Update app.py

app.py

@@ -1,11 +1,16 @@
 import gradio as gr
 from docling.document_converter import DocumentConverter
+from docling.datamodel.pipeline_options import PdfPipelineOptions, TesseractCliOcrOptions
 import spaces
 
 @spaces.GPU
 def convert_document(file, output_format):
+    pdf_opts = PdfPipelineOptions(
+        do_ocr=True,
+        ocr_options=TesseractCliOcrOptions(lang=["eng"]),  # or ["eng", "ara"]
+    )
     # Load document and convert it using Docling
-    converter = DocumentConverter()
+    converter = DocumentConverter(pipeline_options=pdf_opts)
     result = converter.convert(file.name)
 
     # Check available attributes in DoclingDocument