Centralize Scriptorium input source policy across config, analyze, and previous-cache

internal/artifactpolicy/policy.go

@@ -1,6 +1,7 @@
 package artifactpolicy
 
 import (
+	"errors"
 	"fmt"
 	"regexp"
 	"strings"
@@ -19,6 +20,11 @@ const (
 var configuredSourceRE = regexp.MustCompile(`^narratio\.artifact\.([a-z][a-z0-9_]*)$`)
 var previousSourceRE = regexp.MustCompile(`^narratio\.previous_session\.artifact\.([a-z][a-z0-9_]*)$`)
 
+var (
+	ErrUnsupportedScriptoriumInputSource = errors.New("unsupported scriptorium input source")
+	ErrInvalidPreviousSessionSource      = errors.New("invalid previous-session source format")
+)
+
 type SourceKind string
 
 const (
@@ -34,6 +40,28 @@ type Source struct {
 	ConfiguredKey string
 }
 
+// ScriptoriumInputSourceDescriptor describes one validated Scriptorium input source.
+type ScriptoriumInputSourceDescriptor struct {
+	Source          Source
+	PreviousSession *PreviousSessionSourceDescriptor
+}
+
+// PreviousSessionSourceDescriptor describes one canonical previous-session input source.
+type PreviousSessionSourceDescriptor struct {
+	SourceID           string
+	ConfiguredKey      string
+	ConfiguredSourceID string
+}
+
+// UnknownConfiguredArtifactError reports a source that references an undefined configured artifact key.
+type UnknownConfiguredArtifactError struct {
+	ConfiguredKey string
+}
+
+func (e *UnknownConfiguredArtifactError) Error() string {
+	return fmt.Sprintf("references unknown artifact %q", e.ConfiguredKey)
+}
+
 // ConfiguredSourceID converts a configured artifact key into source id form.
 func ConfiguredSourceID(key string) string {
 	return configuredSourcePrefix + strings.TrimSpace(key)
@@ -83,6 +111,70 @@ func ClassifySource(source string) (Source, error) {
 	return Source{}, fmt.Errorf("unsupported artifact source %q", source)
 }
 
+// DescribeScriptoriumInputSource classifies one input source and returns descriptor
+// metadata used by config validation, analyze input resolution, and previous-cache planning.
+func DescribeScriptoriumInputSource(source string) (ScriptoriumInputSourceDescriptor, error) {
+	trimmed := strings.TrimSpace(source)
+	if trimmed == "" {
+		return ScriptoriumInputSourceDescriptor{}, ErrUnsupportedScriptoriumInputSource
+	}
+	if strings.HasPrefix(trimmed, "narratio.previous_session.artifact") {
+		descriptor, err := DescribePreviousSessionSource(trimmed)
+		if err != nil {
+			return ScriptoriumInputSourceDescriptor{}, err
+		}
+		return ScriptoriumInputSourceDescriptor{
+			Source: Source{
+				ID:            descriptor.SourceID,
+				Kind:          SourceKindPreviousArtifact,
+				ConfiguredKey: descriptor.ConfiguredKey,
+			},
+			PreviousSession: &descriptor,
+		}, nil
+	}
+
+	classified, err := ClassifySource(trimmed)
+	if err != nil {
+		return ScriptoriumInputSourceDescriptor{}, ErrUnsupportedScriptoriumInputSource
+	}
+	return ScriptoriumInputSourceDescriptor{Source: classified}, nil
+}
+
+// DescribePreviousSessionSource validates a canonical previous-session source id
+// and returns both previous and configured-source vocabulary descriptors.
+func DescribePreviousSessionSource(source string) (PreviousSessionSourceDescriptor, error) {
+	configuredKey, ok := ParsePreviousSessionSource(source)
+	if !ok {
+		return PreviousSessionSourceDescriptor{}, ErrInvalidPreviousSessionSource
+	}
+	return PreviousSessionSourceDescriptor{
+		SourceID:           PreviousSessionSourceID(configuredKey),
+		ConfiguredKey:      configuredKey,
+		ConfiguredSourceID: ConfiguredSourceID(configuredKey),
+	}, nil
+}
+
+// PreviousSessionSourceDescriptorForConfiguredKey derives a previous-session source descriptor
+// from a configured artifact key.
+func PreviousSessionSourceDescriptorForConfiguredKey(configuredKey string) (PreviousSessionSourceDescriptor, error) {
+	return DescribePreviousSessionSource(PreviousSessionSourceID(configuredKey))
+}
+
+// ValidateInputConfiguredReference checks that configured/previous-session sources
+// reference configured artifacts known to the current Scriptorium config.
+func ValidateInputConfiguredReference(
+	descriptor ScriptoriumInputSourceDescriptor,
+	configured map[string]struct{},
+) error {
+	switch descriptor.Source.Kind {
+	case SourceKindConfiguredArtifact, SourceKindPreviousArtifact:
+		if _, ok := configured[descriptor.Source.ConfiguredKey]; !ok {
+			return &UnknownConfiguredArtifactError{ConfiguredKey: descriptor.Source.ConfiguredKey}
+		}
+	}
+	return nil
+}
+
 // ValidatePublishSource validates that a source is publish-compatible and references a known configured artifact.
 func ValidatePublishSource(source string, configured map[string]string) (Source, error) {
 	classified, err := ClassifySource(source)

internal/artifactpolicy/policy_test.go

@@ -1,6 +1,7 @@
 package artifactpolicy
 
 import (
+	"errors"
 	"strings"
 	"testing"
 )
@@ -90,3 +91,101 @@ func TestResolvePublishedDestinationRejectsTraversal(t *testing.T) {
 		t.Fatal("ResolvePublishedDestination() error = nil, want traversal rejection")
 	}
 }
+
+func TestDescribeScriptoriumInputSource(t *testing.T) {
+	tests := []struct {
+		name        string
+		source      string
+		wantKind    SourceKind
+		wantKey     string
+		wantPrev    bool
+		wantErr     error
+		wantErrLike string
+	}{
+		{name: "built in", source: "narratio.transcript.final_trimmed", wantKind: SourceKindBuiltIn},
+		{name: "configured", source: "narratio.artifact.session_recap", wantKind: SourceKindConfiguredArtifact, wantKey: "session_recap"},
+		{name: "previous", source: "narratio.previous_session.artifact.session_recap", wantKind: SourceKindPreviousArtifact, wantKey: "session_recap", wantPrev: true},
+		{name: "invalid previous", source: "narratio.previous_session.artifact.", wantErr: ErrInvalidPreviousSessionSource},
+		{name: "unsupported", source: "narratio.unknown", wantErr: ErrUnsupportedScriptoriumInputSource},
+		{name: "empty", source: " ", wantErr: ErrUnsupportedScriptoriumInputSource},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			got, err := DescribeScriptoriumInputSource(tt.source)
+			if tt.wantErr != nil {
+				if !errors.Is(err, tt.wantErr) {
+					t.Fatalf("DescribeScriptoriumInputSource() error = %v, want %v", err, tt.wantErr)
+				}
+				return
+			}
+			if tt.wantErrLike != "" {
+				if err == nil || !strings.Contains(err.Error(), tt.wantErrLike) {
+					t.Fatalf("DescribeScriptoriumInputSource() error = %v, want like %q", err, tt.wantErrLike)
+				}
+				return
+			}
+			if err != nil {
+				t.Fatalf("DescribeScriptoriumInputSource() error = %v", err)
+			}
+			if got.Source.Kind != tt.wantKind {
+				t.Fatalf("DescribeScriptoriumInputSource().Source.Kind = %q, want %q", got.Source.Kind, tt.wantKind)
+			}
+			if got.Source.ConfiguredKey != tt.wantKey {
+				t.Fatalf("DescribeScriptoriumInputSource().Source.ConfiguredKey = %q, want %q", got.Source.ConfiguredKey, tt.wantKey)
+			}
+			if tt.wantPrev && got.PreviousSession == nil {
+				t.Fatal("DescribeScriptoriumInputSource().PreviousSession = nil, want descriptor")
+			}
+			if !tt.wantPrev && got.PreviousSession != nil {
+				t.Fatalf("DescribeScriptoriumInputSource().PreviousSession = %#v, want nil", got.PreviousSession)
+			}
+		})
+	}
+}
+
+func TestValidateInputConfiguredReference(t *testing.T) {
+	configured := map[string]struct{}{"session_recap": {}}
+
+	desc, err := DescribeScriptoriumInputSource("narratio.artifact.session_recap")
+	if err != nil {
+		t.Fatalf("DescribeScriptoriumInputSource(configured) error = %v", err)
+	}
+	if err := ValidateInputConfiguredReference(desc, configured); err != nil {
+		t.Fatalf("ValidateInputConfiguredReference(configured) error = %v", err)
+	}
+
+	prevDesc, err := DescribeScriptoriumInputSource("narratio.previous_session.artifact.session_recap")
+	if err != nil {
+		t.Fatalf("DescribeScriptoriumInputSource(previous) error = %v", err)
+	}
+	if err := ValidateInputConfiguredReference(prevDesc, configured); err != nil {
+		t.Fatalf("ValidateInputConfiguredReference(previous) error = %v", err)
+	}
+
+	missingDesc, err := DescribeScriptoriumInputSource("narratio.artifact.quest_log")
+	if err != nil {
+		t.Fatalf("DescribeScriptoriumInputSource(missing configured) error = %v", err)
+	}
+	err = ValidateInputConfiguredReference(missingDesc, configured)
+	var unknown *UnknownConfiguredArtifactError
+	if !errors.As(err, &unknown) || unknown.ConfiguredKey != "quest_log" {
+		t.Fatalf("ValidateInputConfiguredReference(missing configured) error = %v, want UnknownConfiguredArtifactError(quest_log)", err)
+	}
+}
+
+func TestPreviousSessionSourceDescriptorForConfiguredKey(t *testing.T) {
+	got, err := PreviousSessionSourceDescriptorForConfiguredKey("session_recap")
+	if err != nil {
+		t.Fatalf("PreviousSessionSourceDescriptorForConfiguredKey() error = %v", err)
+	}
+	if got.SourceID != "narratio.previous_session.artifact.session_recap" {
+		t.Fatalf("SourceID = %q, want narratio.previous_session.artifact.session_recap", got.SourceID)
+	}
+	if got.ConfiguredSourceID != "narratio.artifact.session_recap" {
+		t.Fatalf("ConfiguredSourceID = %q, want narratio.artifact.session_recap", got.ConfiguredSourceID)
+	}
+	if got.ConfiguredKey != "session_recap" {
+		t.Fatalf("ConfiguredKey = %q, want session_recap", got.ConfiguredKey)
+	}
+}

internal/artifacts/previous_requirements.go

@@ -5,6 +5,7 @@ import (
 	"sort"
 	"strings"
 
+	"gitea.maximumdirect.net/eric/narratio/internal/artifactpolicy"
 	"gitea.maximumdirect.net/eric/narratio/internal/config"
 )
 
@@ -36,10 +37,11 @@ func CollectPreviousArtifactRequirements(
 		inputNames := sortedScriptoriumInputKeys(artifactCfg.Inputs)
 		for _, inputName := range inputNames {
 			inputCfg := artifactCfg.Inputs[inputName]
-			previousName, ok := PreviousSessionArtifactName(inputCfg.Source)
+			descriptor, err := artifactpolicy.DescribeScriptoriumInputSource(inputCfg.Source)
-			if !ok {
+			if err != nil || descriptor.PreviousSession == nil {
 				continue
 			}
+			previousName := descriptor.PreviousSession.ConfiguredKey
 
 			location := fmt.Sprintf(
 				"pipeline.scriptorium.artifacts.%s.inputs.%s.source",

internal/config/validate.go

@@ -9,7 +9,6 @@ import (
 	"strings"
 	"time"
 
-	"gitea.maximumdirect.net/eric/narratio/internal/artifactmodel"
 	"gitea.maximumdirect.net/eric/narratio/internal/artifactpolicy"
 	"gitea.maximumdirect.net/eric/narratio/internal/pathsafe"
 )
@@ -639,13 +638,9 @@ var scriptoriumArtifactKeyRE = regexp.MustCompile(`^[a-z][a-z0-9_]*$`)
 
 func validateScriptoriumInputSource(artifactName, inputName, source string, configuredArtifacts map[string]struct{}) (string, error) {
 	trimmedSource := strings.TrimSpace(source)
-	if isStaticSupportedScriptoriumInputSource(trimmedSource) {
+	descriptor, err := artifactpolicy.DescribeScriptoriumInputSource(trimmedSource)
-		return "", nil
+	if err != nil {
-	}
+		if errors.Is(err, artifactpolicy.ErrInvalidPreviousSessionSource) {
-
-	if strings.HasPrefix(trimmedSource, "narratio.previous_session.artifact") {
-		referenced, ok := artifactpolicy.ParsePreviousSessionSource(trimmedSource)
-		if !ok {
 			return "", fmt.Errorf(
 				"pipeline.scriptorium.artifacts.%s.inputs.%s.source %q must reference configured artifact key matching ^[a-z][a-z0-9_]*$",
 				artifactName,
@@ -653,20 +648,6 @@ func validateScriptoriumInputSource(artifactName, inputName, source string, conf
 				source,
 			)
 		}
-		if _, ok := configuredArtifacts[referenced]; !ok {
-			return "", fmt.Errorf(
-				"pipeline.scriptorium.artifacts.%s.inputs.%s.source %q references unknown artifact %q",
-				artifactName,
-				inputName,
-				source,
-				referenced,
-			)
-		}
-		return "", nil
-	}
-
-	referenced, ok := artifactpolicy.ParseConfiguredSource(trimmedSource)
-	if !ok {
 		return "", fmt.Errorf(
 			"pipeline.scriptorium.artifacts.%s.inputs.%s.source %q is unsupported",
 			artifactName,
@@ -674,28 +655,28 @@ func validateScriptoriumInputSource(artifactName, inputName, source string, conf
 			source,
 		)
 	}
-	if _, ok := configuredArtifacts[referenced]; !ok {
+	if err := artifactpolicy.ValidateInputConfiguredReference(descriptor, configuredArtifacts); err != nil {
+		var unknownConfigured *artifactpolicy.UnknownConfiguredArtifactError
+		if errors.As(err, &unknownConfigured) {
+			return "", fmt.Errorf(
+				"pipeline.scriptorium.artifacts.%s.inputs.%s.source %q references unknown artifact %q",
+				artifactName,
+				inputName,
+				source,
+				unknownConfigured.ConfiguredKey,
+			)
+		}
 		return "", fmt.Errorf(
-			"pipeline.scriptorium.artifacts.%s.inputs.%s.source %q references unknown artifact %q",
+			"pipeline.scriptorium.artifacts.%s.inputs.%s.source %q is unsupported",
 			artifactName,
 			inputName,
 			source,
-			referenced,
 		)
 	}
-	return referenced, nil
+	if descriptor.Source.Kind == artifactpolicy.SourceKindConfiguredArtifact {
-}
+		return descriptor.Source.ConfiguredKey, nil
-
-func isStaticSupportedScriptoriumInputSource(source string) bool {
-	if _, ok := artifactmodel.LookupRuntimeTranscriptArtifact(source); ok {
-		return true
-	}
-	switch source {
-	case "narratio.bounds.session":
-		return true
-	default:
-		return false
 	}
+	return "", nil
 }
 
 func validateEnvVarNameField(fieldName, value string) error {

internal/previouscache/previouscache.go

@@ -229,7 +229,11 @@ func artifactRelativePathCandidates(
 		candidates = append(candidates, normalized)
 	}
 
+	sourceDescriptor, err := artifactpolicy.PreviousSessionSourceDescriptorForConfiguredKey(artifactName)
 	sourceID := artifactpolicy.ConfiguredSourceID(artifactName)
+	if err == nil {
+		sourceID = sourceDescriptor.ConfiguredSourceID
+	}
 	if rel, ok := manifestArtifactRelativePathBySourceID(previousManifest, sourceID); ok {
 		appendCandidate(rel)
 		base := path.Base(rel)

internal/stage/analyze.go

@@ -628,11 +628,11 @@ func resolveScriptoriumInput(
 	runtimeCatalog *artifacts.ArtifactCatalog,
 ) (string, bool, *artifacts.ResolvedSessionArtifact, error) {
 	source := strings.TrimSpace(inputCfg.Source)
-	classified, classifyErr := artifactpolicy.ClassifySource(source)
+	descriptor, describeErr := artifactpolicy.DescribeScriptoriumInputSource(source)
-	if classifyErr != nil {
+	if describeErr != nil {
-		return "", false, nil, classifyErr
+		return "", false, nil, describeErr
 	}
-	if classified.Kind == artifactpolicy.SourceKindPreviousArtifact {
+	if descriptor.Source.Kind == artifactpolicy.SourceKindPreviousArtifact {
 		resolved, err := artifacts.ResolvePreviousSessionArtifactWithCatalog(paths, m, source, runtimeCatalog)
 		if err == nil {
 			copy := resolved
@@ -657,13 +657,13 @@ func resolveScriptoriumInput(
 			return resolved.Path, true, &copy, nil
 		}
 		if errors.Is(err, artifacts.ErrSessionArtifactNotFound) {
-			if classified.Kind == artifactpolicy.SourceKindConfiguredArtifact {
+			if descriptor.Source.Kind == artifactpolicy.SourceKindConfiguredArtifact {
 				if inputCfg.Required {
 					return "", false, nil, fmt.Errorf("configured artifact source %q is unavailable", source)
 				}
 				return "", false, nil, nil
 			}
-			switch classified.ID {
+			switch descriptor.Source.ID {
 			case artifacts.ArtifactTranscriptPolished:
 				return "", false, nil, nil
 			case artifacts.ArtifactTranscriptFinal: