optimize with new transformer rmodel

python/vector/vector_ingest.py

@@ -77,13 +77,14 @@ with open(index_path, "r", encoding="utf-8") as f:
         if not text or not chunk_id:
             continue
 
-        texts.append(text)
+        # -------------------------------------------------
+        # E5 requires "passage:" prefix for document chunks
+        # -------------------------------------------------
+        texts.append(f"passage: {text}")
         ids.append(chunk_id)
 
 if not texts:
     print("No chunks found. Removing vector index.")
-
-    # Entferne final erst später in PHP atomar
     sys.exit(0)
 
 print(f"Loaded {len(texts)} chunks.")