Updated app.py to version 13

app.py

@@ -13,38 +13,69 @@ st.set_page_config(
 st.title("📊 LLM Data Analyzer")
 st.write("*Analyze data and chat with AI - Powered by Hugging Face Spaces*")
 
-# Simple AI responses without API calls
-def get_ai_response(prompt):
-    """Generate simple AI-like responses without external API"""
+# Store dataframe in session state globally
+if "current_df" not in st.session_state:
+    st.session_state.current_df = None
+
+# AI response function with data awareness
+def get_ai_response(prompt, df=None):
+    """Generate AI responses with data awareness"""
     prompt_lower = prompt.lower()
     
-    # Data analysis responses
-    if "average" in prompt_lower or "mean" in prompt_lower:
-        return "Based on the data summary, the average values can be calculated from the statistical measures shown. For more detailed analysis, look at the mean values in the data description."
-    elif "trend" in prompt_lower or "pattern" in prompt_lower:
-        return "The data shows various patterns. Examine the min, max, and std deviation values to understand the distribution and trends in your dataset."
-    elif "correlation" in prompt_lower or "relationship" in prompt_lower:
-        return "To understand relationships between columns, look at how values change together. The standard deviation and percentiles in the summary can give insights."
-    elif "outlier" in prompt_lower or "unusual" in prompt_lower:
-        return "Check the min/max values and compare them to the mean and median. Large differences suggest outliers in your data."
-    elif "summary" in prompt_lower or "overview" in prompt_lower:
-        return "The data summary shows key statistics including count, mean, standard deviation, min, 25%, 50%, 75%, and max values for each column."
-    elif "salary" in prompt_lower:
-        return "Based on the salary column, the average salary is around $61,000. Salaries range from $50,000 to $75,000, with most employees earning between $55,000-$65,000."
-    elif "age" in prompt_lower:
-        return "The average age is 30 years old. Ages range from 25 to 35, showing a diverse age group with good experience spread."
-    elif "department" in prompt_lower:
-        return "The company has employees in Sales, IT, and HR departments. IT has 2 employees, Sales has 2, and HR has 1 person represented in this dataset."
-    
-    # General chat responses
-    elif "hello" in prompt_lower or "hi" in prompt_lower:
-        return "Hello! I'm the LLM Data Analyzer. I can help you understand your data better. Use the tabs to analyze data or paste CSV content!"
+    # If we have data, provide data-specific responses
+    if df is not None and not df.empty:
+        # Try to generate data-aware responses
+        numeric_cols = df.select_dtypes(include=['number']).columns.tolist()
+        
+        if "average" in prompt_lower or "mean" in prompt_lower:
+            if numeric_cols:
+                means = df[numeric_cols].mean()
+                return f"📊 **Average values for numeric columns:**\n" + "\n".join([f"- {col}: {means[col]:.2f}" for col in means.index])
+            return "The data summary shows average values for numeric columns."
+        
+        elif "max" in prompt_lower or "highest" in prompt_lower:
+            if numeric_cols:
+                maxes = df[numeric_cols].max()
+                return f"📊 **Maximum values for numeric columns:**\n" + "\n".join([f"- {col}: {maxes[col]}" for col in maxes.index])
+            return "No numeric columns found to show max values."
+        
+        elif "min" in prompt_lower or "lowest" in prompt_lower:
+            if numeric_cols:
+                mins = df[numeric_cols].min()
+                return f"📊 **Minimum values for numeric columns:**\n" + "\n".join([f"- {col}: {mins[col]}" for col in mins.index])
+            return "No numeric columns found to show min values."
+        
+        elif "count" in prompt_lower or "rows" in prompt_lower or "how many" in prompt_lower:
+            return f"📊 Your dataset has **{len(df)} rows** and **{len(df.columns)} columns**.\n\nColumns: {', '.join(df.columns.tolist())}"
+        
+        elif "columns" in prompt_lower or "fields" in prompt_lower or "headers" in prompt_lower:
+            return f"📊 **Dataset Columns ({len(df.columns)}):**\n" + "\n".join([f"- {col}" for col in df.columns])
+        
+        elif "summary" in prompt_lower or "overview" in prompt_lower or "describe" in prompt_lower:
+            summary = f"📊 **Data Summary:**\n"
+            summary += f"- Rows: {len(df)}\n"
+            summary += f"- Columns: {len(df.columns)}\n"
+            summary += f"- Column names: {', '.join(df.columns.tolist())}\n"
+            if numeric_cols:
+                summary += f"- Numeric columns: {', '.join(numeric_cols)}\n"
+            return summary
+        
+        elif "trend" in prompt_lower or "pattern" in prompt_lower:
+            return "The data shows various patterns. Check the statistics above to identify trends in the numeric columns."
+        
+        else:
+            # Generic response with data context
+            return f"I have access to your data with {len(df)} rows and {len(df.columns)} columns: {', '.join(df.columns.tolist())}. Ask me specific questions like: What's the average? Show me the max values? How many rows?"
+    
+    # Fallback responses (no data)
+    if "hello" in prompt_lower or "hi" in prompt_lower:
+        return "Hello! I'm the LLM Data Analyzer. Load some data first, then ask me questions about it!"
     elif "what can you do" in prompt_lower or "help" in prompt_lower:
-        return "I can help you: 1) Paste CSV data 2) View statistics 3) Answer questions about your data 4) Have conversations. Try pasting CSV content in the Upload tab!"
+        return "I can: 1) Load demo data 2) Analyze your CSV 3) Answer questions about averages, max, min, columns, etc. 4) Chat about your data!"
     elif "thank" in prompt_lower:
-        return "You're welcome! Feel free to ask more questions about your data anytime."
+        return "You're welcome! Ask me anything about your data!"
     else:
-        return "That's an interesting question! To get the most accurate analysis, please provide your data and ask specific questions about the columns and values."
+        return "Please load some data first (click 'Load Demo Data' or paste CSV), then ask me questions about it!"
 
 # Create tabs
 tab1, tab2, tab3 = st.tabs(["📤 Paste Data", "💬 Chat", "📊 About"])
@@ -55,7 +86,7 @@ tab1, tab2, tab3 = st.tabs(["📤 Paste Data", "💬 Chat", "📊 About"])
 with tab1:
     st.header("📤 Analyze Data")
     
-    st.info("💡 HF Spaces file upload has issues. Use one of these methods instead:")
+    st.info("💡 Load demo data or paste your CSV to start analyzing!")
     
     # Demo mode
     if st.button("📌 Load Demo Data (Click to test)", use_container_width=True):
@@ -66,7 +97,7 @@ Charlie,35,75000,HR,8
 David,28,55000,Sales,3
 Eve,32,65000,IT,6"""
         st.session_state.csv_data = demo_csv
-        st.success("✅ Demo data loaded! Scroll down to see analysis.")
+        st.success("✅ Demo data loaded! Check the Chat tab to ask questions about it.")
     
     st.subheader("Or paste your CSV data here:")
     csv_text = st.text_area(
@@ -82,6 +113,9 @@ Eve,32,65000,IT,6"""
             # Parse CSV from text
             df = pd.read_csv(io.StringIO(csv_text))
             
+            # Store in session state for chat to access
+            st.session_state.current_df = df
+            
             st.success(f"✅ Data loaded: {df.shape[0]} rows, {df.shape[1]} columns")
             
             # Display data preview
@@ -114,12 +148,12 @@ Eve,32,65000,IT,6"""
             st.subheader("❓ Ask AI About Your Data")
             question = st.text_input(
                 "What would you like to know about this data?",
-                placeholder="e.g., What is the average salary? What patterns do you see?",
+                placeholder="e.g., What is the average salary? How many rows?",
                 key="data_question"
             )
             
             if question:
-                response = get_ai_response(question)
+                response = get_ai_response(question, df)
                 st.success("✅ Analysis Complete")
                 st.write(response)
         
@@ -132,7 +166,14 @@ Eve,32,65000,IT,6"""
 # ============================================================================
 with tab2:
     st.header("💬 Chat with AI Assistant")
-    st.write("Have a conversation about data analysis and AI.")
+    st.write("Have a conversation about your data.")
+    
+    # Show current data status
+    if st.session_state.current_df is not None:
+        st.success(f"✅ Data loaded: {len(st.session_state.current_df)} rows, {len(st.session_state.current_df.columns)} columns")
+        st.write(f"Columns: {', '.join(st.session_state.current_df.columns.tolist())}")
+    else:
+        st.warning("⚠️ No data loaded yet. Go to 'Paste Data' tab and load data first!")
     
     # Initialize session state for chat history
     if "messages" not in st.session_state:
@@ -146,7 +187,7 @@ with tab2:
     # Chat input
     user_input = st.text_input(
         "Type your message:",
-        placeholder="Ask me anything...",
+        placeholder="Ask me about your data...",
         key="chat_input"
     )
     
@@ -154,8 +195,8 @@ with tab2:
         # Add user message immediately
         st.session_state.messages.append({"role": "user", "content": user_input})
         
-        # Get response
-        response = get_ai_response(user_input)
+        # Get response with data context
+        response = get_ai_response(user_input, st.session_state.current_df)
         
         # Add assistant message
         st.session_state.messages.append({
@@ -188,17 +229,24 @@ with tab3:
     ### ⚡ Features
     
     1. **Data Analysis**: Paste CSV and analyze your data
-    2. **Chat**: Have conversations about data insights
+    2. **Smart Chat**: Chat with AI about your data
     3. **Statistics**: View comprehensive data summaries
     4. **Demo Mode**: Test with sample data instantly
     
     ### 📝 How to Use
     
-    1. **Click "Load Demo Data"** - See it in action
-    2. **Or paste your own CSV** - Headers + comma-separated values
-    3. **Review data preview** - See your data in table format
-    4. **Ask questions** - Get AI-powered analysis
-    5. **Chat** - Have conversations about your analysis
+    1. **Load Data** - Click "Load Demo Data" or paste your CSV
+    2. **View Preview** - See your data in table format
+    3. **Chat** - Go to Chat tab and ask questions about your data
+    4. **Get Insights** - AI analyzes and answers questions
+    
+    ### 💡 Example Questions
+    
+    - "What's the average salary?"
+    - "Show me the maximum values"
+    - "How many rows do I have?"
+    - "What columns are in the data?"
+    - "Give me a summary"
     
     ### 📋 CSV Format Example
     
@@ -217,14 +265,13 @@ with tab3:
     
     ### 🛠 Troubleshooting
     
-    **Why can't I upload files?**
-    - HF Spaces file upload widget has issues in free tier
-    - Solution: Paste CSV content directly instead
+    **Chat can't see my data?**
+    - Make sure to load data in the "Paste Data" tab first
+    - Then go to "Chat" tab - it will show your data status
     
     **How do I format CSV?**
     - First line: column headers separated by commas
     - Following lines: data values separated by commas
-    - No quotes needed unless data contains commas
     
     ### 🔗 Links
     
@@ -233,7 +280,7 @@ with tab3:
     
     ---
     
-    **Version:** 2.0 | **Last Updated:** Dec 2025
+    **Version:** 2.1 | **Last Updated:** Dec 2025
     
-    💡 **Note:** This version uses CSV paste and demo mode to work around HF Spaces limitations.
+    💡 **Note:** Chat now has access to your data! Load data first, then ask questions.
     """)