Update app.py

app.py

@@ -3,6 +3,7 @@ import re
 import json 
 import subprocess 
 import time 
+import shutil
 import img2pdf 
 import gradio as gr 
 from google import genai  # NEW SDK
@@ -11,15 +12,110 @@ from PIL import Image, ImageDraw, ImageFont
 import cv2 
 import numpy as np 
 from PyPDF2 import PdfReader, PdfWriter 
-from prompts import QP_MS_TRANSCRIPTION_PROMPT, get_grading_prompt 
+from prompts import QP_MS_TRANSCRIPTION_PROMPT, get_grading_prompt
+from supabase import create_client, Client 
 
 # ---------------- CONFIG ---------------- 
 # Create client with new SDK
 client = genai.Client(api_key=os.getenv("GEMINI_API_KEY"))
-GRID_ROWS, GRID_COLS = 20, 14 
+GRID_ROWS, GRID_COLS = 20, 14
+
+# Supabase configuration
+SUPABASE_URL = os.getenv("SUPABASE_URL")
+SUPABASE_SERVICE_KEY = os.getenv("SUPABASE_SERVICE_KEY")
+SUPABASE_BUCKET = "examfiles"
+
+# Initialize Supabase client (only if credentials are available)
+supabase_client = None
+if SUPABASE_URL and SUPABASE_SERVICE_KEY:
+    try:
+        supabase_client = create_client(SUPABASE_URL, SUPABASE_SERVICE_KEY)
+        print("✅ Supabase client initialized successfully")
+    except Exception as e:
+        print(f"⚠️ Supabase initialization failed: {e}")
+else:
+    print("⚠️ Supabase credentials not found - file upload to storage disabled") 
 
 # ---------------- PROMPTS ---------------- 
-# Prompts are now imported from prompts.py 
+# Prompts are now imported from prompts.py
+
+# ---------------- SUPABASE HELPERS ---------------- 
+def upload_file_to_supabase(local_path, file_type="unknown"):
+    """
+    Upload a file to Supabase Storage.
+    
+    Args:
+        local_path (str): Local file path
+        file_type (str): Type of file (qp, ms, ans, graded, imprinted)
+    
+    Returns:
+        str: Public URL of uploaded file or None if upload failed
+    """
+    if not supabase_client:
+        print("⚠️ Supabase not configured - skipping upload")
+        return None
+    
+    try:
+        timestamp = str(int(time.time()))
+        original_name = os.path.basename(local_path)
+        remote_path = f"{timestamp}/{file_type}_{original_name}"
+        
+        print(f"📤 Uploading {file_type} to Supabase: {remote_path}")
+        
+        with open(local_path, "rb") as f:
+            supabase_client.storage.from_(SUPABASE_BUCKET).upload(
+                remote_path,
+                f,
+                file_options={"upsert": "true"}
+            )
+        
+        public_url = f"{SUPABASE_URL}/storage/v1/object/public/{SUPABASE_BUCKET}/{remote_path}"
+        print(f"✅ Uploaded successfully: {public_url}")
+        return public_url
+        
+    except Exception as e:
+        print(f"❌ Supabase upload failed for {file_type}: {e}")
+        return None
+
+def process_and_upload_input_files(qp_file_obj, ms_file_obj, ans_file_obj):
+    """
+    Process uploaded files and upload them to Supabase.
+    
+    Args:
+        qp_file_obj: Gradio file object for Question Paper
+        ms_file_obj: Gradio file object for Markscheme
+        ans_file_obj: Gradio file object for Answer Sheet
+    
+    Returns:
+        tuple: (qp_path, ms_path, ans_path, upload_urls_dict)
+    """
+    print("\n" + "="*60)
+    print("📁 PROCESSING INPUT FILES")
+    print("="*60)
+    
+    upload_urls = {
+        "qp_url": None,
+        "ms_url": None,
+        "ans_url": None
+    }
+    
+    # Get local paths from Gradio file objects
+    qp_path = qp_file_obj.name if qp_file_obj else None
+    ms_path = ms_file_obj.name if ms_file_obj else None
+    ans_path = ans_file_obj.name if ans_file_obj else None
+    
+    # Upload to Supabase if configured
+    if supabase_client:
+        if qp_path:
+            upload_urls["qp_url"] = upload_file_to_supabase(qp_path, "qp")
+        if ms_path:
+            upload_urls["ms_url"] = upload_file_to_supabase(ms_path, "ms")
+        if ans_path:
+            upload_urls["ans_url"] = upload_file_to_supabase(ans_path, "ans")
+    
+    print("="*60 + "\n")
+    
+    return qp_path, ms_path, ans_path, upload_urls 
 
 
 
@@ -1050,19 +1146,37 @@ def align_and_grade_pipeline(qp_path, ms_path, ans_path, subject="Maths", imprin
             imprinted_pdf_path = imprint_marks_using_mapping(ans_path, grading_json, imprinted_pdf_path, extracted_ids)
             print("✅ Imprinting finished. Imprinted PDF at:", imprinted_pdf_path)
 
+        # Upload output files to Supabase
+        output_urls = {
+            "graded_pdf_url": None,
+            "imprinted_pdf_url": None
+        }
+        
+        if supabase_client:
+            print("\n📤 Uploading output files to Supabase...")
+            if grading_pdf_path:
+                output_urls["graded_pdf_url"] = upload_file_to_supabase(grading_pdf_path, "graded")
+            if imprinted_pdf_path:
+                output_urls["imprinted_pdf_url"] = upload_file_to_supabase(imprinted_pdf_path, "imprinted")
+
         print("🏁 Pipeline finished successfully.")
-        return qpms_text, as_text, grading_text, grading_pdf_path, imprinted_pdf_path
+        return qpms_text, as_text, grading_text, grading_pdf_path, imprinted_pdf_path, output_urls
 
     except Exception as e:
         print("❌ Pipeline error:", e)
         import traceback
         traceback.print_exc()
-        return f"❌ Error: {e}", None, None, None, None
+        return f"❌ Error: {e}", None, None, None, None, {}
 
 # ---------------- GRADIO UI ---------------- 
 with gr.Blocks(title="AI Grading (Pandoc + pdflatex)") as demo: 
     gr.Markdown("## 📘 AI Grading — Using Pandoc + pdflatex for PDF Generation") 
     gr.Markdown("**✅ Now using Pandoc with pdflatex for professional-quality PDF outputs!**")
+    
+    if supabase_client:
+        gr.Markdown("**☁️ Supabase Storage: Enabled** - All files will be uploaded to cloud storage")
+    else:
+        gr.Markdown("**⚠️ Supabase Storage: Disabled** - Files will only be processed locally")
 
     with gr.Row(): 
         qp_file = gr.File(label="📄 Upload Question Paper (PDF)") 
@@ -1080,6 +1194,11 @@ with gr.Blocks(title="AI Grading (Pandoc + pdflatex)") as demo:
     
     run_button = gr.Button("🚀 Run Pipeline") 
 
+    # File URLs section (only shown if Supabase is enabled)
+    if supabase_client:
+        with gr.Accordion("☁️ Uploaded File URLs", open=False):
+            file_urls_box = gr.Textbox(label="Cloud Storage URLs", lines=8, interactive=False)
+
     with gr.Row(): 
         qpms_box = gr.Textbox(label="📑 QP+MS Transcript", lines=12) 
         as_box = gr.Textbox(label="📝 AS Transcript", lines=12) 
@@ -1090,23 +1209,74 @@ with gr.Blocks(title="AI Grading (Pandoc + pdflatex)") as demo:
 
     def run_pipeline(qp_file_obj, ms_file_obj, ans_file_obj, subject_choice, imprint_flag): 
         if not qp_file_obj or not ms_file_obj or not ans_file_obj:
-            return "❌ Please upload all three files", "", "", None, None
-            
-        qp_path = qp_file_obj.name 
-        ms_path = ms_file_obj.name 
-        ans_path = ans_file_obj.name 
+            error_msg = "❌ Please upload all three files"
+            if supabase_client:
+                return error_msg, "", "", None, None, ""
+            else:
+                return error_msg, "", "", None, None
+        
+        # Process and upload input files
+        qp_path, ms_path, ans_path, input_urls = process_and_upload_input_files(
+            qp_file_obj, ms_file_obj, ans_file_obj
+        )
 
-        qpms_text, as_text, grading_text, grading_pdf_path, imprinted_pdf_path = align_and_grade_pipeline( 
+        # Run the grading pipeline
+        qpms_text, as_text, grading_text, grading_pdf_path, imprinted_pdf_path, output_urls = align_and_grade_pipeline( 
             qp_path, ms_path, ans_path, subject=subject_choice, imprint=imprint_flag 
-        ) 
-
-        return qpms_text or "", as_text or "", grading_text or "", grading_pdf_path, imprinted_pdf_path 
+        )
+        
+        # Build URLs summary
+        urls_summary = ""
+        if supabase_client:
+            urls_summary = "📤 UPLOADED FILES:\n\n"
+            urls_summary += "INPUT FILES:\n"
+            if input_urls.get("qp_url"):
+                urls_summary += f"• Question Paper: {input_urls['qp_url']}\n"
+            if input_urls.get("ms_url"):
+                urls_summary += f"• Markscheme: {input_urls['ms_url']}\n"
+            if input_urls.get("ans_url"):
+                urls_summary += f"• Answer Sheet: {input_urls['ans_url']}\n"
+            
+            urls_summary += "\nOUTPUT FILES:\n"
+            if output_urls.get("graded_pdf_url"):
+                urls_summary += f"• Graded PDF: {output_urls['graded_pdf_url']}\n"
+            if output_urls.get("imprinted_pdf_url"):
+                urls_summary += f"• Imprinted PDF: {output_urls['imprinted_pdf_url']}\n"
+            
+            if not any(input_urls.values()) and not any(output_urls.values()):
+                urls_summary += "\n⚠️ No files were uploaded to Supabase"
+
+        if supabase_client:
+            return (
+                qpms_text or "", 
+                as_text or "", 
+                grading_text or "", 
+                grading_pdf_path, 
+                imprinted_pdf_path,
+                urls_summary
+            )
+        else:
+            return (
+                qpms_text or "", 
+                as_text or "", 
+                grading_text or "", 
+                grading_pdf_path, 
+                imprinted_pdf_path
+            )
 
-    run_button.click( 
-        fn=run_pipeline, 
-        inputs=[qp_file, ms_file, ans_file, subject_dropdown, imprint_toggle], 
-        outputs=[qpms_box, as_box, grading_output_box, grading_pdf_file, imprint_pdf_file] 
-    ) 
+    # Set up the click handler based on whether Supabase is enabled
+    if supabase_client:
+        run_button.click( 
+            fn=run_pipeline, 
+            inputs=[qp_file, ms_file, ans_file, subject_dropdown, imprint_toggle], 
+            outputs=[qpms_box, as_box, grading_output_box, grading_pdf_file, imprint_pdf_file, file_urls_box] 
+        )
+    else:
+        run_button.click( 
+            fn=run_pipeline, 
+            inputs=[qp_file, ms_file, ans_file, subject_dropdown, imprint_toggle], 
+            outputs=[qpms_box, as_box, grading_output_box, grading_pdf_file, imprint_pdf_file] 
+        )
 
 if __name__ == "__main__": 
     demo.launch()