FDN-01: repository & projektgerüst

Git-Repository für bestehenden archivdms-Code initialisiert, Branch-/Commit-Konvention (feature/<ticket>-<slug>-Branches, Ticket-Prefix in Commit-Nachricht) etabliert.
2026-08-11 21:27:53 +02:00
parent 40ed80da71
commit 9a24ea29e1
274 changed files with 53708 additions and 0 deletions
@@ -0,0 +1,75 @@
+package api
+
+import (
+	"net/http"
+	"strconv"
+)
+
+// ocrWordResponse is the wire shape of one OCR word box. It is a dedicated DTO
+// (not storage.OCRWord, which carries no json tags and exposes the internal row
+// id / document_id) so the overlay renderer gets a compact, stable payload.
+// Coordinates are in the original file's coordinate space — see
+// internal/ocr/coords.go.
+type ocrWordResponse struct {
+	Text       string  `json:"text"`
+	Left       int     `json:"left"`
+	Top        int     `json:"top"`
+	Width      int     `json:"width"`
+	Height     int     `json:"height"`
+	Confidence float64 `json:"confidence"`
+	Page       int     `json:"page"`
+	Block      int     `json:"block"`
+	Par        int     `json:"par"`
+	Line       int     `json:"line"`
+}
+
+// handleListDocumentOCRWords returns the stored word-level bounding boxes of a
+// document (GET /api/documents/{id}/ocr-words), Phase 3 of the OCR
+// text-highlight/overlay feature.
+//
+// Tenant/ACL: ocr_words has no tenant_id column, access is only ever mediated
+// through document_id. The handler therefore performs the exact same ownership
+// check as handleDocumentAuditLog / handleGetDocumentFile — GetDocument(id,
+// tenantID) filters WHERE tenant_id and yields 404 for both "unknown id" and
+// "foreign tenant", so the endpoint never reveals whether a document exists
+// outside the caller's tenant — BEFORE any ocr_words row is read.
+//
+// Pure read: no audit entry, consistent with the other document GET handlers.
+// Empty result serializes as [] (never null).
+func (s *Server) handleListDocumentOCRWords(w http.ResponseWriter, r *http.Request) {
+	sess := sessionFromCtx(r.Context())
+	id, err := strconv.ParseInt(r.PathValue("id"), 10, 64)
+	if err != nil || sess.TenantID == nil {
+		writeError(w, http.StatusBadRequest, "invalid document id")
+		return
+	}
+	// ACL/tenant check first — must precede the ocr_words lookup.
+	if _, err := s.store.GetDocument(r.Context(), id, *sess.TenantID); err != nil {
+		writeError(w, http.StatusNotFound, "document not found")
+		return
+	}
+
+	words, err := s.store.ListOCRWords(r.Context(), id)
+	if err != nil {
+		s.logger.Error("list ocr words failed", "document_id", id, "tenant_id", *sess.TenantID, "err", err)
+		writeError(w, http.StatusInternalServerError, "list ocr words failed")
+		return
+	}
+
+	out := make([]ocrWordResponse, 0, len(words))
+	for _, wd := range words {
+		out = append(out, ocrWordResponse{
+			Text:       wd.Word,
+			Left:       wd.Left,
+			Top:        wd.Top,
+			Width:      wd.Width,
+			Height:     wd.Height,
+			Confidence: wd.Confidence,
+			Page:       wd.Page,
+			Block:      wd.Block,
+			Par:        wd.Par,
+			Line:       wd.Line,
+		})
+	}
+	writeJSON(w, http.StatusOK, out)
+}