feat(training): add segmentation training pipeline and complete Part 6

- Add /segtrain endpoint to OCR service (ZIP upload, ketos.segtrain,
  backup rotation, in-process model reload)
- Add segtrainModel() to OcrClient and RestClientOcrClient (10-min timeout,
  X-Training-Token header)
- Add SegmentationTrainingExportService: PAGE XML export with polygon
  de-normalization and per-page PNG rendering via PDFBox
- Add GET /api/ocr/segmentation-training-data/export endpoint
- Make TranscriptionBlock.text nullable for segmentation-only blocks
  (V31 migration)
- Add Paraglide i18n translation keys for all training UI strings (de/en/es)
- Pass source prop from TranscriptionEditView to TranscriptionBlock

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

frontend/messages/de.json

@@ -533,5 +533,28 @@
 	"ocr_status_error": "OCR fehlgeschlagen",
 	"transcription_block_review": "Als geprüft markieren",
 	"transcription_block_unreview": "Markierung aufheben",
-	"transcription_reviewed_count": "{reviewed} von {total} geprüft"
+	"transcription_reviewed_count": "{reviewed} von {total} geprüft",
+	"training_ocr_heading": "Kurrent-Erkennung trainieren",
+	"training_ocr_description": "Starte ein neues Training mit den bisher geprüften OCR-Blöcken, um die Erkennungsgenauigkeit für Kurrentschrift zu verbessern.",
+	"training_ocr_blocks_ready": "{blocks} geprüfte Blöcke bereit / {docs} Dokumente",
+	"training_ocr_blocks_total": "(von {total} OCR-Blöcken gesamt)",
+	"training_start_btn": "Training starten",
+	"training_in_progress": "…",
+	"training_success": "Training wurde gestartet und abgeschlossen.",
+	"training_too_few_blocks": "Mindestens 5 geprüfte Blöcke erforderlich (aktuell: {available}).",
+	"training_service_down": "OCR-Dienst ist nicht erreichbar.",
+	"training_history_heading": "Verlauf",
+	"training_history_empty": "Noch keine Trainings-Läufe.",
+	"training_history_col_date": "Datum",
+	"training_history_col_status": "Status",
+	"training_history_col_blocks": "Blöcke",
+	"training_history_col_docs": "Dokumente",
+	"training_status_done": "Fertig",
+	"training_status_failed": "Fehler",
+	"training_status_running": "Läuft…",
+	"training_seg_heading": "Segmentierung trainieren",
+	"training_seg_description": "Starte ein neues Training mit annotierten Segmentierungsbereichen, um die Texterkennung zu verbessern.",
+	"training_seg_blocks_ready": "{blocks} Segmentierungsblöcke bereit",
+	"training_seg_too_few_blocks": "Mindestens 5 Segmentierungsblöcke erforderlich (aktuell: {available}).",
+	"transcription_block_segmentation_only": "Nur Segmentierung"
 }

frontend/messages/en.json

@@ -533,5 +533,28 @@
 	"ocr_status_error": "OCR failed",
 	"transcription_block_review": "Mark as reviewed",
 	"transcription_block_unreview": "Unmark as reviewed",
-	"transcription_reviewed_count": "{reviewed} of {total} reviewed"
+	"transcription_reviewed_count": "{reviewed} of {total} reviewed",
+	"training_ocr_heading": "Train Kurrent recognition",
+	"training_ocr_description": "Start a new training run using the reviewed OCR blocks to improve recognition accuracy for Kurrent script.",
+	"training_ocr_blocks_ready": "{blocks} reviewed blocks ready / {docs} documents",
+	"training_ocr_blocks_total": "(of {total} OCR blocks total)",
+	"training_start_btn": "Start training",
+	"training_in_progress": "…",
+	"training_success": "Training started and completed.",
+	"training_too_few_blocks": "At least 5 reviewed blocks required (currently: {available}).",
+	"training_service_down": "OCR service is unavailable.",
+	"training_history_heading": "History",
+	"training_history_empty": "No training runs yet.",
+	"training_history_col_date": "Date",
+	"training_history_col_status": "Status",
+	"training_history_col_blocks": "Blocks",
+	"training_history_col_docs": "Documents",
+	"training_status_done": "Done",
+	"training_status_failed": "Failed",
+	"training_status_running": "Running…",
+	"training_seg_heading": "Train segmentation",
+	"training_seg_description": "Start a new training run using annotated segmentation regions to improve text detection.",
+	"training_seg_blocks_ready": "{blocks} segmentation blocks ready",
+	"training_seg_too_few_blocks": "At least 5 segmentation blocks required (currently: {available}).",
+	"transcription_block_segmentation_only": "Segmentation only"
 }

frontend/messages/es.json

@@ -533,5 +533,28 @@
 	"ocr_status_error": "OCR fallido",
 	"transcription_block_review": "Marcar como revisado",
 	"transcription_block_unreview": "Desmarcar como revisado",
-	"transcription_reviewed_count": "{reviewed} de {total} revisados"
+	"transcription_reviewed_count": "{reviewed} de {total} revisados",
+	"training_ocr_heading": "Entrenar reconocimiento Kurrent",
+	"training_ocr_description": "Inicia un nuevo entrenamiento con los bloques OCR revisados para mejorar la precisión de reconocimiento del script Kurrent.",
+	"training_ocr_blocks_ready": "{blocks} bloques revisados listos / {docs} documentos",
+	"training_ocr_blocks_total": "(de {total} bloques OCR en total)",
+	"training_start_btn": "Iniciar entrenamiento",
+	"training_in_progress": "…",
+	"training_success": "Entrenamiento iniciado y completado.",
+	"training_too_few_blocks": "Se requieren al menos 5 bloques revisados (actualmente: {available}).",
+	"training_service_down": "El servicio OCR no está disponible.",
+	"training_history_heading": "Historial",
+	"training_history_empty": "Todavía no hay ejecuciones de entrenamiento.",
+	"training_history_col_date": "Fecha",
+	"training_history_col_status": "Estado",
+	"training_history_col_blocks": "Bloques",
+	"training_history_col_docs": "Documentos",
+	"training_status_done": "Listo",
+	"training_status_failed": "Error",
+	"training_status_running": "Ejecutando…",
+	"training_seg_heading": "Entrenar segmentación",
+	"training_seg_description": "Inicia un nuevo entrenamiento con regiones de segmentación anotadas para mejorar la detección de texto.",
+	"training_seg_blocks_ready": "{blocks} bloques de segmentación listos",
+	"training_seg_too_few_blocks": "Se requieren al menos 5 bloques de segmentación (actualmente: {available}).",
+	"transcription_block_segmentation_only": "Solo segmentación"
 }

frontend/src/lib/components/TranscriptionEditView.svelte

@@ -344,6 +344,7 @@ $effect(() => {
 							onMoveDown={() => handleMoveDown(block.id)}
 							isFirst={i === 0}
 							isLast={i === sortedBlocks.length - 1}
+							source={block.source}
 						/>
 					</div>
 				{/each}