add application file

app.py

@@ -1,4 +1,5 @@
 import gradio as gr
+from gradio_pdf import PDF
 
 import fitz
 import os
@@ -1010,15 +1011,7 @@ def find_reference_pages(pdf_path):
     doc.close()
     return ref_pages, start_page, end_page, ref_text
 
-def encode_pdf_to_base64(path):
-    """Encode PDF file to base64 string for embedding."""
-    try:
-        with open(path, "rb") as f:
-            encoded_string = base64.b64encode(f.read()).decode('utf-8')
-        return encoded_string
-    except Exception as e:
-        print(f"Error encoding PDF: {e}")
-        return None
+
 
 def process_pdf_initial(pdf_file, state_pdf_path, state_ref_pages, state_citations, state_removed_citations, state_appendix_header, state_ref_text):
     """Initial PDF processing - find references and show PDF immediately."""
@@ -1068,18 +1061,10 @@ def process_pdf_initial(pdf_file, state_pdf_path, state_ref_pages, state_citatio
         
     status += "\n⏳ Starting automatic extraction... Please wait."
     
-    basename = os.path.basename(new_pdf_path)
-    
-    # Generate Base64 for the full PDF view
-    # This bypasses browser security restrictions on local files
-    b64_pdf = encode_pdf_to_base64(new_pdf_path)
-    if b64_pdf:
-        iframe_html = f'<iframe src="data:application/pdf;base64,{b64_pdf}" width="100%" height="700px" style="border: none;"></iframe>'
-    else:
-        iframe_html = '<div style="padding:20px">Error loading PDF viewer.</div>'
     
+    # Return directly to the PDF component
     return (new_pdf_path, status, 
-            gr.update(value=iframe_html, visible=True), 
+            gr.update(value=new_pdf_path, visible=True), 
             gr.update(visible=True, value="Show Full PDF"),
             gr.update(visible=False),  # Citations display
             gr.update(interactive=False, visible=False), # Verify Button
@@ -1941,13 +1926,7 @@ def update_view(view_mode, state_pdf_path, state_ref_pages, state_citations, sta
         
         # Check cache first
         if state_ref_pdf_path and os.path.exists(state_ref_pdf_path):
-            # Return iframe HTML with Base64
-            b64_pdf = encode_pdf_to_base64(state_ref_pdf_path)
-            if b64_pdf:
-                iframe_html = f'<iframe src="data:application/pdf;base64,{b64_pdf}" width="100%" height="700px" style="border: none;"></iframe>'
-            else:
-                iframe_html = "Error loading PDF."
-            upd_ref_pdf = gr.update(value=iframe_html)
+            upd_ref_pdf = gr.update(value=state_ref_pdf_path)
         else:
             # Generate the Subset PDF if needed.
             if state_ref_pages and state_pdf_path:
@@ -1962,13 +1941,7 @@ def update_view(view_mode, state_pdf_path, state_ref_pages, state_citations, sta
                 doc.close()
                 
                 state_ref_pdf_path = output_path
-                # Return iframe HTML with Base64
-                b64_pdf = encode_pdf_to_base64(output_path)
-                if b64_pdf:
-                    iframe_html = f'<iframe src="data:application/pdf;base64,{b64_pdf}" width="100%" height="700px" style="border: none;"></iframe>'
-                else:
-                    iframe_html = "Error loading PDF."
-                upd_ref_pdf = gr.update(value=iframe_html)
+                upd_ref_pdf = gr.update(value=output_path)
              
         yield (vis_full, vis_ref, vis_cit, vis_ver, upd_ref_pdf, upd_cit_disp, upd_ver_disp, upd_load, state_ref_pdf_path)
 
@@ -2207,13 +2180,13 @@ with gr.Blocks(title="CiteAudit", css="""
             
             # 1. Full PDF View
             with gr.Group(visible=True) as view_full_pdf:
-                # Use standard HTML iframe for better browser compatibility (specifically Safari)
-                pdf_viewer_full = gr.HTML(label="Full PDF", elem_id="pdf-viewer-full")
+                # Use standard PDF component
+                pdf_viewer_full = PDF(label="Full PDF", height=700, elem_id="pdf-viewer-full", interactive=False)
             
             # 2. Reference Pages View
             with gr.Group(visible=False) as view_ref_pages:
-                # Use standard HTML iframe
-                pdf_viewer_ref = gr.HTML(label="Reference Pages", elem_id="pdf-viewer-ref")
+                # Use standard PDF component
+                pdf_viewer_ref = PDF(label="Reference Pages", height=700, elem_id="pdf-viewer-ref", interactive=False)
             
             # 3. Citations View
             with gr.Group(visible=False, elem_id="view-citations") as view_citations:

requirements.txt

@@ -1,6 +1,7 @@
 PyMuPDF
 gradio==4.31.0
 
+gradio_pdf
 pydantic==2.10.6
 grobid-client-python
 huggingface_hub<1.0.0