Update app.py

app.py

@@ -4,10 +4,10 @@ import json
 import os
 import tempfile
 
-# 1) Initialize DocumentConverter
+# Initialize DocumentConverter
 converter = DocumentConverter()
 
-# 2) Conversion function
+# Conversion function
 def convert(file, out_format):
     if file is None:
         return "Please upload a file first."
@@ -18,52 +18,39 @@ def convert(file, out_format):
 
         # Export based on the selected format
         if out_format == "Markdown":
-            # Save as Markdown file
-            with tempfile.NamedTemporaryFile(delete=False, suffix=".md") as temp_file:
-                temp_filename = temp_file.name
-                with open(temp_filename, 'w') as f:
-                    f.write(doc.document.export_to_markdown())
-                return temp_filename  # Return file for download
-
+            # Show Markdown content directly in the output box
+            return doc.document.export_to_markdown()
         elif out_format == "HTML":
-            # Save as HTML file
-            with tempfile.NamedTemporaryFile(delete=False, suffix=".html") as temp_file:
-                temp_filename = temp_file.name
-                with open(temp_filename, 'w') as f:
-                    f.write(doc.document.export_to_html())
-                return temp_filename  # Return file for download
-
+            # Show HTML content directly in the output box
+            return doc.document.export_to_html()
         elif out_format == "JSON":
-            # Save as JSON file (lossless serialization)
-            doctags = doc.document.save_as_doctags()
-            with tempfile.NamedTemporaryFile(delete=False, suffix=".json") as temp_file:
-                temp_filename = temp_file.name
-                with open(temp_filename, 'w') as f:
-                    json.dump(doctags, f, indent=4)
-                return temp_filename  # Return file for download
-
+            # Export to JSON format
+            doctags = doc.document.export_to_dict()
+            return json.dumps(doctags, indent=4)
         elif out_format == "Text":
-            # Save as Text file (plain text)
-            with tempfile.NamedTemporaryFile(delete=False, suffix=".txt") as temp_file:
-                temp_filename = temp_file.name
-                with open(temp_filename, 'w') as f:
-                    f.write(doc.document.text)
-                return temp_filename  # Return file for download
-
+            # Extract plain text
+            return doc.document.export_to_markdown()  # Updated to the correct method
         elif out_format == "Doctags":
-            # Save as Doctags file
+            # Create a temporary file to save Doctags and allow download
             with tempfile.NamedTemporaryFile(delete=False, suffix=".doctags") as temp_file:
                 temp_filename = temp_file.name
+
+                # Save the document as Doctags (JSON-like format)
                 doc.document.save_as_doctags(temp_filename)
-                return temp_filename  # Return file for download
 
+                # Ensure the file is not empty
+                if os.path.getsize(temp_filename) > 0:
+                    # Return the file path for download
+                    return temp_filename
+                else:
+                    return "Error: The Doctags file is empty."
         else:
             return "Unsupported output format"
 
     except Exception as e:
         return f"Error converting document: {str(e)}"
 
-# 3) Build Gradio interface
+# Build Gradio interface
 with gr.Blocks() as demo:
     # Top: explanatory text about Docling and a link to the official repository
     gr.Markdown(
@@ -102,12 +89,12 @@ with gr.Blocks() as demo:
         )
         btn = gr.Button("Convert")
 
-    # Output panel (file output for all formats)
+    # Output panel (file output for Doctags)
     out = gr.File(label="📄 Download Output File")
 
     # Wire it up
     btn.click(fn=convert, inputs=[inp, fmt], outputs=out)
 
-# 4) Launch
+# Launch
 if __name__ == "__main__":
     demo.launch()