Update app.py

app.py

@@ -226,29 +226,31 @@ def identify_headers_with_openrouter(pdf_path, model,LLM_prompt, pages_to_check=
         conf = float(obj.get('confidence') or 0)
         if t and page is not None:
             out.append({'text': t, 'page': page-1, 'suggested_level': level, 'confidence': conf})
-    df = pd.DataFrame(out)
-    
-    # Save to Excel
-    df.to_excel("output.xlsx", index=False)
-    
-    print("Saved JSON output to output.xlsx")
     return out
 
-# Wrapper function to convert JSON to a dataframe-friendly format
+import gradio as gr
+import pandas as pd
+from io import BytesIO
+
 def identify_headers_with_table(pdf_path, model, LLM_prompt):
     # Call your existing function
     result = identify_headers_with_openrouter(pdf_path, model, LLM_prompt)
     
-    # Convert list of dicts to list of lists for Gradio Dataframe
     if not result:
-        return []  # empty table if no results
+        return None  # nothing to return
     
-    table_data = [[item['text'], item['page']+1, item['suggested_level'], item['confidence']] for item in result]
-    return table_data
-
-# Column names for the table
-columns = ["Text", "Page", "Suggested Level", "Confidence"]
-
+    # Convert to DataFrame
+    df = pd.DataFrame(result)
+    
+    # Optional: adjust page numbers for human-readable format
+    df['page'] = df['page'] + 1
+    
+    # Save to in-memory Excel file
+    output = BytesIO()
+    df.to_excel(output, index=False)
+    output.seek(0)
+    
+    return output  # return file-like object for download
 
 # Gradio Interface
 iface = gr.Interface(
@@ -258,7 +260,7 @@ iface = gr.Interface(
         gr.Textbox(label="Model Type"),
         gr.Textbox(label="LLM Prompt")
     ],
-    outputs=gr.Dataframe(headers=columns)
+    outputs=gr.File(label="Download Excel")
 )
 
 iface.launch()