fixing bug

app.py

@@ -5,15 +5,19 @@ from langchain_community.document_loaders import PyMuPDFLoader
 import requests
 from groq import Groq
 import os
-# from dotenv import load_dotenv
+from dotenv import load_dotenv
 import tempfile
 from googlesearch import search
 from bs4 import BeautifulSoup
 from urllib.parse import urljoin, urlparse
 import re
 
-# load_dotenv()
-print(os.getenv("GROQ_API_KEY"))
+
+
+load_dotenv()
+# Check if GROQ_API_KEY is available
+if not os.getenv("GROQ_API_KEY"):
+    print("Warning: GROQ_API_KEY not found in environment variables")
 class DocumentState(TypedDict):
     documents: list[dict]
     summaries: list[str]
@@ -299,6 +303,108 @@ def process_pdf_and_chat(pdf_file, message, history, system_message, max_tokens,
     except Exception as e:
         return history + [(message, f"Error processing PDF: {str(e)}")]
 
+def respond_messages(message, history, system_message, max_tokens, temperature, top_p, enable_search=False):
+    """Enhanced chat function with optional Google search - returns just the response text"""
+    if enable_search:
+        # Use search workflow
+        search_graph = create_search_graph()
+        state = {
+            'documents': [],
+            'summaries': [],
+            'search_results': [],
+            'search_query': message,
+            'needs_search': False
+        }
+        
+        final_state = search_graph.invoke(state)
+        
+        if final_state['summaries']:
+            response = final_state['summaries'][0]
+        else:
+            # Fallback to regular LLM response
+            prompt = f"{system_message}\n\nUser: {message}"
+            llm_response = get_llm_response(prompt)
+            response = llm_response["response"]
+    else:
+        # Regular chat without search
+        prompt = f"{system_message}\n\nUser: {message}"
+        llm_response = get_llm_response(prompt)
+        response = llm_response["response"]
+    
+    return response
+
+def process_pdf_and_chat_messages(pdf_file, message, history, system_message, max_tokens, temperature, top_p, enable_search=False):
+    """Enhanced PDF processing function - returns just the response text"""
+    if pdf_file is None:
+        return "Please upload a PDF file first."
+    
+    try:
+        # Create a temporary file path for the uploaded PDF
+        with tempfile.NamedTemporaryFile(delete=False, suffix='.pdf') as tmp_file:
+            tmp_file.write(pdf_file.read())
+            tmp_pdf_path = tmp_file.name
+        
+        # Check if user wants to search for additional information
+        search_keywords = ['search', 'find more', 'additional info', 'more information', 'research']
+        if enable_search and any(keyword in message.lower() for keyword in search_keywords):
+            # Use search workflow for additional information
+            search_graph = create_search_graph()
+            search_state = {
+                'documents': [],
+                'summaries': [],
+                'search_results': [],
+                'search_query': message,
+                'needs_search': True
+            }
+            
+            search_result = search_graph.invoke(search_state)
+            
+            # Also process the PDF
+            def document_extractor_with_path(state: DocumentState) -> DocumentState:
+                return document_extractor_agent(state, tmp_pdf_path)
+            
+            talking_documents = StateGraph(DocumentState)
+            talking_documents.add_node('document_extractor', document_extractor_with_path)
+            talking_documents.add_node('document_summarizer', document_summarizer_agent)
+            talking_documents.set_entry_point('document_extractor')
+            talking_documents.add_edge('document_extractor', 'document_summarizer')
+            pdf_graph = talking_documents.compile()
+            
+            pdf_state = {'documents': [], 'summaries': []}
+            pdf_result = pdf_graph.invoke(pdf_state)
+            
+            # Combine PDF and search results
+            combined_response = f"**PDF Summary:**\n{pdf_result['summaries'][0] if pdf_result['summaries'] else 'No summary available'}\n\n**Additional Information from Web:**\n{search_result['summaries'][0] if search_result['summaries'] else 'No additional information found'}"
+            
+            response = combined_response
+        else:
+            # Regular PDF processing
+            def document_extractor_with_path(state: DocumentState) -> DocumentState:
+                return document_extractor_agent(state, tmp_pdf_path)
+            
+            talking_documents = StateGraph(DocumentState)
+            talking_documents.add_node('document_extractor', document_extractor_with_path)
+            talking_documents.add_node('document_summarizer', document_summarizer_agent)
+            talking_documents.set_entry_point('document_extractor')
+            talking_documents.add_edge('document_extractor', 'document_summarizer')
+            graph = talking_documents.compile()
+            
+            state = {'documents': [], 'summaries': []}
+            final_state = graph.invoke(state)
+            
+            if final_state['summaries']:
+                response = final_state['summaries'][0]
+            else:
+                response = "Unable to process the PDF. Please check the file format."
+        
+        # Clean up temporary file
+        os.unlink(tmp_pdf_path)
+        
+        return response
+    
+    except Exception as e:
+        return f"Error processing PDF: {str(e)}"
+
 def respond(message, history, system_message, max_tokens, temperature, top_p, enable_search=False):
     """Enhanced chat function with optional Google search"""
     if enable_search:
@@ -347,20 +453,20 @@ with gr.Blocks() as demo:
             top_p = gr.Slider(minimum=0.1, maximum=1.0, value=0.95, step=0.05, label="Top-p (nucleus sampling)")
         
         with gr.Column(scale=2):
-            chatbot = gr.Chatbot()
+            chatbot = gr.Chatbot(type='messages')
             msg = gr.Textbox(label="Message")
             clear = gr.Button("Clear")
     
     def user_input(message, history):
-        return "", history + [(message, None)]
+        return "", history + [{"role": "user", "content": message}]
     
     def bot_response(history, pdf_file, enable_search, system_message, max_tokens, temperature, top_p):
-        message = history[-1][0]
+        message = history[-1]["content"]
         if pdf_file is not None:
-            new_history = process_pdf_and_chat(pdf_file, message, history[:-1], system_message, max_tokens, temperature, top_p, enable_search)
+            response = process_pdf_and_chat_messages(pdf_file, message, history[:-1], system_message, max_tokens, temperature, top_p, enable_search)
         else:
-            new_history = respond(message, history[:-1], system_message, max_tokens, temperature, top_p, enable_search)
-        return new_history
+            response = respond_messages(message, history[:-1], system_message, max_tokens, temperature, top_p, enable_search)
+        return history[:-1] + [{"role": "user", "content": message}, {"role": "assistant", "content": response}]
     
     msg.submit(user_input, [msg, chatbot], [msg, chatbot], queue=False).then(
         bot_response, [chatbot, pdf_upload, enable_search, system_message, max_tokens, temperature, top_p], chatbot
@@ -368,4 +474,4 @@ with gr.Blocks() as demo:
     clear.click(lambda: None, None, chatbot, queue=False)
 
 if __name__ == "__main__":
-    demo.launch()
+    demo.launch(share=True)