Add deterministic analysis input identities

This commit is contained in:
2026-08-29 19:02:16 +00:00
parent 62de6abdbf
commit ab5751459a
12 changed files with 815 additions and 150 deletions

View File

@@ -6,6 +6,7 @@ import (
"path/filepath"
"strings"
"gitea.maximumdirect.net/eric/narratio/internal/artifactmodel"
"gitea.maximumdirect.net/eric/narratio/internal/fileops"
"gitea.maximumdirect.net/eric/narratio/internal/manifest"
)
@@ -23,7 +24,15 @@ const (
type ExtractionEvidence struct {
State ExtractionEvidenceState
Reason, ProducerRunID string
Outputs map[string]string
Outputs map[string]ExtractionEvidenceOutput
}
// ExtractionEvidenceOutput is the verified semantic identity of one lane.
type ExtractionEvidenceOutput struct {
Path string
Checksum string
Size int64
Contract *artifactmodel.ContractMetadata
}
// InspectExtractionEvidence verifies structure, confinement, identities, contracts, and payload bytes.
@@ -72,7 +81,7 @@ func InspectExtractionEvidence(
for key, d := range configured {
expected[ExtractionArtifactSourceID(key)] = d
}
seen, outputs := map[string]struct{}{}, map[string]string{}
seen, outputs := map[string]struct{}{}, map[string]ExtractionEvidenceOutput{}
indexSeen := false
for _, out := range r.Outputs {
if strings.TrimSpace(out.ProducerRunID) != runID {
@@ -82,7 +91,7 @@ func InspectExtractionEvidence(
if indexSeen || out.Kind != extractionIndexKind || filepath.Clean(out.LocalPath) != filepath.Join(root, "index.json") {
return ExtractionEvidence{State: ExtractionEvidenceObsolete, Reason: "extract index path is not canonical"}
}
if state, reason := inspectExtractionPayload(root, out.LocalPath, out.Checksum); state != ExtractionEvidenceValid {
if state, reason, _ := inspectExtractionPayload(root, out.LocalPath, out.Checksum); state != ExtractionEvidenceValid {
return ExtractionEvidence{State: state, Reason: reason}
}
indexSeen = true
@@ -98,11 +107,15 @@ func InspectExtractionEvidence(
if !compatibleCatalogExtractionContract(out.Contract, d) || !compatibleCatalogExtractionProvenance(out.ExternalProvenance, receiptRunID, receiptPipelineID, d) {
return ExtractionEvidence{State: ExtractionEvidenceObsolete, Reason: "extract output contract or provenance is incompatible"}
}
if state, reason := inspectExtractionPayload(root, out.LocalPath, out.Checksum); state != ExtractionEvidenceValid {
state, reason, size := inspectExtractionPayload(root, out.LocalPath, out.Checksum)
if state != ExtractionEvidenceValid {
return ExtractionEvidence{State: state, Reason: reason}
}
seen[out.SourceID] = struct{}{}
outputs[out.SourceID] = out.LocalPath
outputs[out.SourceID] = ExtractionEvidenceOutput{
Path: out.LocalPath, Checksum: out.Checksum, Size: size,
Contract: cloneArtifactContract(out.Contract),
}
}
if !indexSeen || len(seen) != len(expected) || len(r.Outputs) != len(expected)+1 {
return ExtractionEvidence{State: ExtractionEvidenceObsolete, Reason: "extract result is incomplete"}
@@ -110,30 +123,30 @@ func InspectExtractionEvidence(
return ExtractionEvidence{State: ExtractionEvidenceValid, ProducerRunID: runID, Outputs: outputs}
}
func inspectExtractionPayload(root, path, checksum string) (ExtractionEvidenceState, string) {
func inspectExtractionPayload(root, path, checksum string) (ExtractionEvidenceState, string, int64) {
if !filepath.IsAbs(path) || !pathWithinExtractionRoot(root, path) || strings.TrimSpace(checksum) == "" {
return ExtractionEvidenceUnsafe, "extract output path or checksum is unsafe"
return ExtractionEvidenceUnsafe, "extract output path or checksum is unsafe", 0
}
info, err := os.Lstat(path)
if os.IsNotExist(err) {
return ExtractionEvidenceObsolete, "extract output is missing"
return ExtractionEvidenceObsolete, "extract output is missing", 0
}
if !safeExtractionComponents(root, path) {
return ExtractionEvidenceUnsafe, "extract output path contains unsafe components"
return ExtractionEvidenceUnsafe, "extract output path contains unsafe components", 0
}
if err != nil || !info.Mode().IsRegular() || info.Mode()&os.ModeSymlink != 0 {
return ExtractionEvidenceUnsafe, "extract output is not a regular file"
return ExtractionEvidenceUnsafe, "extract output is not a regular file", 0
}
actual, err := SHA256File(path)
if err != nil {
return ExtractionEvidenceUnsafe, "extract output checksum cannot be read"
return ExtractionEvidenceUnsafe, "extract output checksum cannot be read", 0
}
if actual != checksum {
return ExtractionEvidenceObsolete, "extract output checksum does not match durable bytes"
return ExtractionEvidenceObsolete, "extract output checksum does not match durable bytes", 0
}
body, err := fileops.ReadRegularFile(path, MaxExtractionPayloadBytes)
if err != nil || !json.Valid(body) {
return ExtractionEvidenceObsolete, "extract output is not valid JSON"
return ExtractionEvidenceObsolete, "extract output is not valid JSON", 0
}
return ExtractionEvidenceValid, ""
return ExtractionEvidenceValid, "", info.Size()
}