationResult{Approved: true, Warnings: []contracts.Warning{{Scope: "current", ReasonCode: "observed", Message: "current warning"}}}},
|
|
{name: "rejection", result: contracts.ValidationResult{Approved: false, ReasonCode: "rejected", Message: "rejected hit"}, wantReject: true},
|
|
{name: "error", validatorErr: errors.New("validator failed"), wantError: "validator failed"},
|
|
}
|
|
for _, tc := range tests {
|
|
t.Run(tc.name, func(t *testing.T) {
|
|
prepared, plan := preparedTerminalDebugPipeline(t)
|
|
calls := 0
|
|
llmCalls := 0
|
|
prepared.chunker = llmCountingChunker{terminalChunker: terminalChunker{key: prepared.resolved.Chunk.Module, plan: plan, calls: &calls}, llmCalls: &llmCalls}
|
|
validator := &countingChunkValidator{result: tc.result, err: tc.validatorErr}
|
|
prepared.chunkValidators.validators = []preparedValidator{{resolved: ResolvedValidator{Binding: Binding(validator.Name()), Target: ValidatorTargetChunk}, chunk: validator}}
|
|
record := chunkPlanRecord(t, prepared, plan)
|
|
record.Warnings = []contracts.Warning{{Scope: "stored", ReasonCode: "observed", Message: "stored warning"}}
|
|
store := &recordingChunkPlanStore{record: record, decision: ChunkPlanDecision{Status: ChunkPlanHit}}
|
|
debug := newCapturedDebugRecorder()
|
|
output, err := New().Run(context.Background(), RunInput{Prepared: prepared, RawInput: []byte("input"), ChunkCacheMode: ChunkCacheAuto, ChunkPlans: store, Debug: debug})
|
|
if tc.wantError != "" {
|
|
if err == nil || !strings.Contains(err.Error(), tc.wantError) {
|
|
t.Fatalf("Run() error = %v, want %q", err, tc.wantError)
|
|
}
|
|
} else if err != nil {
|
|
t.Fatalf("Run() error = %v", err)
|
|
}
|
|
if validator.calls != 1 || calls != 0 || llmCalls != 0 || store.loads != 1 || store.saves != 0 {
|
|
t.Fatalf("calls = validator %d module %d llm %d load %d save %d", validator.calls, calls, llmCalls, store.loads, store.saves)
|
|
}
|
|
assertAttemptEnvelopeSequence(t, debug, "chunk")
|
|
if tc.wantError == "" {
|
|
encoded := string(debug.json["chunk/output.json"])
|
|
if !strings.Contains(encoded, `"status":"hit"`) || !strings.Contains(encoded, `"plan":`) || !strings.Contains(encoded, `"materialized_chunks":`) {
|
|
t.Fatalf("chunk hit debug = %s", encoded)
|
|
}
|
|
}
|
|
if tc.wantReject && (len(output.Rejected) != 1 || output.Rejected[0].ReasonCode != "rejected") {
|
|
t.Fatalf("rejected = %#v", output.Rejected)
|
|
}
|
|
if tc.wantError == "" && len(output.Warnings) != 1+len(tc.result.Warnings) {
|
|
t.Fatalf("warnings = %#v, want stored warning once plus current warnings", output.Warnings)
|
|
}
|
|
})
|
|
}
|
|
}
|
|
|
|
func TestRunnerPublishesOnlyAcceptedGeneratedPlans(t *testing.T) {
|
|
for _, tc := range []struct {
|
|
name string
|
|
moduleErr error
|
|
validator contracts.ValidationResult
|
|
cancel bool
|
|
wantCalls int
|
|
wantReject bool
|
|
}{
|
|
{name: "module error", moduleErr: errors.New("generation failed"), wantCalls: 2},
|
|
{name: "validator rejection", validator: contracts.ValidationResult{Approved: false, ReasonCode: "rejected", Message: "rejected generated plan"}, wantCalls: 2, wantReject: true},
|
|
{name: "cancellation", cancel: true},
|
|
} {
|
|
t.Run(tc.name, func(t *testing.T) {
|
|
prepared, plan := preparedTerminalDebugPipeline(t)
|
|
prepared.resolved.Chunk.Retries = 1
|
|
calls := 0
|
|
prepared.chunker = terminalChunker{key: prepared.resolved.Chunk.Module, plan: plan, calls: &calls, err: tc.moduleErr}
|
|
if tc.wantReject {
|
|
validator := &countingChunkValidator{result: tc.validator}
|
|
prepared.chunkValidators.validators = []preparedValidator{{resolved: ResolvedValidator{Binding: Binding(validator.Name()), Target: ValidatorTargetChunk}, chunk: validator}}
|
|
}
|
|
ctx := context.Background()
|
|
if tc.cancel {
|
|
cancelled, cancel := context.WithCancel(ctx)
|
|
cancel()
|
|
ctx = cancelled
|
|
}
|
|
store := &recordingChunkPlanStore{decision: ChunkPlanDecision{Status: ChunkPlanInvalid}}
|
|
output, err := New().Run(ctx, RunInput{Prepared: prepared, RawInput: []byte("input"), ChunkCacheMode: ChunkCacheAuto, ChunkPlans: store})
|
|
if tc.wantReject {
|
|
if err != nil || len(output.Rejected) != 1 {
|
|
t.Fatalf("Run() error = %v rejected = %#v", err, output.Rejected)
|
|
}
|
|
} else if err == nil {
|
|
t.Fatal("Run() error = nil")
|
|
}
|
|
if calls != tc.wantCalls || store.saves != 0 {
|
|
t.Fatalf("module calls = %d, saves = %d; want %d, 0", calls, store.saves, tc.wantCalls)
|
|
}
|
|
})
|
|
}
|
|
}
|
|
|
|
func TestRunnerStoresProducerProvenanceAndProducerWarnings(t *testing.T) {
|
|
prepared, plan := preparedTerminalDebugPipeline(t)
|
|
prepared.resolved.Chunk.LLMProfile = "chunk-profile"
|
|
prepared.resolved.ChunkExecutionClass = contracts.ExecutionClassLLMBacked
|
|
prepared.resolved.ChunkReferences = ResolvedReferenceTarget{
|
|
Stage: StageChunk,
|
|
ReferenceSet: contracts.ReferenceSet{Slots: map[string]contracts.ResolvedReferenceSlot{
|
|
"guide": {Items: []contracts.ReferenceItem{{SlotName: "guide", Digest: "sha256:guide", Origin: contracts.ReferenceOrigin{Type: "file", URI: "file:///guide.txt"}, Content: []byte("sensitive")}}},
|
|
}},
|
|
}
|
|
prepared.chunker = manifestChunker{terminalChunker: terminalChunker{key: prepared.resolved.Chunk.Module, plan: plan, warnings: []contracts.Warning{{Scope: "producer", ReasonCode: "observed", Message: "producer warning"}}}, metadata: map[string]any{"prompt_id": "chunk/prompt"}}
|
|
validator := &countingChunkValidator{result: contracts.ValidationResult{Approved: true, Warnings: []contracts.Warning{{Scope: "validator", ReasonCode: "observed", Message: "validator warning"}}}}
|
|
prepared.chunkValidators.validators = []preparedValidator{{resolved: ResolvedValidator{Binding: Binding(validator.Name()), Target: ValidatorTargetChunk}, chunk: validator}}
|
|
store := &recordingChunkPlanStore{}
|
|
if _, err := New().Run(context.Background(), RunInput{Prepared: prepared, RawInput: []byte("input"), ChunkCacheMode: ChunkCacheRefresh, ChunkPlans: store}); err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
producer := store.saved.Producer
|
|
if producer.InputModule != prepared.input.Key() || producer.ChunkModule != prepared.chunker.Key() || producer.LLMProfile != "chunk-profile" || producer.Metadata["prompt_id"] != "chunk/prompt" {
|
|
t.Fatalf("producer = %#v", producer)
|
|
}
|
|
if len(producer.References) != 1 || producer.References[0].Digest != "sha256:guide" {
|
|
t.Fatalf("producer references = %#v", producer.References)
|
|
}
|
|
if len(store.saved.Warnings) != 1 || store.saved.Warnings[0].Scope != "producer" {
|
|
t.Fatalf("stored warnings = %#v, want only producer warning", store.saved.Warnings)
|
|
}
|
|
}
|
|
|
|
func TestRunnerRejectsUncloneableModuleManifestMetadata(t *testing.T) {
|
|
prepared, plan := preparedTerminalDebugPipeline(t)
|
|
cyclic := make(map[string]any)
|
|
cyclic["self"] = cyclic
|
|
prepared.chunker = manifestChunker{
|
|
terminalChunker: terminalChunker{key: prepared.resolved.Chunk.Module, plan: plan},
|
|
metadata: map[string]any{"cyclic": cyclic},
|
|
}
|
|
_, err := New().Run(context.Background(), RunInput{Prepared: prepared, RawInput: []byte("input")})
|
|
if err == nil || !strings.Contains(err.Error(), `clone manifest metadata for module "chunker": metadata.cyclic.self contains a cycle`) {
|
|
t.Fatalf("Run() error = %v, want contextual module metadata clone failure", err)
|
|
}
|
|
}
|
|
|
|
func TestRunnerOmitsProducerProfileForDeterministicChunker(t *testing.T) {
|
|
prepared, plan := preparedTerminalDebugPipeline(t)
|
|
prepared.resolved.Chunk.LLMProfile = "configured-but-unused"
|
|
prepared.resolved.ChunkExecutionClass = contracts.ExecutionClassDeterministic
|
|
prepared.chunker = terminalChunker{key: prepared.resolved.Chunk.Module, plan: plan}
|
|
store := &recordingChunkPlanStore{}
|
|
output, err := New().Run(context.Background(), RunInput{Prepared: prepared, RawInput: []byte("input"), ChunkCacheMode: ChunkCacheRefresh, ChunkPlans: store})
|
|
if err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
if store.saved.Producer.LLMProfile != "" || output.Manifest.ChunkPlan.ProducerLLMProfile != "" {
|
|
t.Fatalf("deterministic producer profile = stored %q manifest %q", store.saved.Producer.LLMProfile, output.Manifest.ChunkPlan.ProducerLLMProfile)
|
|
}
|
|
}
|
|
|
|
func TestRunnerDoesNotInventProducerProfileForLLMChunker(t *testing.T) {
|
|
prepared, plan := preparedTerminalDebugPipeline(t)
|
|
prepared.resolved.ChunkExecutionClass = contracts.ExecutionClassLLMBacked
|
|
prepared.resolved.Chunk.LLMProfile = ""
|
|
prepared.chunker = terminalChunker{key: prepared.resolved.Chunk.Module, plan: plan}
|
|
store := &recordingChunkPlanStore{}
|
|
|
|
output, err := New().Run(context.Background(), RunInput{Prepared: prepared, RawInput: []byte("input"), ChunkCacheMode: ChunkCacheRefresh, ChunkPlans: store})
|
|
if err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
if store.saved.Producer.LLMProfile != "" || output.Manifest.ChunkPlan.ProducerLLMProfile != "" {
|
|
t.Fatalf("LLM producer profile = stored %q manifest %q, want empty", store.saved.Producer.LLMProfile, output.Manifest.ChunkPlan.ProducerLLMProfile)
|
|
}
|
|
}
|
|
|
|
func TestRunnerRefreshChangesDownstreamChunkFingerprint(t *testing.T) {
|
|
doc := typedTestDocumentWithUnits(2)
|
|
prepared := preparedConcurrentPipeline(t, 1)
|
|
prepared.Steps[0].lanes = prepared.Steps[0].lanes[:1]
|
|
prepared.resolved.Steps[0].ArtifactLanes = prepared.resolved.Steps[0].ArtifactLanes[:1]
|
|
prepared.Steps[0].ArtifactLanes = prepared.Steps[0].ArtifactLanes[:1]
|
|
prepared.input = &typedTestInput{key: prepared.resolved.Input.Module, doc: doc}
|
|
|
|
run := func(plan source.ChunkPlan) []CheckpointFingerprint {
|
|
prepared.chunker = terminalChunker{key: prepared.resolved.Chunk.Module, plan: plan}
|
|
loader := &dependencyLoader{CheckpointLoader: NoopCheckpointLoader()}
|
|
if _, err := New().Run(context.Background(), RunInput{Prepared: prepared, ChunkCacheMode: ChunkCacheRefresh, ChunkPlans: &recordingChunkPlanStore{}, Checkpoint: loader}); err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
return loader.extractDependencies
|
|
}
|
|
separate := typedTestPlan(doc)
|
|
combined := source.ChunkPlan{SourceDigest: doc.Digest, Ranges: []source.ChunkRange{{StartUnitID: 1, EndUnitID: 2}}}
|
|
if first, second := run(separate), run(combined); reflect.DeepEqual(first, second) {
|
|
t.Fatalf("downstream fingerprints did not change: %#v", first)
|
|
}
|
|
}
|
|
|
|
func chunkPlanRecord(t *testing.T, prepared *PreparedPipeline, plan source.ChunkPlan) ChunkPlanRecord {
|
|
t.Helper()
|
|
digest, err := source.DigestChunkPlan(plan)
|
|
if err != nil {
|
|
t.Fatal(err)
|
|
}
|
|
return ChunkPlanRecord{
|
|
SchemaVersion: ChunkPlanSchemaVersion,
|
|
SourceDigest: plan.SourceDigest,
|
|
PlanDigest: digest,
|
|
Plan: source.CloneChunkPlan(plan),
|
|
Producer: ChunkPlanProducer{InputModule: prepared.input.Key(), ChunkModule: prepared.chunker.Key()},
|
|
CreatedAt: time.Now().UTC(),
|
|
}
|
|
}
|