Centralize extraction bundle evidence

This commit is contained in:
2026-08-10 22:06:48 +00:00
parent 8ff1b4fa66
commit 4e4eff6ba7
7 changed files with 255 additions and 286 deletions

View File

@@ -1,13 +1,11 @@
package artifacts
import (
"encoding/json"
"os"
"path/filepath"
"strings"
"gitea.maximumdirect.net/eric/narratio/internal/artifactmodel"
"gitea.maximumdirect.net/eric/narratio/internal/fileops"
"gitea.maximumdirect.net/eric/narratio/internal/manifest"
)
@@ -22,11 +20,6 @@ const (
extractionMetadataRoot = "bundle_root"
)
type hydratedExtraction struct {
sourceID string
path string
}
// HydrateExtractionArtifacts marks extraction sources available only when the current
// manifest contains one complete, internally consistent, succeeded extraction bundle.
// Invalid, stale, incomplete, or unsafe records leave every extraction source unavailable.
@@ -38,71 +31,12 @@ func (c *ArtifactCatalog) HydrateExtractionArtifacts(
if c == nil || m == nil || len(configured) == 0 {
return
}
record := m.Stages[extractStageName]
if record == nil || record.Name != extractStageName || record.Status != manifest.StatusSucceeded {
proof := InspectExtractionEvidence(paths, m, configured)
if proof.State != ExtractionEvidenceValid {
return
}
producerRunID := extractionMetadataString(record.Metadata, extractionMetadataRun)
if !safeExtractionPathSegment(producerRunID) {
return
}
bundleRoot := filepath.Clean(filepath.Join(paths.ArtifactsDir, "notarius", producerRunID))
if !filepath.IsAbs(bundleRoot) || extractionMetadataString(record.Metadata, extractionMetadataRoot) != bundleRoot {
return
}
if !safeExistingExtractionDirectory(paths.Root, bundleRoot) {
return
}
receiptRunID, receiptPipelineID := extractionReceiptIdentity(record.Metadata)
if receiptRunID == "" || receiptPipelineID == "" {
return
}
expected := make(map[string]ExtractionArtifactDefinition, len(configured))
for key, definition := range configured {
sourceID, ok := c.SourceIDForExtractionKey(key)
if !ok {
return
}
expected[sourceID] = definition
}
seen := make(map[string]struct{}, len(expected))
hydrated := make([]hydratedExtraction, 0, len(expected))
indexSeen := false
for _, output := range record.Outputs {
if strings.TrimSpace(output.ProducerRunID) != producerRunID {
return
}
if output.SourceID == "" {
if indexSeen || output.Kind != extractionIndexKind || filepath.Clean(output.LocalPath) != filepath.Join(bundleRoot, "index.json") ||
!validExtractionPayload(bundleRoot, output.LocalPath, output.Checksum) {
return
}
indexSeen = true
continue
}
definition, ok := expected[output.SourceID]
if !ok || output.Kind != extractionLaneKind {
return
}
if _, duplicate := seen[output.SourceID]; duplicate {
return
}
if !compatibleCatalogExtractionContract(output.Contract, definition) ||
!compatibleCatalogExtractionProvenance(output.ExternalProvenance, receiptRunID, receiptPipelineID, definition) ||
!validExtractionPayload(bundleRoot, output.LocalPath, output.Checksum) {
return
}
seen[output.SourceID] = struct{}{}
hydrated = append(hydrated, hydratedExtraction{sourceID: output.SourceID, path: output.LocalPath})
}
if !indexSeen || len(seen) != len(expected) || len(record.Outputs) != len(expected)+1 {
return
}
for _, item := range hydrated {
_ = c.markAvailableFromExtractManifest(item.sourceID, item.path, producerRunID)
for sourceID, path := range proof.Outputs {
_ = c.markAvailableFromExtractManifest(sourceID, path, proof.ProducerRunID)
}
}
@@ -120,25 +54,6 @@ func compatibleCatalogExtractionProvenance(
pipelineID == strings.TrimSpace(want.PipelineID) && got.ArtifactID == want.LaneID
}
func validExtractionPayload(bundleRoot, path, checksum string) bool {
if !filepath.IsAbs(path) || !pathWithinExtractionRoot(bundleRoot, path) || strings.TrimSpace(checksum) == "" {
return false
}
info, err := os.Lstat(path)
if err != nil || !info.Mode().IsRegular() || info.Mode()&os.ModeSymlink != 0 {
return false
}
if !safeExtractionComponents(bundleRoot, path) {
return false
}
actual, err := SHA256File(path)
if err != nil || actual != checksum {
return false
}
body, err := fileops.ReadRegularFile(path, MaxExtractionPayloadBytes)
return err == nil && json.Valid(body)
}
func safeExistingExtractionDirectory(sessionRoot, bundleRoot string) bool {
if !pathWithinExtractionRoot(sessionRoot, bundleRoot) || !safeExtractionComponents(sessionRoot, bundleRoot) {
return false