From 9b0e31097893ea489fc9485634ea488b1bf6452a Mon Sep 17 00:00:00 2001
From: Benjamin Admin <benjaminadmin@MacBook-Pro.local>
Date: Fri, 17 Apr 2026 00:46:15 +0200
Subject: [PATCH] Fix: reprocess button works after session resume + apply
 merge logic

Two bugs fixed:
1. reprocessPages() failed silently after session resume because
   successfulPages was empty. Now derives pages from vocabulary
   source_page or selectedPages as fallback.

2. process-single-page endpoint built vocabulary entries WITHOUT
   applying merge logic (_merge_wrapped_rows, _merge_continuation_rows).
   Now applies full merge pipeline after vocabulary extraction.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
---
 .../backend/vocab_worksheet_api.py            | 41 +++++++++++++++++++
 .../app/vocab-worksheet/useVocabWorksheet.ts  | 21 +++++++++-
 2 files changed, 60 insertions(+), 2 deletions(-)

diff --git a/klausur-service/backend/vocab_worksheet_api.py b/klausur-service/backend/vocab_worksheet_api.py
index 43d2a45..91e7307 100644
--- a/klausur-service/backend/vocab_worksheet_api.py
+++ b/klausur-service/backend/vocab_worksheet_api.py
@@ -77,6 +77,11 @@ try:
         render_pdf_high_res,
         PageRegion, RowGeometry,
     )
+    from cv_cell_grid import (
+        _merge_wrapped_rows,
+        _merge_phonetic_continuation_rows,
+        _merge_continuation_rows,
+    )
     from ocr_pipeline_session_store import (
         create_session_db as create_pipeline_session_db,
         update_session_db as update_pipeline_session_db,
@@ -1696,6 +1701,42 @@ async def _run_ocr_pipeline_for_page(
                 })
             extraction_source = f"generic ({len(all_ci)} cols)"
 
+    # --- Post-processing: merge cell-wrap continuation rows ---
+    if len(page_vocabulary) >= 2:
+        try:
+            # Convert to internal format (example_sentence → example)
+            internal = []
+            for v in page_vocabulary:
+                internal.append({
+                    'row_index': len(internal),
+                    'english': v.get('english', ''),
+                    'german': v.get('german', ''),
+                    'example': v.get('example_sentence', ''),
+                })
+
+            n_before = len(internal)
+            internal = _merge_wrapped_rows(internal)
+            internal = _merge_phonetic_continuation_rows(internal)
+            internal = _merge_continuation_rows(internal)
+
+            if len(internal) < n_before:
+                # Rebuild page_vocabulary from merged entries
+                merged_vocab = []
+                for entry in internal:
+                    if not entry.get('english') and not entry.get('german'):
+                        continue
+                    merged_vocab.append({
+                        'id': str(uuid.uuid4()),
+                        'english': entry.get('english', ''),
+                        'german': entry.get('german', ''),
+                        'example_sentence': entry.get('example', ''),
+                        'source_page': page_number + 1,
+                    })
+                logger.info(f"  row merging: {n_before} → {len(merged_vocab)} entries")
+                page_vocabulary = merged_vocab
+        except Exception as e:
+            logger.warning(f"  row merging failed (non-critical): {e}")
+
     logger.info(f"  vocab extraction: {len(page_vocabulary)} entries via {extraction_source}")
 
     total_duration = _time.time() - t_total
diff --git a/studio-v2/app/vocab-worksheet/useVocabWorksheet.ts b/studio-v2/app/vocab-worksheet/useVocabWorksheet.ts
index da95877..72a2ee3 100644
--- a/studio-v2/app/vocab-worksheet/useVocabWorksheet.ts
+++ b/studio-v2/app/vocab-worksheet/useVocabWorksheet.ts
@@ -760,11 +760,28 @@ export function useVocabWorksheet(): VocabWorksheetHook {
 
   // Reprocess all successful pages with new IPA/syllable modes
   const reprocessPages = (ipa: IpaMode, syllable: SyllableMode) => {
-    if (!session || successfulPages.length === 0) return
+    if (!session) return
+
+    // Determine pages to reprocess: use successfulPages if available,
+    // otherwise derive from vocabulary source_page or selectedPages
+    let pagesToReprocess: number[]
+    if (successfulPages.length > 0) {
+      pagesToReprocess = successfulPages.map(p => p - 1)
+    } else if (vocabulary.length > 0) {
+      // Derive from vocabulary entries' source_page (1-indexed → 0-indexed)
+      const pageSet = new Set(vocabulary.map(v => (v.source_page || 1) - 1))
+      pagesToReprocess = [...pageSet].sort((a, b) => a - b)
+    } else if (selectedPages.length > 0) {
+      pagesToReprocess = [...selectedPages]
+    } else {
+      // Fallback: try page 0
+      pagesToReprocess = [0]
+    }
+
+    if (pagesToReprocess.length === 0) return
 
     setIsExtracting(true)
     setExtractionStatus('Verarbeite mit neuen Einstellungen...')
-    const pagesToReprocess = successfulPages.map(p => p - 1)
     const API_BASE = getApiBase()
 
     ;(async () => {