Files
narratio/internal/stage/archive.go

583 lines
17 KiB
Go

package stage
import (
"context"
"encoding/json"
"fmt"
"io/fs"
"os"
"path/filepath"
"sort"
"strings"
"time"
"gitea.maximumdirect.net/eric/narratio/internal/adapters/storage"
"gitea.maximumdirect.net/eric/narratio/internal/artifacts"
"gitea.maximumdirect.net/eric/narratio/internal/config"
"gitea.maximumdirect.net/eric/narratio/internal/manifest"
)
type archiveStage struct{}
type archiveUploadFile struct {
RelativePath string
LocalPath string
}
var archivePrerequisiteStages = []string{
"prepare",
"transcribe",
"merge",
"polish",
"normalize",
"trim",
"analyze",
}
func (archiveStage) Name() string { return "archive" }
func (archiveStage) Declares() IODecl {
return IODecl{
Inputs: []artifacts.Ref{
{Kind: "manifest", Category: "input", RelativePath: "manifest.json"},
},
}
}
func (archiveStage) Run(ctx context.Context, env *Env, m *manifest.Manifest) (*StageResult, error) {
if env == nil || env.Config == nil || env.Config.Pipeline == nil || env.Config.Session == nil {
return nil, fmt.Errorf("archive: resolved config must include pipeline and session")
}
if archiveDisabled(env) {
return &StageResult{
Metadata: map[string]any{
"stage": "archive",
"skipped": true,
"archive_enabled": false,
"audio_upload_skipped": true,
"current_pointer_written": false,
},
}, nil
}
if archiveRunUploadDisabled(env) {
return &StageResult{
Metadata: map[string]any{
"stage": "archive",
"skipped": true,
"upload_run_enabled": false,
"audio_upload_skipped": true,
"current_pointer_written": false,
},
}, nil
}
if err := validateArchivePrerequisites(m); err != nil {
return nil, fmt.Errorf("archive: %w", err)
}
if env.ObjectStore == nil {
return nil, fmt.Errorf("archive: remote object store backend is required when archive run upload is enabled")
}
runRoot, err := resolveArchiveRunRoot(env, m)
if err != nil {
return nil, fmt.Errorf("archive: resolve run root: %w", err)
}
runRootInfo, err := os.Stat(runRoot)
if err != nil {
return nil, fmt.Errorf("archive: run root %q: %w", runRoot, err)
}
if !runRootInfo.IsDir() {
return nil, fmt.Errorf("archive: run root %q is not a directory", runRoot)
}
runPrefix, err := archiveRunPrefix(env, m)
if err != nil {
return nil, fmt.Errorf("archive: resolve s3 run prefix: %w", err)
}
sessionPrefix, err := archiveSessionPrefix(env, m)
if err != nil {
return nil, fmt.Errorf("archive: resolve s3 session prefix: %w", err)
}
bucket := archiveBucket(env, m)
if bucket == "" {
return nil, fmt.Errorf("archive: resolve s3 bucket: bucket is required")
}
runID := strings.TrimSpace(m.RunID)
if runID == "" {
return nil, fmt.Errorf("archive: run id is required")
}
manifestSource, err := resolveArchiveRunManifestSource(runRoot)
if err != nil {
return nil, fmt.Errorf("archive: resolve run manifest source: %w", err)
}
runFiles, err := collectArchiveRunFiles(runRoot, manifestSource)
if err != nil {
return nil, fmt.Errorf("archive: collect run files: %w", err)
}
sessionRoot, err := resolveArchiveSessionRoot(env, m)
if err != nil {
return nil, fmt.Errorf("archive: resolve session root for promotions: %w", err)
}
promotions, err := resolveArchivePromotions(sessionRoot, env.Config.Pipeline.Archive.PromoteArtifacts)
if err != nil {
return nil, fmt.Errorf("archive: resolve promotion rules: %w", err)
}
runUploaded := make([]string, 0, len(runFiles))
for _, file := range runFiles {
key := artifacts.S3RunRelativeDestinationKey(runPrefix, file.RelativePath)
if _, err := env.ObjectStore.Upload(ctx, file.LocalPath, key, storage.UploadOptions{}); err != nil {
return nil, fmt.Errorf("archive: upload run file %q to %q: %w", file.RelativePath, key, err)
}
runUploaded = append(runUploaded, file.RelativePath)
}
promotedUploaded := make([]string, 0, len(promotions))
skippedOptional := make([]string, 0)
for _, promotion := range promotions {
if !promotion.Exists {
if promotion.Required {
return nil, fmt.Errorf("archive: required promotion source missing: %q", promotion.From)
}
skippedOptional = append(skippedOptional, promotion.To)
continue
}
key := artifacts.S3PromotedArtifactKey(sessionPrefix, promotion.To)
if _, err := env.ObjectStore.Upload(ctx, promotion.LocalPath, key, storage.UploadOptions{}); err != nil {
return nil, fmt.Errorf("archive: upload promoted output %q to %q: %w", promotion.From, key, err)
}
promotedUploaded = append(promotedUploaded, promotion.To)
}
currentManifestKey := artifacts.S3CurrentManifestKey(sessionPrefix)
manifestTempPath, err := writeCurrentManifestSnapshot(m, archiveMetadataPreview(
bucket,
runPrefix,
sessionPrefix,
runUploaded,
promotedUploaded,
skippedOptional,
currentManifestKey,
))
if err != nil {
return nil, fmt.Errorf("archive: build current manifest snapshot: %w", err)
}
defer func() { _ = os.Remove(manifestTempPath) }()
if _, err := env.ObjectStore.Upload(ctx, manifestTempPath, currentManifestKey, storage.UploadOptions{
ContentType: "application/json",
}); err != nil {
return nil, fmt.Errorf("archive: upload current manifest to %q: %w", currentManifestKey, err)
}
currentRunPointerKey := artifacts.S3CurrentRunPointerKey(sessionPrefix)
runIDTempPath, err := writeCurrentRunIDPointer(runID)
if err != nil {
return nil, fmt.Errorf("archive: build current run id pointer: %w", err)
}
defer func() { _ = os.Remove(runIDTempPath) }()
if _, err := env.ObjectStore.Upload(ctx, runIDTempPath, currentRunPointerKey, storage.UploadOptions{
ContentType: "text/plain; charset=utf-8",
}); err != nil {
return nil, fmt.Errorf("archive: upload current run pointer to %q: %w", currentRunPointerKey, err)
}
return &StageResult{
Metadata: map[string]any{
"stage": "archive",
"uploaded": true,
"s3_bucket": bucket,
"s3_run_prefix": runPrefix,
"run_files_uploaded": len(runUploaded),
"run_uploaded_paths": runUploaded,
"promoted_files_uploaded": len(promotedUploaded),
"promoted_paths": promotedUploaded,
"skipped_optional_promotions": skippedOptional,
"current_manifest_key": currentManifestKey,
"current_run_id_key": currentRunPointerKey,
"current_pointer_written": true,
"audio_upload_skipped": true,
},
}, nil
}
type archivePromotion struct {
From string
To string
Required bool
LocalPath string
Exists bool
}
func archiveDisabled(env *Env) bool {
cfg := env.Config.Pipeline.Archive
if cfg == nil {
return true
}
return cfg.Enabled != nil && !*cfg.Enabled
}
func archiveRunUploadDisabled(env *Env) bool {
cfg := env.Config.Pipeline.Archive
if cfg == nil {
return true
}
return cfg.UploadRun != nil && !*cfg.UploadRun
}
func validateArchivePrerequisites(m *manifest.Manifest) error {
if m == nil {
return fmt.Errorf("manifest is required")
}
for _, stageName := range archivePrerequisiteStages {
sr := m.Stages[stageName]
if sr == nil {
return fmt.Errorf("prerequisite stage %q has not succeeded", stageName)
}
if sr.Status != manifest.StatusSucceeded {
return fmt.Errorf("prerequisite stage %q status is %q (want %q)", stageName, sr.Status, manifest.StatusSucceeded)
}
}
return nil
}
func resolveArchiveRunRoot(env *Env, m *manifest.Manifest) (string, error) {
sessionID := strings.TrimSpace(env.Config.Session.SessionID)
if sessionID == "" && m != nil {
sessionID = strings.TrimSpace(m.SessionID)
}
campaign := strings.TrimSpace(env.Config.Session.Campaign)
if campaign == "" && m != nil {
campaign = strings.TrimSpace(m.Campaign)
}
runID := ""
if m != nil {
runID = strings.TrimSpace(m.RunID)
}
if sessionID == "" || campaign == "" {
return "", fmt.Errorf("campaign and session id are required")
}
if runID == "" {
return "", fmt.Errorf("run id is required")
}
canonical := filepath.Clean(artifacts.SessionRunRootForCampaign(env.Config.Pipeline.Workspace.Root, campaign, sessionID, runID))
canonicalExists, err := directoryExists(canonical)
if err != nil {
return "", fmt.Errorf("check canonical run root %q: %w", canonical, err)
}
if !canonicalExists {
return "", fmt.Errorf("run root not found for campaign %q session %q run %q at canonical path %q", campaign, sessionID, runID, canonical)
}
return canonical, nil
}
func resolveArchiveSessionRoot(env *Env, m *manifest.Manifest) (string, error) {
sessionID := strings.TrimSpace(env.Config.Session.SessionID)
if sessionID == "" && m != nil {
sessionID = strings.TrimSpace(m.SessionID)
}
campaign := strings.TrimSpace(env.Config.Session.Campaign)
if campaign == "" && m != nil {
campaign = strings.TrimSpace(m.Campaign)
}
if sessionID == "" {
return "", fmt.Errorf("session id is required")
}
if campaign == "" {
return "", fmt.Errorf("campaign is required")
}
return filepath.Clean(artifacts.SessionWorkDirForCampaign(env.Config.Pipeline.Workspace.Root, campaign, sessionID)), nil
}
func archiveRunPrefix(env *Env, m *manifest.Manifest) (string, error) {
runPrefix := strings.TrimSpace(m.S3RunPrefix)
if runPrefix != "" {
return runPrefix, nil
}
sessionPrefix, err := archiveSessionPrefix(env, m)
if err != nil {
return "", err
}
runID := strings.TrimSpace(m.RunID)
if runID == "" {
return "", fmt.Errorf("run id is required")
}
return artifacts.S3RunPrefix(sessionPrefix, runID), nil
}
func archiveSessionPrefix(env *Env, m *manifest.Manifest) (string, error) {
if m != nil && strings.TrimSpace(m.S3SessionPrefix) != "" {
return strings.TrimSpace(m.S3SessionPrefix), nil
}
sessionID := strings.TrimSpace(env.Config.Session.SessionID)
if sessionID == "" {
sessionID = strings.TrimSpace(m.SessionID)
}
campaign := strings.TrimSpace(env.Config.Session.Campaign)
if campaign == "" {
campaign = strings.TrimSpace(m.Campaign)
}
if env.Config.Pipeline.Storage.S3 == nil {
return "", fmt.Errorf("pipeline.storage.s3 configuration is required")
}
sessionPrefix := artifacts.S3SessionPrefix(env.Config.Pipeline.Storage.S3.RootPrefix, campaign, sessionID)
if strings.TrimSpace(sessionPrefix) == "" {
return "", fmt.Errorf("session prefix is required")
}
return sessionPrefix, nil
}
func archiveBucket(env *Env, m *manifest.Manifest) string {
if m != nil && strings.TrimSpace(m.S3Bucket) != "" {
return strings.TrimSpace(m.S3Bucket)
}
if env == nil || env.Config == nil || env.Config.Pipeline == nil || env.Config.Pipeline.Storage.S3 == nil {
return ""
}
return strings.TrimSpace(env.Config.Pipeline.Storage.S3.Bucket)
}
func resolveArchivePromotions(sessionRoot string, rules []config.ArchivePromotionRule) ([]archivePromotion, error) {
sessionRoot = filepath.Clean(strings.TrimSpace(sessionRoot))
if sessionRoot == "" {
return nil, fmt.Errorf("session root is required")
}
out := make([]archivePromotion, 0, len(rules))
for _, rule := range rules {
from := strings.TrimSpace(rule.From)
to := strings.TrimSpace(rule.To)
required := rule.Required == nil || *rule.Required
resolvedPath, err := resolveWorkDirRelativePath(sessionRoot, from)
if err != nil {
return nil, fmt.Errorf("promotion from %q: %w", from, err)
}
info, err := os.Stat(resolvedPath)
exists := err == nil && !info.IsDir()
if err != nil && !os.IsNotExist(err) {
return nil, fmt.Errorf("promotion source %q: %w", from, err)
}
localPath := resolvedPath
out = append(out, archivePromotion{
From: from,
To: to,
Required: required,
LocalPath: localPath,
Exists: exists,
})
}
return out, nil
}
func resolveWorkDirRelativePath(workDir, rel string) (string, error) {
rel = filepath.Clean(filepath.FromSlash(strings.TrimSpace(rel)))
if rel == "." || rel == "" {
return "", fmt.Errorf("relative path is required")
}
full := filepath.Join(workDir, rel)
cleanedWork := filepath.Clean(workDir)
cleanedFull := filepath.Clean(full)
relative, err := filepath.Rel(cleanedWork, cleanedFull)
if err != nil {
return "", fmt.Errorf("compute relative path: %w", err)
}
if relative == ".." || strings.HasPrefix(relative, ".."+string(filepath.Separator)) {
return "", fmt.Errorf("path escapes workdir")
}
return cleanedFull, nil
}
func collectArchiveRunFiles(runRoot, manifestPath string) ([]archiveUploadFile, error) {
files := make([]archiveUploadFile, 0, 64)
err := filepath.WalkDir(runRoot, func(path string, d fs.DirEntry, walkErr error) error {
if walkErr != nil {
return walkErr
}
if d.IsDir() {
if path == runRoot {
return nil
}
relDir, err := filepath.Rel(runRoot, path)
if err != nil {
return fmt.Errorf("relative dir from %q to %q: %w", runRoot, path, err)
}
relDir = filepath.ToSlash(relDir)
// Preserve existing behavior: audio is not uploaded in archive run record.
if relDir == "audio" || strings.HasPrefix(relDir, "audio/") {
return filepath.SkipDir
}
return nil
}
rel, err := filepath.Rel(runRoot, path)
if err != nil {
return fmt.Errorf("relative path from %q to %q: %w", runRoot, path, err)
}
rel = filepath.ToSlash(rel)
files = append(files, archiveUploadFile{
RelativePath: rel,
LocalPath: path,
})
return nil
})
if err != nil {
return nil, fmt.Errorf("walk %q: %w", runRoot, err)
}
manifestInfo, err := os.Stat(manifestPath)
if err != nil {
if os.IsNotExist(err) {
return nil, fmt.Errorf("manifest.json not found (checked path %q)", manifestPath)
}
return nil, fmt.Errorf("stat %q: %w", manifestPath, err)
}
if manifestInfo.IsDir() {
return nil, fmt.Errorf("manifest path %q is a directory", manifestPath)
}
files = append(files, archiveUploadFile{
RelativePath: "manifest.json",
LocalPath: manifestPath,
})
seen := map[string]archiveUploadFile{}
for _, file := range files {
seen[file.RelativePath] = file
}
files = files[:0]
for _, file := range seen {
files = append(files, file)
}
sort.Slice(files, func(i, j int) bool {
return files[i].RelativePath < files[j].RelativePath
})
return files, nil
}
func resolveArchiveRunManifestSource(runRoot string) (string, error) {
path := filepath.Join(filepath.Clean(runRoot), "manifest.json")
info, err := os.Stat(path)
if err != nil {
if os.IsNotExist(err) {
return "", fmt.Errorf("manifest.json not found in run root %q", runRoot)
}
return "", fmt.Errorf("stat %q: %w", path, err)
}
if info.IsDir() {
return "", fmt.Errorf("manifest path %q is a directory", path)
}
return path, nil
}
func directoryExists(path string) (bool, error) {
info, err := os.Stat(path)
if err == nil {
return info.IsDir(), nil
}
if os.IsNotExist(err) {
return false, nil
}
return false, err
}
func writeCurrentManifestSnapshot(m *manifest.Manifest, archiveMetadata map[string]any) (string, error) {
if m == nil {
return "", fmt.Errorf("manifest is required")
}
clone := *m
clone.Stages = make(map[string]*manifest.StageRecord, len(m.Stages))
for name, sr := range m.Stages {
if sr == nil {
continue
}
stageCopy := *sr
if sr.Outputs != nil {
stageCopy.Outputs = append([]manifest.ArtifactRecord(nil), sr.Outputs...)
}
if sr.Logs != nil {
stageCopy.Logs = append([]string(nil), sr.Logs...)
}
if sr.GeneratedConfigs != nil {
stageCopy.GeneratedConfigs = append([]string(nil), sr.GeneratedConfigs...)
}
if sr.Metadata != nil {
metaCopy := make(map[string]any, len(sr.Metadata))
for k, v := range sr.Metadata {
metaCopy[k] = v
}
stageCopy.Metadata = metaCopy
}
clone.Stages[name] = &stageCopy
}
now := time.Now().UTC()
clone.MarkStageSucceeded("archive", now, nil)
if sr := clone.Stages["archive"]; sr != nil {
sr.Metadata = archiveMetadata
}
data, err := json.MarshalIndent(&clone, "", " ")
if err != nil {
return "", fmt.Errorf("marshal manifest: %w", err)
}
data = append(data, '\n')
tmp, err := os.CreateTemp("", "narratio-current-manifest-*.json")
if err != nil {
return "", fmt.Errorf("create temp file: %w", err)
}
path := tmp.Name()
if _, err := tmp.Write(data); err != nil {
_ = tmp.Close()
return "", fmt.Errorf("write temp manifest: %w", err)
}
if err := tmp.Close(); err != nil {
return "", fmt.Errorf("close temp manifest: %w", err)
}
return path, nil
}
func writeCurrentRunIDPointer(runID string) (string, error) {
tmp, err := os.CreateTemp("", "narratio-current-run-id-*.txt")
if err != nil {
return "", fmt.Errorf("create temp file: %w", err)
}
path := tmp.Name()
if _, err := tmp.WriteString(runID + "\n"); err != nil {
_ = tmp.Close()
return "", fmt.Errorf("write temp run id pointer: %w", err)
}
if err := tmp.Close(); err != nil {
return "", fmt.Errorf("close temp run id pointer: %w", err)
}
return path, nil
}
func archiveMetadataPreview(
bucket, runPrefix, sessionPrefix string,
runUploaded []string,
promotedUploaded []string,
skippedOptional []string,
currentManifestKey string,
) map[string]any {
return map[string]any{
"stage": "archive",
"uploaded": true,
"s3_bucket": bucket,
"s3_run_prefix": runPrefix,
"run_files_uploaded": len(runUploaded),
"run_uploaded_paths": append([]string(nil), runUploaded...),
"promoted_files_uploaded": len(promotedUploaded),
"promoted_paths": append([]string(nil), promotedUploaded...),
"skipped_optional_promotions": append([]string(nil), skippedOptional...),
"current_manifest_key": currentManifestKey,
"current_run_id_key": artifacts.S3CurrentRunPointerKey(sessionPrefix),
"current_pointer_written": false,
"audio_upload_skipped": true,
}
}