Update app.py

app.py

@@ -140,8 +140,16 @@ def simple_process_pdfs(pdf_paths):
     """Process PDF documents and return document objects"""
     documents = []
     
+    print(f"Attempting to process PDFs: {pdf_paths}")
+    print(f"Current directory contents: {os.listdir('.')}")
+    
     for pdf_path in pdf_paths:
         try:
+            if not os.path.exists(pdf_path):
+                print(f"Warning: {pdf_path} does not exist")
+                continue
+                
+            print(f"Processing {pdf_path}...")
             text = ""
             with open(pdf_path, 'rb') as file:
                 reader = PyPDF2.PdfReader(file)
@@ -161,6 +169,8 @@ def simple_process_pdfs(pdf_paths):
                 print(f"Warning: No text extracted from {pdf_path}")
         except Exception as e:
             print(f"Error processing {pdf_path}: {e}")
+            import traceback
+            traceback.print_exc()
     
     print(f"Processed {len(documents)} PDF documents")
     return documents
@@ -414,26 +424,21 @@ comprehensive_evaluation_data = [
 # Gradio Interface
 def initialize_system():
     """Initialize the Vision 2030 Assistant system"""
-    # This would normally process PDFs and load models
-    # For Hugging Face Space, we'll need to check if models are already downloaded
-    # and if vector stores are already created
-    
-    # Define paths
-    model_dir = "models"
-    vector_store_dir = "vector_stores"
-    pdf_dir = "pdf_data"
+    # Define paths for PDF files in the root directory
+    pdf_files = ["saudi_vision203.pdf", "saudi_vision2030_ar.pdf"]
     
-    os.makedirs(model_dir, exist_ok=True)
-    os.makedirs(vector_store_dir, exist_ok=True)
-    os.makedirs(pdf_dir, exist_ok=True)
-    
-    # Check if we need to download PDFs
-    pdf_files = ["vision2030_docs/saudi_vision203.pdf", "vision2030_docs/saudi_vision2030_ar.pdf"]
+    # Print available files for debugging
+    print("Files in current directory:", os.listdir("."))
     
-    # This is where you would normally download the PDFs if they don't exist
-    # For Hugging Face Space, you would need to upload these files
+    # Check if PDFs exist
+    for pdf_file in pdf_files:
+        if not os.path.exists(pdf_file):
+            print(f"Warning: {pdf_file} not found")
     
     # Process PDFs and create vector store
+    vector_store_dir = "vector_stores"
+    os.makedirs(vector_store_dir, exist_ok=True)
+    
     if os.path.exists(os.path.join(vector_store_dir, "index.faiss")):
         print("Loading existing vector store...")
         embedding_function = HuggingFaceEmbeddings(
@@ -443,6 +448,8 @@ def initialize_system():
     else:
         print("Creating new vector store...")
         documents = simple_process_pdfs(pdf_files)
+        if not documents:
+            raise ValueError("No documents were processed successfully. Cannot continue.")
         vector_store = create_vector_store(documents)
         vector_store.save_local(vector_store_dir)
     
@@ -505,14 +512,19 @@ def run_evaluation_on_sample(assistant, sample_index=0):
     # Evaluate response
     evaluation_results = evaluate_response(query, response, reference)
     
-    # Format for display
-    metrics_str = "\n".join([f"{k}: {v}" for k, v in evaluation_results.items()])
-    
     return query, response, reference, evaluation_results, sources, category, language
 
-def qualitative_evaluation_interface(assistant):
+def qualitative_evaluation_interface(assistant=None):
     """Create a Gradio interface for qualitative evaluation"""
     
+    # If assistant is None, create a simplified interface
+    if assistant is None:
+        with gr.Blocks(title="Vision 2030 Assistant - Initialization Error") as interface:
+            gr.Markdown("# Vision 2030 Assistant - Initialization Error")
+            gr.Markdown("There was an error initializing the assistant. Please check the logs for details.")
+            gr.Textbox(label="Status", value="System initialization failed")
+        return interface
+    
     sample_options = [f"{i+1}. {item['query'][:50]}..." for i, item in enumerate(comprehensive_evaluation_data)]
     
     with gr.Blocks(title="Vision 2030 Assistant - Qualitative Evaluation") as interface:
@@ -596,8 +608,11 @@ def qualitative_evaluation_interface(assistant):
                 query, response, reference, metrics, sources, category, language = run_evaluation_on_sample(assistant, index)
                 sources_str = ", ".join(sources)
                 return query, response, reference, metrics, sources_str, category, language
-            except:
-                return "Error processing selection", "", "", {}, "", "", ""
+            except Exception as e:
+                print(f"Error in handle_sample_selection: {e}")
+                import traceback
+                traceback.print_exc()
+                return f"Error processing selection: {e}", "", "", {}, "", "", ""
         
         eval_button.click(
             handle_sample_selection,
@@ -614,6 +629,7 @@ def qualitative_evaluation_interface(assistant):
         )
         
         # Custom evaluation event handlers
+        @spaces.GPU  # Use GPU for custom evaluation
         def handle_custom_evaluation(query, reference):
             if not query:
                 return "Please enter a query", "", {}
@@ -639,6 +655,7 @@ def qualitative_evaluation_interface(assistant):
         )
         
         # Conversation mode event handlers
+        @spaces.GPU  # Use GPU for conversation handling
         def handle_conversation(message, history):
             if not message:
                 return history, "", ""
@@ -672,19 +689,72 @@ def qualitative_evaluation_interface(assistant):
 
 # Main function to run in Hugging Face Space
 def main():
+    # Start with a debugging report
+    print("=" * 50)
+    print("SYSTEM INITIALIZATION")
+    print("=" * 50)
+    print("Current directory:", os.getcwd())
+    print("Files in directory:", os.listdir("."))
+    print("=" * 50)
+    
     # Initialize the system
     try:
+        # First check if PDF files exist
+        pdf_files = ["saudi_vision203.pdf", "saudi_vision2030_ar.pdf"]
+        for pdf_file in pdf_files:
+            if not os.path.exists(pdf_file):
+                print(f"Warning: {pdf_file} not found!")
+                
+        # Process with initialization
+        print("Starting system initialization...")
         assistant = initialize_system()
+        
+        print("Creating interface...")
         interface = qualitative_evaluation_interface(assistant)
+        
+        print("Launching interface...")
         interface.launch()
     except Exception as e:
-        print(f"Error initializing system: {e}")
+        print(f"Error during initialization: {e}")
+        import traceback
+        traceback.print_exc()
+        
         # Create a simple error interface
-        gr.Interface(
-            fn=lambda x: f"System initialization failed: {str(e)}",
-            inputs=gr.Textbox(placeholder="System failed to initialize"),
-            outputs=gr.Textbox()
-        ).launch()
+        with gr.Blocks(title="Vision 2030 Assistant - Error") as debug_interface:
+            gr.Markdown("# Vision 2030 Assistant - Initialization Error")
+            gr.Markdown("There was an error initializing the assistant.")
+            
+            # Display error details
+            gr.Textbox(
+                value=f"Error: {str(e)}", 
+                label="Error Details", 
+                lines=5
+            )
+            
+            # Show file system status
+            files_list = "\n".join(os.listdir("."))
+            gr.Textbox(
+                value=files_list,
+                label="Files in Directory",
+                lines=10
+            )
+            
+            # Add a button to check PDFs
+            def check_pdfs():
+                result = []
+                for pdf_file in ["saudi_vision203.pdf", "saudi_vision2030_ar.pdf"]:
+                    if os.path.exists(pdf_file):
+                        size = os.path.getsize(pdf_file) / (1024 * 1024)  # Size in MB
+                        result.append(f"{pdf_file}: Found ({size:.2f} MB)")
+                    else:
+                        result.append(f"{pdf_file}: Not found")
+                return "\n".join(result)
+            
+            check_btn = gr.Button("Check PDF Files")
+            pdf_status = gr.Textbox(label="PDF Status", lines=3)
+            check_btn.click(check_pdfs, inputs=[], outputs=[pdf_status])
+            
+        debug_interface.launch()
 
 if __name__ == "__main__":
     main()