Use artifact source IDs for archive promotion

This commit is contained in:
2026-05-19 20:05:24 -05:00
parent c5c35cd3b4
commit d001baa660
14 changed files with 432 additions and 124 deletions

View File

@@ -79,8 +79,8 @@ type ArchiveConfig struct {
// ArchivePromotionRule configures one artifact promotion mapping.
type ArchivePromotionRule struct {
From string `yaml:"from"`
To string `yaml:"to"`
Source string `yaml:"source"`
Dest string `yaml:"dest"`
Required *bool `yaml:"required"`
}

View File

@@ -74,8 +74,7 @@ const (
// DefaultArchivePromoteArtifacts defines the default archive promotion rules.
// Callers should copy this slice before mutating.
var DefaultArchivePromoteArtifacts = []ArchivePromotionRule{
{From: PathTranscriptTrimmed, To: PathTranscriptTrimmed},
{From: "artifacts/session_recap.md", To: "artifacts/session_recap.md"},
{Source: "narratio.transcript.trimmed", Dest: PathTranscriptTrimmed},
}
// DefaultPipelineConfigSearchPaths defines the default search order for

View File

@@ -136,40 +136,86 @@ func TestSpoolAndArchiveDefaults(t *testing.T) {
if cfg.Pipeline.Archive.UploadRun == nil || !*cfg.Pipeline.Archive.UploadRun {
t.Fatalf("archive.upload_run = %#v, want true", cfg.Pipeline.Archive.UploadRun)
}
if len(cfg.Pipeline.Archive.PromoteArtifacts) != 2 {
t.Fatalf("archive.promote_artifacts len = %d, want 2 defaults", len(cfg.Pipeline.Archive.PromoteArtifacts))
if len(cfg.Pipeline.Archive.PromoteArtifacts) != 1 {
t.Fatalf("archive.promote_artifacts len = %d, want 1 default", len(cfg.Pipeline.Archive.PromoteArtifacts))
}
for i, item := range cfg.Pipeline.Archive.PromoteArtifacts {
if item.Required == nil || !*item.Required {
t.Fatalf("archive.promote_artifacts[%d].required = %#v, want true", i, item.Required)
}
item := cfg.Pipeline.Archive.PromoteArtifacts[0]
if item.Required == nil || !*item.Required {
t.Fatalf("archive.promote_artifacts[0].required = %#v, want true", item.Required)
}
if item.Source != "narratio.transcript.trimmed" {
t.Fatalf("archive.promote_artifacts[0].source = %q, want narratio.transcript.trimmed", item.Source)
}
if item.Dest != "transcripts/trimmed.json" {
t.Fatalf("archive.promote_artifacts[0].dest = %q, want transcripts/trimmed.json", item.Dest)
}
}
func TestArchivePromotionPathValidation(t *testing.T) {
func TestArchivePromotionValidation(t *testing.T) {
tests := []struct {
name string
ruleYML string
wantErr string
}{
{
name: "absolute from path rejected",
name: "absolute dest path rejected",
ruleYML: `archive:
promote_artifacts:
- from: "/transcripts/trimmed.json"
to: "transcripts/trimmed.json"
- source: "narratio.transcript.trimmed"
dest: "/transcripts/trimmed.json"
`,
wantErr: "must be a relative path",
},
{
name: "traversal to path rejected",
name: "traversal dest path rejected",
ruleYML: `archive:
promote_artifacts:
- from: "transcripts/trimmed.json"
to: "../trimmed.json"
- source: "narratio.transcript.trimmed"
dest: "../trimmed.json"
`,
wantErr: "must not contain path traversal",
},
{
name: "invalid source rejected",
ruleYML: `archive:
promote_artifacts:
- source: "narratio.unknown"
dest: "transcripts/trimmed.json"
`,
wantErr: "source \"narratio.unknown\" is unsupported",
},
{
name: "duplicate destination rejected",
ruleYML: `archive:
promote_artifacts:
- source: "narratio.transcript.trimmed"
dest: "artifacts/shared.md"
- source: "narratio.transcript.full"
dest: "artifacts/shared.md"
`,
wantErr: "duplicates another archive promotion destination",
},
{
name: "configured source requires configured artifact key",
ruleYML: `archive:
promote_artifacts:
- source: "narratio.artifact.session_recap"
dest: "artifacts/session_recap.md"
`,
wantErr: "configured artifact \"session_recap\" is not defined in pipeline.scriptorium.artifacts",
},
{
name: "configured source without output path fails when dest omitted",
ruleYML: `scriptorium:
artifacts:
session_recap:
enabled: false
archive:
promote_artifacts:
- source: "narratio.artifact.session_recap"
`,
wantErr: "destination cannot be derived",
},
}
for _, tt := range tests {
@@ -189,6 +235,72 @@ func TestArchivePromotionPathValidation(t *testing.T) {
}
}
func TestArchivePromotionDerivesDestinationWhenOmitted(t *testing.T) {
tests := []struct {
name string
pipelineYML string
wantDest string
}{
{
name: "built in source derives canonical destination",
pipelineYML: testPipelineBaseYAML + `
archive:
promote_artifacts:
- source: narratio.transcript.full
`,
wantDest: "transcripts/normalized.json",
},
{
name: "configured source derives configured output path",
pipelineYML: testPipelineBaseYAML + `
scriptorium:
artifacts:
session_recap:
enabled: true
prompt_id: dnd.session_recap
output_path: artifacts/session_recap.md
archive:
promote_artifacts:
- source: narratio.artifact.session_recap
`,
wantDest: "artifacts/session_recap.md",
},
}
for _, tt := range tests {
t.Run(tt.name, func(t *testing.T) {
pipelinePath, sessionPath := writeConfigFiles(t, tt.pipelineYML, testSessionBaseYAML)
cfg, err := Load(pipelinePath, sessionPath)
if err != nil {
t.Fatalf("Load() error = %v", err)
}
if err := Validate(cfg); err != nil {
t.Fatalf("Validate() error = %v", err)
}
if len(cfg.Pipeline.Archive.PromoteArtifacts) != 1 {
t.Fatalf("archive.promote_artifacts len = %d, want 1", len(cfg.Pipeline.Archive.PromoteArtifacts))
}
if cfg.Pipeline.Archive.PromoteArtifacts[0].Dest != tt.wantDest {
t.Fatalf("archive.promote_artifacts[0].dest = %q, want %q", cfg.Pipeline.Archive.PromoteArtifacts[0].Dest, tt.wantDest)
}
})
}
}
func TestArchiveLegacyFromToFailsStrictDecode(t *testing.T) {
pipelineYAML := testPipelineBaseYAML + `
archive:
promote_artifacts:
- from: transcripts/trimmed.json
to: transcripts/trimmed.json
`
pipelinePath, sessionPath := writeConfigFiles(t, pipelineYAML, testSessionBaseYAML)
_, err := Load(pipelinePath, sessionPath)
if err == nil || !strings.Contains(err.Error(), "strict decode failed") {
t.Fatalf("Load() error = %v, want strict decode failed", err)
}
}
func TestSessionAudioS3Validation(t *testing.T) {
tests := []struct {
name string

View File

@@ -47,7 +47,7 @@ func validatePipeline(cfg *PipelineConfig) error {
if err := validateSpool(cfg.Spool); err != nil {
return err
}
if err := validateArchive(cfg.Archive); err != nil {
if err := validateArchive(cfg.Archive, cfg.Scriptorium); err != nil {
return err
}
if err := validateWhisperX(cfg.WhisperX); err != nil {
@@ -104,28 +104,91 @@ func validateSpool(cfg SpoolConfig) error {
return nil
}
func validateArchive(cfg *ArchiveConfig) error {
func validateArchive(cfg *ArchiveConfig, scriptorium *ScriptoriumConfig) error {
if cfg == nil {
return nil
}
seenDest := map[string]struct{}{}
for i, item := range cfg.PromoteArtifacts {
prefix := fmt.Sprintf("pipeline.archive.promote_artifacts[%d]", i)
if strings.TrimSpace(item.From) == "" {
return fmt.Errorf("%s.from is required", prefix)
source := strings.TrimSpace(item.Source)
if source == "" {
return fmt.Errorf("%s.source is required", prefix)
}
if strings.TrimSpace(item.To) == "" {
return fmt.Errorf("%s.to is required", prefix)
if _, err := archiveSourceKnown(source, scriptorium); err != nil {
return fmt.Errorf("%s.source %q is unsupported: %w", prefix, item.Source, err)
}
if err := validateRelativeSafePath(prefix+".from", item.From); err != nil {
dest := strings.TrimSpace(item.Dest)
if dest == "" {
derivedDest, err := deriveArchivePromotionDest(source, scriptorium)
if err != nil {
return fmt.Errorf("%s.dest is required when destination cannot be derived from %q: %w", prefix, source, err)
}
dest = derivedDest
cfg.PromoteArtifacts[i].Dest = derivedDest
}
if err := validateRelativeSafePath(prefix+".dest", dest); err != nil {
return err
}
if err := validateRelativeSafePath(prefix+".to", item.To); err != nil {
return err
normalizedDest := filepath.ToSlash(filepath.Clean(dest))
if _, ok := seenDest[normalizedDest]; ok {
return fmt.Errorf("%s.dest %q duplicates another archive promotion destination", prefix, dest)
}
seenDest[normalizedDest] = struct{}{}
}
return nil
}
func archiveSourceKnown(source string, scriptorium *ScriptoriumConfig) (string, error) {
trimmed := strings.TrimSpace(source)
switch trimmed {
case "narratio.transcript.merged",
"narratio.transcript.polished",
"narratio.transcript.full",
"narratio.transcript.trimmed",
"narratio.bounds.session":
return "", nil
}
matches := narratioArtifactSourceRE.FindStringSubmatch(trimmed)
if len(matches) != 2 {
return "", fmt.Errorf("must be a built-in source id or narratio.artifact.<name>")
}
artifactKey := matches[1]
if scriptorium == nil || len(scriptorium.Artifacts) == 0 {
return "", fmt.Errorf("configured artifact %q is not defined in pipeline.scriptorium.artifacts", artifactKey)
}
if _, ok := scriptorium.Artifacts[artifactKey]; !ok {
return "", fmt.Errorf("configured artifact %q is not defined in pipeline.scriptorium.artifacts", artifactKey)
}
return artifactKey, nil
}
func deriveArchivePromotionDest(source string, scriptorium *ScriptoriumConfig) (string, error) {
trimmed := strings.TrimSpace(source)
switch trimmed {
case "narratio.transcript.merged":
return PathTranscriptMerged, nil
case "narratio.transcript.polished":
return PathTranscriptProcessed, nil
case "narratio.transcript.full":
return PathTranscriptNormalized, nil
case "narratio.transcript.trimmed":
return PathTranscriptTrimmed, nil
case "narratio.bounds.session":
return filepath.ToSlash(filepath.Join(PathArtifactsDirSegment, "session_bounds.json")), nil
}
artifactKey, err := archiveSourceKnown(trimmed, scriptorium)
if err != nil {
return "", err
}
artifactCfg := scriptorium.Artifacts[artifactKey]
outputPath := strings.TrimSpace(artifactCfg.OutputPath)
if outputPath == "" {
return "", fmt.Errorf("pipeline.scriptorium.artifacts.%s.output_path is empty", artifactKey)
}
return outputPath, nil
}
func validateSecrets(cfg *SecretsConfig) error {
if cfg == nil {
return nil