Update helpers.py

helpers.py

@@ -147,9 +147,11 @@ def get_web_documents(base_urls=['https://nct.neu.edu.vn/']):
 
 def load_text_data(file_path):
     """Load text content from a DOCX file (tables removed)."""
+    text_splitter = RecursiveCharacterTextSplitter(chunk_size=512, chunk_overlap=100)
     loader = DoclingLoader(
         file_path=file_path,
-        export_type=ExportType.DOC_CHUNKS  # Enable internal chunking
+        export_type=ExportType.MARKDOWN,  # Enable internal chunking,
+        chunker = text_splitter
     )
     return loader.load()