Create app.py

app.py

@@ -0,0 +1,330 @@
+import streamlit as st
+import pandas as pd
+import matplotlib.pyplot as plt
+import seaborn as sns
+import altair as alt
+import google.generativeai as genai
+from datetime import datetime
+import os
+import re
+import json
+
+# App title and configuration
+st.set_page_config(page_title="Expense Tracker", layout="wide")
+
+# Initialize session state
+if 'expenses' not in st.session_state:
+    st.session_state.expenses = []
+if 'df' not in st.session_state:
+    st.session_state.df = pd.DataFrame(columns=['Date', 'Category', 'Amount', 'Description'])
+if 'chat_history' not in st.session_state:
+    st.session_state.chat_history = []
+
+# Load Gemini API key from secrets
+def configure_genai():
+    # For local development, use st.secrets
+    # For Hugging Face deployment, use environment variables
+    if 'GEMINI_API_KEY' in st.secrets:
+        api_key = st.secrets['GEMINI_API_KEY']
+    else:
+        api_key = os.environ.get('GEMINI_API_KEY')
+    
+    if not api_key:
+        st.error("Gemini API key not found. Please add it to the secrets or environment variables.")
+        st.stop()
+    
+    genai.configure(api_key=api_key)
+    return genai.GenerativeModel('gemini-2.0-flash')
+
+model = configure_genai()
+
+# Function to extract expense data using Gemini
+def extract_expense_data(text):
+    prompt = f"""
+    Extract expense information from the following text. 
+    Return a JSON object with these fields:
+    - date: in YYYY-MM-DD format (use today's date if not specified)
+    - category: the expense category (e.g., food, transport, entertainment)
+    - amount: the numerical amount (just the number, no currency symbol)
+    - description: brief description of the expense
+
+    Example output format:
+    {{
+        "date": "2025-03-19",
+        "category": "food",
+        "amount": 25.50,
+        "description": "lunch at cafe"
+    }}
+
+    If multiple expenses are mentioned, return an array of such objects.
+    
+    Text: {text}
+    """
+    
+    try:
+        response = model.generate_content(prompt)
+        response_text = response.text
+        
+        # Extract JSON from the response
+        json_match = re.search(r'```json\n(.*?)```', response_text, re.DOTALL)
+        if json_match:
+            json_str = json_match.group(1)
+        else:
+            # If no code block, try to find JSON directly
+            json_str = response_text
+        
+        # Parse the JSON
+        data = json.loads(json_str)
+        return data
+    except Exception as e:
+        st.error(f"Error extracting expense data: {e}")
+        return None
+
+# Function to add expenses to the dataframe
+def add_expense_to_df(expense_data):
+    if isinstance(expense_data, list):
+        # Handle multiple expenses
+        for expense in expense_data:
+            add_single_expense(expense)
+    else:
+        # Handle single expense
+        add_single_expense(expense_data)
+    
+    # Sort by date
+    st.session_state.df = st.session_state.df.sort_values(by='Date', ascending=False)
+
+def add_single_expense(expense):
+    # Convert amount to float
+    try:
+        amount = float(expense['amount'])
+    except:
+        amount = 0.0
+    
+    # Create a new row
+    new_row = pd.DataFrame({
+        'Date': [expense.get('date', datetime.now().strftime('%Y-%m-%d'))],
+        'Category': [expense.get('category', 'Other')],
+        'Amount': [amount],
+        'Description': [expense.get('description', '')]
+    })
+    
+    # Append to the dataframe
+    st.session_state.df = pd.concat([st.session_state.df, new_row], ignore_index=True)
+
+# Function to get AI insights about expenses
+def get_expense_insights(query):
+    if st.session_state.df.empty:
+        return "No expense data available yet. Please add some expenses first."
+    
+    # Convert dataframe to string representation
+    df_str = st.session_state.df.to_string()
+    
+    prompt = f"""
+    Here is a dataset of expenses:
+    {df_str}
+    
+    User query: {query}
+    
+    Please analyze this expense data and answer the query.
+    Provide your analysis in a clear and concise way.
+    If the query is about visualizations, describe what kind of chart would be helpful.
+    """
+    
+    try:
+        response = model.generate_content(prompt)
+        return response.text
+    except Exception as e:
+        return f"Error getting insights: {e}"
+
+# Function to create visualizations
+def create_visualizations():
+    if st.session_state.df.empty:
+        st.info("Add some expenses to see visualizations")
+        return
+    
+    # Create a copy of the dataframe for visualization
+    df = st.session_state.df.copy()
+    
+    # Ensure Date is datetime
+    df['Date'] = pd.to_datetime(df['Date'])
+    
+    # Create tabs for different visualizations
+    tab1, tab2, tab3 = st.tabs(["Expenses by Category", "Expenses Over Time", "Recent Expenses"])
+    
+    with tab1:
+        st.subheader("Expenses by Category")
+        category_totals = df.groupby('Category')['Amount'].sum().reset_index()
+        
+        # Create a pie chart
+        fig, ax = plt.subplots(figsize=(8, 8))
+        ax.pie(category_totals['Amount'], labels=category_totals['Category'], autopct='%1.1f%%')
+        ax.set_title('Expenses by Category')
+        st.pyplot(fig)
+        
+        # Create a bar chart
+        category_chart = alt.Chart(category_totals).mark_bar().encode(
+            x=alt.X('Category:N', sort='-y'),
+            y=alt.Y('Amount:Q'),
+            color='Category:N'
+        ).properties(
+            title='Total Expenses by Category'
+        )
+        st.altair_chart(category_chart, use_container_width=True)
+    
+    with tab2:
+        st.subheader("Expenses Over Time")
+        # Group by date and sum amounts
+        daily_totals = df.groupby(df['Date'].dt.date)['Amount'].sum().reset_index()
+        
+        # Create a line chart
+        time_chart = alt.Chart(daily_totals).mark_line(point=True).encode(
+            x='Date:T',
+            y='Amount:Q',
+            tooltip=['Date:T', 'Amount:Q']
+        ).properties(
+            title='Daily Expenses Over Time'
+        )
+        st.altair_chart(time_chart, use_container_width=True)
+    
+    with tab3:
+        st.subheader("Recent Expenses")
+        # Sort by date and get the last 10 expenses
+        recent = df.sort_values('Date', ascending=False).head(10)
+        
+        # Create a bar chart
+        recent_chart = alt.Chart(recent).mark_bar().encode(
+            x=alt.X('Description:N', sort='-y'),
+            y='Amount:Q',
+            color='Category:N',
+            tooltip=['Date:T', 'Category:N', 'Amount:Q', 'Description:N']
+        ).properties(
+            title='Most Recent Expenses'
+        )
+        st.altair_chart(recent_chart, use_container_width=True)
+
+# App layout
+st.title("💰 Expense Tracker with AI")
+
+# Sidebar for app navigation
+page = st.sidebar.radio("Navigation", ["Add Expenses", "View & Analyze", "Chat with your Data"])
+
+if page == "Add Expenses":
+    st.header("Add Your Expenses")
+    st.write("Describe your expenses in natural language, and AI will extract the details.")
+    
+    with st.form("expense_form"):
+        user_input = st.text_area(
+            "Enter your expenses:", 
+            height=100,
+            placeholder="Example: I spent $25 on lunch today, $15 on transport yesterday, and $50 on groceries on March 15th"
+        )
+        submit_button = st.form_submit_button("Add Expenses")
+    
+    if submit_button and user_input:
+        with st.spinner("Processing your expenses..."):
+            expense_data = extract_expense_data(user_input)
+            
+            if expense_data:
+                add_expense_to_df(expense_data)
+                st.success("Expenses added successfully!")
+                st.write("Extracted information:")
+                st.json(expense_data)
+            else:
+                st.error("Failed to extract expense data. Please try again with a clearer description.")
+    
+    # Show the current expenses
+    if not st.session_state.df.empty:
+        st.subheader("Your Recent Expenses")
+        st.dataframe(st.session_state.df.sort_values(by='Date', ascending=False), use_container_width=True)
+
+elif page == "View & Analyze":
+    st.header("Your Expense Data")
+    
+    # Show the current expenses as a table
+    if not st.session_state.df.empty:
+        st.dataframe(st.session_state.df.sort_values(by='Date', ascending=False), use_container_width=True)
+        
+        # Add download button
+        csv = st.session_state.df.to_csv(index=False)
+        st.download_button(
+            label="Download CSV",
+            data=csv,
+            file_name="expenses.csv",
+            mime="text/csv"
+        )
+        
+        # Show summary statistics
+        st.subheader("Summary Statistics")
+        col1, col2, col3 = st.columns(3)
+        with col1:
+            st.metric("Total Expenses", f"${st.session_state.df['Amount'].sum():.2f}")
+        with col2:
+            st.metric("Average Expense", f"${st.session_state.df['Amount'].mean():.2f}")
+        with col3:
+            st.metric("Number of Expenses", f"{len(st.session_state.df)}")
+        
+        # Create visualizations
+        st.subheader("Visualizations")
+        create_visualizations()
+    else:
+        st.info("No expense data available yet. Please add some expenses first.")
+
+elif page == "Chat with your Data":
+    st.header("Chat with Your Expense Data")
+    
+    if st.session_state.df.empty:
+        st.info("No expense data available yet. Please add some expenses first.")
+    else:
+        st.write("Ask questions about your expenses to get insights.")
+        
+        # Display chat history
+        for message in st.session_state.chat_history:
+            with st.chat_message(message["role"]):
+                st.write(message["content"])
+        
+        # Get user input
+        user_query = st.chat_input("Ask about your expenses...")
+        
+        if user_query:
+            # Add user message to chat history
+            st.session_state.chat_history.append({"role": "user", "content": user_query})
+            
+            # Display user message
+            with st.chat_message("user"):
+                st.write(user_query)
+            
+            # Get AI response
+            with st.spinner("Thinking..."):
+                response = get_expense_insights(user_query)
+            
+            # Add AI response to chat history
+            st.session_state.chat_history.append({"role": "assistant", "content": response})
+            
+            # Display AI response
+            with st.chat_message("assistant"):
+                st.write(response)
+
+# Add instructions for Hugging Face deployment in the sidebar
+with st.sidebar.expander("Deployment Instructions"):
+    st.write("""
+    ### How to deploy to Hugging Face:
+    
+    1. Save this code as `app.py`
+    2. Create a `requirements.txt` file with these dependencies:
+       ```
+       streamlit
+       pandas
+       matplotlib
+       seaborn
+       altair
+       google-generativeai
+       ```
+    3. Create a `README.md` file describing your app
+    4. Add your Gemini API key to your Hugging Face Space secrets with the name `GEMINI_API_KEY`
+    5. Push your code to a GitHub repository
+    6. Create a new Hugging Face Space, select Streamlit as the SDK, and connect your GitHub repository
+    """)
+
+# Bottom credits
+st.sidebar.markdown("---")
+st.sidebar.caption("Built with Streamlit and Gemini AI")