Update app.py

app.py

@@ -45,8 +45,7 @@ sidebar_profiles()
 def get_api_key():
     api_key = os.getenv("GROQ_API_KEY")
     if not api_key:
-        if "GROQ_API_KEY" in st.session_state:
-            api_key = st.session_state["GROQ_API_KEY"]
+        st.error("GROQ_API_KEY environment variable is not set. Please set it before running the application.")
     return api_key
 
 # Session state initialization
@@ -126,13 +125,10 @@ def retrieve_chunks(query, top_k=5):
 def build_prompt(system_prompt, context_chunks, question):
     context = "\n\n".join(context_chunks)
     return f"""{system_prompt}
-
 Context:
 {context}
-
 Question:
 {question}
-
 Answer: Please provide a comprehensive answer based only on the context provided."""
 
 def generate_answer(prompt):
@@ -213,19 +209,7 @@ st.title("📄 RAG Explorer:  AI-Powered Document Assistant")
 st.markdown("Upload a document and ask questions to get AI-powered answers with translation capabilities.")
 
 # Add API key input in sidebar
-with st.sidebar:
-    st.header("API Configuration")
-    api_key_input = st.text_input(
-        "Groq API Key", 
-        value=get_api_key() or "", 
-        type="password",
-        help="Enter your Groq API key here if not set as environment variable"
-    )
-    
-    if api_key_input:
-        st.session_state["GROQ_API_KEY"] = api_key_input
-        st.success("API key saved for this session!")
-    
+with st.sidebar:    
     # Add model selection
     st.subheader("Model Selection")
     model_choice = st.selectbox(
@@ -244,58 +228,14 @@ with st.sidebar:
     st.session_state.debug_mode = st.checkbox("Show Debug Information", value=st.session_state.debug_mode)
     
     if st.session_state.last_query_time:
-        st.subheader("Performance")
-        # Main UI for file upload
-        uploaded_file = st.file_uploader("Upload a PDF document", type="pdf")
-        
-        if uploaded_file:
-            text = extract_text_from_pdf(uploaded_file)
-            chunk_count = index_uploaded_text(text)
-            st.success(f"Indexed {chunk_count} text chunks from the uploaded PDF.")
-        
-            # User query input
-            question = st.text_input("Ask a question based on the document")
-        
-            if question:
-                context_chunks = retrieve_chunks(question)
-                system_prompt = "You are a helpful assistant answering only from the document context provided."
-                prompt = build_prompt(system_prompt, context_chunks, question)
-                answer = generate_answer(prompt)
-                
-                # Show response
-                st.markdown("### 💬 AI Answer")
-                st.write(answer)
+         st.subheader("About")
+         st.markdown("""
+         This app uses Retrieval-Augmented Generation (RAG) to answer questions about uploaded documents.
+         1. Upload a document
+         2. Ask a question
+         3. Translate responses to other languages
+         """)
                 
-                # Optional: Translate
-                translate_option = st.selectbox("Translate the answer to", ["None", "Urdu", "Hindi", "Spanish", "French"])
-                language_codes = {
-                    "Urdu": "ur",
-                    "Hindi": "hi",
-                    "Spanish": "es",
-                    "French": "fr"
-                }
-        
-                if translate_option != "None":
-                    translated_text = GoogleTranslator(source='auto', target='fr').translate(text)
-                    st.markdown(f"### 🌐 Translated Answer ({translate_option})")
-                    st.write(translated_text)
-        
-                    # Audio playback
-                    audio_file_path = text_to_speech(translated_text, language_codes.get(translate_option, "en"))
-                    if audio_file_path:
-                        st.audio(audio_file_path, format="audio/mp3")
-
-            elif uploaded_file is None:
-                st.info("Please upload a PDF document to begin.")
-            
-                
-                st.subheader("About")
-                st.markdown("""
-                This app uses Retrieval-Augmented Generation (RAG) to answer questions about uploaded documents.
-                1. Upload a document
-                2. Ask a question
-                3. Optionally translate responses to other languages
-                """)
 
 # Main content area
 col1, col2 = st.columns([2, 1])
@@ -314,13 +254,41 @@ with col1:
             st.success(f"Document indexed successfully! Created {total_chunks} chunks.")
             
             # Display document preview
-            with st.expander("Document Preview"):
-                st.text_area("First 1000 characters of document", raw_text[:20000], height=200)
+            with st.expander("Document Preview"):          
+                # Extract and display key points
+                st.subheader("Key Points")
+                
+                # Simple algorithm to extract potential key points (sentences that might be important)
+                sentences = raw_text.split('. ')
+                key_points = []
+                
+                # Look for sentences that might be key points (contains keywords, not too long/short)
+                for sentence in sentences[:50]:  # Check first 50 sentences
+                    sentence = sentence.strip()
+                    if len(sentence) > 15 and len(sentence) < 200:  # Reasonable length for a key point
+                        # Keywords that might indicate important information
+                        important_keywords = ["important", "key", "significant", "main", "primary", "essential", 
+                                             "critical", "crucial", "fundamental", "major", "summary", "conclusion"]
+                        
+                        if any(keyword in sentence.lower() for keyword in important_keywords) or sentence.endswith(':'):
+                            key_points.append(sentence)
+                
+                # If we didn't find obvious key points, just take some representative sentences
+                if len(key_points) < 3:
+                    key_points = [s.strip() for s in sentences[:50:10] if len(s.strip()) > 15][:5]  # Every 10th sentence from first 50
+                
+                # Display the key points as bullets
+                for point in key_points[:5]:  # Show up to 5 key points
+                    st.markdown(f"• {point}")
+                
+                if not key_points:
+                    st.info("No clear key points detected. Try exploring the full document.")
 
 with col2:
     if st.session_state.chunks:
         st.info(f"Document chunks: {len(st.session_state.chunks)}")
 
+# Query and answer section
 # Query and answer section
 st.divider()
 query = st.text_input("Ask a question about the document")
@@ -340,7 +308,11 @@ with col2:
     }
     lang_code = language_codes[language]
 
-if query:
+# Add a submit button
+submit_button = st.button("Get Answer", type="primary", key="submit_query")
+
+# Only process when the button is clicked and there's a query
+if submit_button and query:
     if index.ntotal == 0:
         st.warning("Please upload and index a document first.")
     else: