feat: wire persistence into the web layer (check → archive → dossier)

This is the piece that turns a Pre-Publish-Prüfung into an actual
archived, provable record instead of a one-off form response.

extract.Client.Extract now returns Result{Facts, RawJSON} instead of
just Facts — RawJSON is the model's exact, unmodified JSON, which is
what belongs in extraction.payload (the audit trail), not a re-encoded
view through our own Facts struct. extract.ParsePayload reconstructs
Facts from a stored payload later, reusing the same parsing/validation
path Extract uses (including the enum guard), so a previously-saved
extraction can be read back exactly as it would have been the first
time.

internal/dossier.BuildContent no longer requires AssetHash: most checks
right now are caption-only (no image/video upload wired yet), and
inventing a placeholder hash for a nonexistent asset would itself be an
integrity problem in an evidence tool. Content shows "kein Asset
hinterlegt" instead.

internal/web gains a narrow Store interface (mirroring the Extractor
pattern — only the methods these handlers use, not the full
*store.Store) so its test suite stays network/DB-free via an in-memory
fake:

- POST /pruefen persists submission + extraction + findings and marks
  the submission "checked". A needsClarification result persists the
  extraction (there's something worth keeping) but no findings and no
  status change, and the template omits the archive option entirely.
- POST /veroeffentlichen re-derives Facts from the stored payload, hashes
  the canonical submission+facts+findings metadata, gets an RFC-3161
  timestamp, generates the PDF dossier to disk, and persists the
  evidence_package row before marking the submission "published".
- GET /dossier/{id} serves the generated PDF.

Tested end-to-end offline: a fake Timestamper builds a real, structurally
valid self-signed RFC-3161 response so the full check→archive→download
flow runs against an in-memory store, verifying the downloaded bytes are
an actual PDF and the dossier file lands on disk — without hitting a
real database, TSA, or the Claude API.

cmd/deklarix/main.go now wires store.Store, evidence.NewHTTPTimestamper
(TSA_URL, default FreeTSA), and DOSSIER_DIR (default "dossiers") into
web.NewServer alongside the extractor and rule set.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
This commit is contained in:
noroot
2026-08-27 14:39:38 +02:00
parent 72005fc326
commit f5ad08cebd
14 changed files with 724 additions and 105 deletions

View File

@@ -14,8 +14,10 @@ import (
"html/template"
"net/http"
"github.com/netcell-it/deklarix/internal/evidence"
"github.com/netcell-it/deklarix/internal/extract"
"github.com/netcell-it/deklarix/internal/rules"
"github.com/netcell-it/deklarix/internal/store"
)
//go:embed templates/*.html
@@ -28,31 +30,61 @@ var staticFS embed.FS
// *extract.Client erfüllt sie; Tests injizieren einen Fake statt echte
// Claude-API-Aufrufe zu machen.
type Extractor interface {
Extract(ctx context.Context, in extract.Input) (rules.Facts, error)
Extract(ctx context.Context, in extract.Input) (extract.Result, error)
ModelVersion() string
}
// Store ist die Schnittstelle, die der Server zur Persistenz braucht —
// bewusst schmal (nur was diese Handler tatsächlich nutzen), nicht der
// komplette *store.Store. *store.Store erfüllt sie; Tests injizieren
// einen Fake statt eine echte Datenbank zu brauchen.
type Store interface {
CreateSubmission(ctx context.Context, platform, postType, caption string) (store.Submission, error)
GetSubmission(ctx context.Context, id string) (store.Submission, error)
SetSubmissionStatus(ctx context.Context, id, status string) error
CreateExtraction(ctx context.Context, submissionID string, payload []byte, modelVersion, promptVersion string) (store.Extraction, error)
GetLatestExtraction(ctx context.Context, submissionID string) (store.Extraction, error)
CreateFinding(ctx context.Context, submissionID string, extractionID *string, ruleID string, ruleVersion int, severity, title, fix string, sources []string) (store.Finding, error)
ListCurrentFindings(ctx context.Context, submissionID string) ([]store.Finding, error)
CreateEvidencePackage(ctx context.Context, submissionID, dossierPath, sha256Hex string, timestampToken []byte) (store.EvidencePackage, error)
GetLatestEvidencePackage(ctx context.Context, submissionID string) (store.EvidencePackage, error)
}
// Server bündelt Routing und Abhängigkeiten der Web-Schicht.
type Server struct {
mux *http.ServeMux
extractor Extractor
ruleSet []rules.Rule
templates *template.Template
mux *http.ServeMux
extractor Extractor
ruleSet []rules.Rule
store Store
timestamper evidence.Timestamper
dossierDir string
templates *template.Template
}
// NewServer erstellt den Server. ruleSet kommt von rules.Load und wird
// einmal beim Start geladen, nicht pro Request.
func NewServer(extractor Extractor, ruleSet []rules.Rule) (*Server, error) {
// einmal beim Start geladen, nicht pro Request. dossierDir ist das
// Verzeichnis, in das erzeugte Nachweis-Dossiers geschrieben werden.
func NewServer(extractor Extractor, ruleSet []rules.Rule, st Store, timestamper evidence.Timestamper, dossierDir string) (*Server, error) {
tmpl, err := template.ParseFS(templatesFS, "templates/*.html")
if err != nil {
return nil, fmt.Errorf("web: templates parsen: %w", err)
}
s := &Server{extractor: extractor, ruleSet: ruleSet, templates: tmpl}
s := &Server{
extractor: extractor,
ruleSet: ruleSet,
store: st,
timestamper: timestamper,
dossierDir: dossierDir,
templates: tmpl,
}
mux := http.NewServeMux()
mux.HandleFunc("GET /health", s.handleHealth)
mux.HandleFunc("GET /{$}", s.handleIndex)
mux.HandleFunc("POST /pruefen", s.handleCheck)
mux.HandleFunc("POST /veroeffentlichen", s.handleArchive)
mux.HandleFunc("GET /dossier/{id}", s.handleDossierDownload)
mux.Handle("GET /static/", http.FileServerFS(staticFS))
s.mux = mux