Data table tunned

app.py

@@ -2,11 +2,17 @@ import streamlit as st
 import os
 import pandas as pd
 import plotly.express as px
+import ast
 import google.generativeai as genai
 from io import StringIO
+from dotenv import load_dotenv
+
+# Load environment variables
+load_dotenv()
 
 # Configure Genai Key
-genai.configure(api_key=os.environ.get("GOOGLE_API_KEY"))
+# genai.configure(api_key=os.environ.get("GOOGLE_API_KEY"))
+genai.configure(api_key=os.getenv("GOOGLE_API_KEY"))
 
 # Function to load Google Gemini Model and provide queries as response
 def get_gemini_response(question, prompt):
@@ -36,12 +42,116 @@ id,product_name,category,price,stock_quantity,supplier,last_restock_date
     df['last_restock_date'] = pd.to_datetime(df['last_restock_date'], errors='coerce')
     return df
 
+# # Function to execute pandas query
+# def execute_pandas_query(df, query):
+#     try:
+#         # This is a very simple and unsafe way to execute queries.
+#         # In a real application, you'd need to parse the SQL and translate it to pandas operations.
+#         result = eval(f"df.{query}")
+#         return result
+#     except Exception as e:
+#         st.error(f"An error occurred: {e}")
+#         return pd.DataFrame()
+
+# # Define Your Prompt
+# prompt = [
+#     """
+#     You are an expert in converting English questions to pandas DataFrame operations!
+#     The DataFrame 'df' has the following columns: 
+#     id, product_name, category, price, stock_quantity, supplier, last_restock_date.
+    
+#     Examples:
+#     - How many products do we have in total? 
+#       The pandas operation will be: len()
+#     - What are all the products in the Electronics category? 
+#       The pandas operation will be: query("category == 'Electronics'")
+    
+#     The pandas operation should be a valid Python expression that can be applied to a DataFrame 'df'.
+#     """
+# ]
+
+# Function to execute pandas query
+# def execute_pandas_query(df, query):
+#     try:
+#         # Remove any 'df.' prefixes from the query
+#         query = query.replace('df.', '')
+        
+#         # Execute the query
+#         if query.startswith('query'):
+#             # For filtering operations
+#             result = df.query(query.split('(', 1)[1].rsplit(')', 1)[0].strip('"\''))
+#         elif query.startswith('groupby'):
+#             # For groupby operations
+#             group_col, agg_func = query.split('.', 2)[1:]
+#             result = eval(f"df.groupby('{group_col}').{agg_func}")
+#         else:
+#             # For other operations
+#             result = eval(f"df.{query}")
+        
+#         return result
+#     except Exception as e:
+#         st.error(f"An error occurred: {e}")
+#         return pd.DataFrame()
+
+# # Define Your Prompt
+# prompt = [
+#     """
+#     You are an expert in converting English questions to pandas DataFrame operations!
+#     The DataFrame 'df' has the following columns: 
+#     id, product_name, category, price, stock_quantity, supplier, last_restock_date.
+    
+#     Examples:
+#     - How many products do we have in total? 
+#       The pandas operation will be: shape[0]
+#     - What are all the products in the Electronics category? 
+#       The pandas operation will be: query("category == 'Electronics'")
+#     - What is the average price of products in each category?
+#       The pandas operation will be: groupby('category').mean()['price']
+    
+#     The pandas operation should be a valid Python expression that can be applied to a DataFrame without the 'df.' prefix.
+#     """
+# ]
+
+# Function to safely evaluate a string as a Python expression
+def safe_eval(expr, df):
+    try:
+        # Parse the expression
+        parsed = ast.parse(expr, mode='eval')
+        
+        # Define allowed names
+        allowed_names = {
+            'df': df,
+            'query': df.query,
+            'groupby': df.groupby,
+            'mean': pd.DataFrame.mean,
+            'sum': pd.DataFrame.sum,
+            'count': pd.DataFrame.count,
+            'max': pd.DataFrame.max,
+            'min': pd.DataFrame.min
+        }
+        
+        # Evaluate the expression
+        return eval(compile(parsed, '<string>', 'eval'), allowed_names)
+    except Exception as e:
+        st.error(f"Error in query execution: {e}")
+        return pd.DataFrame()
+
 # Function to execute pandas query
 def execute_pandas_query(df, query):
     try:
-        # This is a very simple and unsafe way to execute queries.
-        # In a real application, you'd need to parse the SQL and translate it to pandas operations.
-        result = eval(f"df.{query}")
+        # Remove any 'df.' prefixes from the query
+        query = query.replace('df.', '')
+        
+        # Execute the query
+        result = safe_eval(query, df)
+        
+        # Convert result to DataFrame if it's not already
+        if not isinstance(result, pd.DataFrame):
+            if isinstance(result, pd.Series):
+                result = result.to_frame()
+            else:
+                result = pd.DataFrame({'Result': [result]})
+        
         return result
     except Exception as e:
         st.error(f"An error occurred: {e}")
@@ -56,32 +166,35 @@ prompt = [
     
     Examples:
     - How many products do we have in total? 
-      The pandas operation will be: len()
+      The pandas operation will be: len(df)
     - What are all the products in the Electronics category? 
-      The pandas operation will be: query("category == 'Electronics'")
+      The pandas operation will be: df.query("category == 'Electronics'")
+    - What is the average price of products in each category?
+      The pandas operation will be: df.groupby('category')['price'].mean()
     
-    The pandas operation should be a valid Python expression that can be applied to a DataFrame 'df'.
+    The pandas operation should be a valid Python expression that can be applied to a DataFrame named 'df'.
+    Always include 'df.' at the beginning of your operations unless you're using a function like len().
     """
 ]
 
 # Streamlit App
-st.set_page_config(page_title="Department Store Analytics", layout="wide")
+st.set_page_config(page_title="AutomatiX - Department Store Analytics", layout="wide")
 
 # Load data
 df = load_data()
 
 # Sidebar for user input
-st.sidebar.title("Department Store Query Interface")
+st.sidebar.title("AutomatiX - Department Store Chat Interface")
 question = st.sidebar.text_area("Enter your question:", key="input")
 submit = st.sidebar.button("Ask Me")
 
 # Main content area
-st.title("Department Store Dashboard")
+st.title("AutomatiX - Department Store Dashboard")
 
 if submit:
-    with st.spinner("Generating query and fetching data..."):
+    with st.spinner("Generating and Fetching the data..."):
         pandas_query = get_gemini_response(question, prompt)
-        st.code(pandas_query, language="python")
+        # st.code(pandas_query, language="python")
         
         result_df = execute_pandas_query(df, pandas_query)
         
@@ -92,7 +205,7 @@ if submit:
             st.subheader("Data Table")
             st.dataframe(result_df)
             
-            # Create visualizations based on the data
+            # # Create visualizations based on the data
             st.subheader("Data Visualizations")
             
             col1, col2 = st.columns(2)
@@ -126,4 +239,9 @@ else:
 
 # Footer
 st.sidebar.markdown("---")
-st.sidebar.warning("AutomatiX - Department Store Analytics - Powered by Streamlit and Google Gemini")
+st.sidebar.info("You can ask questions like:\n"
+"1.What are all the products in the Electronics category?\n"
+"2.What is the average price of products in each category?\n"
+"3.Which products have a stock quantity less than 30?\n"
+"4.What are the top 5 most expensive products?")
+st.sidebar.warning("CopyRights@AutomatiX - Powered by Streamlit and Google Gemini")