Code Editor for app.py

import streamlit as st
import pandas as pd
import numpy as np

# Create sample medicine prescription data
@st.cache_data
def create_sample_data():
    """Create sample prescription data for demonstration"""
    medicines = [
        'Aspirin', 'Ibuprofen', 'Acetaminophen', 'Lisinopril', 'Metformin',
        'Atorvastatin', 'Omeprazole', 'Amoxicillin', 'Metoprolol', 'Amlodipine',
        'Hydrochlorothiazide', 'Simvastatin', 'Losartan', 'Levothyroxine', 'Warfarin',
        'Prednisone', 'Albuterol', 'Insulin', 'Clopidogrel', 'Gabapentin'
    ]
    
    # Generate combinations of medicines that are commonly prescribed together
    data = []
    np.random.seed(42)  # For reproducible results
    
    for primary_med in medicines:
        # Each medicine can be prescribed with 3-8 other medicines
        num_combinations = np.random.randint(3, 9)
        secondary_meds = np.random.choice(
            [m for m in medicines if m != primary_med], 
            size=num_combinations, 
            replace=False
        )
        
        for secondary_med in secondary_meds:
            # Add some realistic frequency data
            frequency = np.random.randint(1, 100)
            data.append({
                'Primary_Medicine': primary_med,
                'Prescribed_With': secondary_med,
                'Frequency': frequency,
                'Condition': np.random.choice([
                    'Hypertension', 'Diabetes', 'Heart Disease', 'Arthritis', 
                    'Infection', 'Pain Management', 'Cholesterol', 'Thyroid'
                ])
            })
    
    return pd.DataFrame(data)

def load_data():
    """Load data from uploaded file or use sample data"""
    uploaded_file = st.file_uploader(
        "Upload your medicine prescription Excel file",
        type=['xlsx', 'xls', 'csv'],
        help="Upload an Excel/CSV file with columns: Primary_Medicine, Prescribed_With, Frequency (optional), Condition (optional)"
    )
    
    if uploaded_file is not None:
        try:
            if uploaded_file.name.endswith('.csv'):
                df = pd.read_csv(uploaded_file)
            else:
                df = pd.read_excel(uploaded_file)
            
            # Validate required columns
            required_cols = ['Primary_Medicine', 'Prescribed_With']
            if not all(col in df.columns for col in required_cols):
                st.error(f"Please ensure your file has these columns: {required_cols}")
                return None
                
            return df
        except Exception as e:
            st.error(f"Error reading file: {str(e)}")
            return None
    else:
        st.info("Using sample data. Upload your own Excel file to use real prescription data.")
        return create_sample_data()

def main():
    st.title("🏥 Medicine Prescription Finder")
    st.markdown("Enter a medicine name to find which other medicines are commonly prescribed with it.")
    
    # Load data
    df = load_data()
    
    if df is None:
        return
    
    # Display data info
    with st.expander("📊 View Data Summary"):
        st.write(f"**Total prescription combinations:** {len(df)}")
        st.write(f"**Unique medicines:** {df['Primary_Medicine'].nunique()}")
        
        # Show sample of data
        st.write("**Sample data:**")
        st.dataframe(df.head(10))
    
    # Medicine search
    st.subheader("🔍 Search for Medicine Combinations")
    
    # Get unique medicines for dropdown
    all_medicines = sorted(df['Primary_Medicine'].unique())
    
    # Medicine selection
    selected_medicine = st.selectbox(
        "Select or type a medicine name:",
        options=all_medicines,
        index=0
    )
    
    if selected_medicine:
        # Filter data for selected medicine
        filtered_data = df[df['Primary_Medicine'] == selected_medicine].copy()
        
        if not filtered_data.empty:
            st.success(f"Found {len(filtered_data)} medicines commonly prescribed with **{selected_medicine}**")
            
            # Sort by frequency if available
            if 'Frequency' in filtered_data.columns:
                filtered_data = filtered_data.sort_values('Frequency', ascending=False)
            
            # Display results
            col1, col2 = st.columns([2, 1])
            
            with col1:
                st.subheader("💊 Medicines Prescribed Together")
                
                # Create a clean display dataframe
                display_df = filtered_data[['Prescribed_With']].copy()
                display_df.index = range(1, len(display_df) + 1)
                
                if 'Frequency' in filtered_data.columns:
                    display_df['Frequency'] = filtered_data['Frequency'].values
                
                if 'Condition' in filtered_data.columns:
                    display_df['Common Condition'] = filtered_data['Condition'].values
                
                st.dataframe(display_df, use_container_width=True)
            
            with col2:
                st.subheader("📈 Statistics")
                
                if 'Frequency' in filtered_data.columns:
                    st.metric("Most Common Combination", 
                             filtered_data.iloc[0]['Prescribed_With'],
                             f"Frequency: {filtered_data.iloc[0]['Frequency']}")
                    
                    st.metric("Average Frequency", 
                             f"{filtered_data['Frequency'].mean():.1f}")
                
                if 'Condition' in filtered_data.columns:
                    st.write("**Common Conditions:**")
                    condition_counts = filtered_data['Condition'].value_counts()
                    for condition, count in condition_counts.head(3).items():
                        st.write(f"• {condition}: {count}")
        
        else:
            st.warning(f"No prescription combinations found for **{selected_medicine}**")
    
    # Additional features
    st.subheader("🔄 Reverse Search")
    st.markdown("Find which medicines commonly include a specific medicine in their combinations:")
    
    reverse_medicine = st.selectbox(
        "Select medicine to find what it's prescribed with:",
        options=sorted(df['Prescribed_With'].unique()),
        key="reverse_search"
    )
    
    if reverse_medicine:
        reverse_data = df[df['Prescribed_With'] == reverse_medicine].copy()
        
        if not reverse_data.empty:
            st.write(f"**{reverse_medicine}** is commonly prescribed WITH:")
            
            if 'Frequency' in reverse_data.columns:
                reverse_data = reverse_data.sort_values('Frequency', ascending=False)
            
            reverse_display = reverse_data[['Primary_Medicine']].copy()
            reverse_display.index = range(1, len(reverse_display) + 1)
            
            if 'Frequency' in reverse_data.columns:
                reverse_display['Frequency'] = reverse_data['Frequency'].values
            
            st.dataframe(reverse_display.head(10), use_container_width=True)

if __name__ == "__main__":
    main()