# app_eda.py
import streamlit as st
import pandas as pd
import numpy as np
import plotly.express as px
import plotly.graph_objects as go
from scipy import stats
import seaborn as sns
import matplotlib.pyplot as plt

# Page configuration
st.set_page_config(
    page_title="EDA Dashboard",
    page_icon="📊",
    layout="wide",
    initial_sidebar_state="expanded"
)

# Custom CSS
st.markdown("""
<style>
    .main-header {
        font-size: 2.5rem;
        color: #1f77b4;
        text-align: center;
        margin-bottom: 2rem;
    }
    .metric-card {
        background-color: #f0f2f6;
        padding: 1rem;
        border-radius: 10px;
        margin: 0.5rem 0;
    }
</style>
""", unsafe_allow_html=True)

@st.cache_data
def load_data(uploaded_file):
    """Load data dengan caching untuk performance"""
    try:
        if uploaded_file.name.endswith('.csv'):
            return pd.read_csv(uploaded_file)
        elif uploaded_file.name.endswith('.xlsx'):
            return pd.read_excel(uploaded_file)
        else:
            st.error("Format file tidak didukung")
            return None
    except Exception as e:
        st.error(f"Error loading file: {e}")
        return None

def show_data_overview(df):
    """Menampilkan overview data"""
    st.header("📈 Data Overview")
    
    col1, col2, col3, col4 = st.columns(4)
    
    with col1:
        st.metric("Total Rows", df.shape[0])
    with col2:
        st.metric("Total Columns", df.shape[1])
    with col3:
        st.metric("Missing Values", df.isnull().sum().sum())
    with col4:
        st.metric("Memory Usage", f"{df.memory_usage(deep=True).sum() / 1024**2:.2f} MB")
    
    # Data preview
    st.subheader("Data Preview")
    st.dataframe(df, use_container_width=True)
    
    # Data types
    st.subheader("Data Types")
    dtype_df = pd.DataFrame(df.dtypes, columns=['Data Type'])
    st.dataframe(dtype_df)

def show_missing_analysis(df):
    """Analisis missing values"""
    st.header("🔍 Missing Values Analysis")
    
    missing_data = df.isnull().sum()
    missing_percent = (missing_data / len(df)) * 100
    
    missing_df = pd.DataFrame({
        'Column': missing_data.index,
        'Missing Count': missing_data.values,
        'Missing Percentage': missing_percent.values
    }).sort_values('Missing Count', ascending=False)
    
    # Filter hanya yang ada missing values
    missing_df = missing_df[missing_df['Missing Count'] > 0]
    
    if len(missing_df) > 0:
        st.dataframe(missing_df, use_container_width=True)
        
        # Visualisasi missing values
        fig = px.bar(missing_df, x='Column', y='Missing Percentage', 
                     title='Missing Values Percentage by Column')
        st.plotly_chart(fig, use_container_width=True)
    else:
        st.success("🎉 Tidak ada missing values dalam dataset!")

def show_statistical_analysis(df):
    """Analisis statistik"""
    st.header("📊 Statistical Analysis")
    
    # Pilih kolom numerik
    numeric_cols = df.select_dtypes(include=[np.number]).columns.tolist()
    
    if numeric_cols:
        selected_col = st.selectbox("Pilih kolom untuk analisis:", numeric_cols)
        
        if selected_col:
            col_data = df[selected_col].dropna()
            
            col1, col2 = st.columns(2)
            
            with col1:
                st.subheader("Descriptive Statistics")
                stats_df = pd.DataFrame({
                    'Statistic': ['Count', 'Mean', 'Std Dev', 'Min', '25%', '50%', '75%', 'Max', 'Skewness', 'Kurtosis'],
                    'Value': [
                        len(col_data),
                        col_data.mean(),
                        col_data.std(),
                        col_data.min(),
                        col_data.quantile(0.25),
                        col_data.median(),
                        col_data.quantile(0.75),
                        col_data.max(),
                        col_data.skew(),
                        col_data.kurtosis()
                    ]
                })
                st.dataframe(stats_df, use_container_width=True)
            
            with col2:
                st.subheader("Distribution")
                fig = px.histogram(col_data, x=selected_col, 
                                 title=f'Distribution of {selected_col}',
                                 marginal="box")
                st.plotly_chart(fig, use_container_width=True)

def show_correlation_analysis(df):
    """Analisis korelasi"""
    st.header("🔗 Correlation Analysis")
    
    numeric_df = df.select_dtypes(include=[np.number])
    
    if len(numeric_df.columns) > 1:
        # Heatmap korelasi
        corr_matrix = numeric_df.corr()
        
        fig = px.imshow(corr_matrix,
                       title="Correlation Heatmap",
                       color_continuous_scale="RdBu",
                       aspect="auto")
        st.plotly_chart(fig, use_container_width=True)
        
        # Nilai korelasi detail
        st.subheader("Detailed Correlation Values")
        st.dataframe(corr_matrix.style.background_gradient(cmap='RdBu'), 
                    use_container_width=True)

def main():
    st.markdown('<div class="main-header">📊 Exploratory Data Analysis Dashboard</div>', 
                unsafe_allow_html=True)
    
    # File upload
    uploaded_file = st.sidebar.file_uploader(
        "Upload dataset Anda", 
        type=['csv', 'xlsx'],
        help="Support CSV dan Excel files"
    )
    
    if uploaded_file is not None:
        df = load_data(uploaded_file)
        
        if df is not None:
            # Sidebar controls
            st.sidebar.header("Analysis Options")
            show_overview = st.sidebar.checkbox("Data Overview", True)
            show_missing = st.sidebar.checkbox("Missing Analysis", True)
            show_stats = st.sidebar.checkbox("Statistical Analysis", True)
            show_corr = st.sidebar.checkbox("Correlation Analysis", True)
            
            # Main analysis sections
            if show_overview:
                show_data_overview(df)
            
            if show_missing:
                show_missing_analysis(df)
            
            if show_stats:
                show_statistical_analysis(df)
            
            if show_corr:
                show_correlation_analysis(df)
            
            # Download processed data
            st.sidebar.header("Export Data")
            if st.sidebar.button("Download Processed Data as CSV"):
                csv = df.to_csv(index=False)
                st.sidebar.download_button(
                    label="Download CSV",
                    data=csv,
                    file_name="processed_data.csv",
                    mime="text/csv"
                )
    else:
        st.info("👆 Please upload your dataset to begin analysis")
        st.markdown("""
        ### Supported Features:
        - **Data Overview**: Basic statistics and data preview
        - **Missing Analysis**: Identify and visualize missing values
        - **Statistical Analysis**: Detailed descriptive statistics
        - **Correlation Analysis**: Heatmaps and correlation values
        - **Data Export**: Download processed data
        """)

if __name__ == "__main__":
    main()

# app_ml.py
import streamlit as st
import pandas as pd
import numpy as np
from sklearn.model_selection import train_test_split, cross_val_score
from sklearn.ensemble import RandomForestClassifier, RandomForestRegressor
from sklearn.linear_model import LogisticRegression, LinearRegression
from sklearn.svm import SVC, SVR
from sklearn.metrics import (accuracy_score, classification_report, 
                           confusion_matrix, mean_squared_error, r2_score)
import plotly.express as px
import plotly.figure_factory as ff
import joblib
import io

# Page config
st.set_page_config(
    page_title="ML Model Trainer",
    page_icon="🤖",
    layout="wide"
)

@st.cache_data
def load_data(uploaded_file):
    """Load dataset dengan caching"""
    if uploaded_file.name.endswith('.csv'):
        return pd.read_csv(uploaded_file)
    elif uploaded_file.name.endswith('.xlsx'):
        return pd.read_excel(uploaded_file)
    return None

def evaluate_model(model, X_test, y_test, problem_type):
    """Evaluate model performance"""
    y_pred = model.predict(X_test)
    
    if problem_type == "classification":
        accuracy = accuracy_score(y_test, y_pred)
        report = classification_report(y_test, y_pred, output_dict=True)
        cm = confusion_matrix(y_test, y_pred)
        return accuracy, report, cm, y_pred
    else:
        mse = mean_squared_error(y_test, y_pred)
        r2 = r2_score(y_test, y_pred)
        return mse, r2, y_pred

def main():
    st.title("🤖 Machine Learning Model Trainer")
    st.markdown("Train and evaluate multiple ML models with ease")
    
    # File upload
    uploaded_file = st.sidebar.file_uploader(
        "Upload Dataset", 
        type=['csv', 'xlsx'],
        help="Dataset harus mengandung features dan target variable"
    )
    
    if uploaded_file is not None:
        df = load_data(uploaded_file)
        
        if df is not None:
            st.success(f"✅ Dataset loaded successfully: {df.shape[0]} rows, {df.shape[1]} columns")
            
            # Problem type selection
            problem_type = st.sidebar.radio(
                "Select Problem Type:",
                ["Classification", "Regression"]
            )
            
            # Feature and target selection
            st.sidebar.header("Model Configuration")
            target_col = st.sidebar.selectbox(
                "Select Target Variable:",
                df.columns
            )
            
            feature_cols = st.sidebar.multiselect(
                "Select Features:",
                [col for col in df.columns if col != target_col],
                default=[col for col in df.columns if col != target_col]
            )
            
            # Model selection based on problem type
            if problem_type == "Classification":
                models = {
                    "Random Forest": RandomForestClassifier(),
                    "Logistic Regression": LogisticRegression(),
                    "Support Vector Machine": SVC()
                }
            else:
                models = {
                    "Random Forest": RandomForestRegressor(),
                    "Linear Regression": LinearRegression(),
                    "Support Vector Regression": SVR()
                }
            
            selected_models = st.sidebar.multiselect(
                "Select Models to Train:",
                list(models.keys()),
                default=list(models.keys())[0]
            )
            
            # Training parameters
            st.sidebar.header("Training Parameters")
            test_size = st.sidebar.slider("Test Size Ratio:", 0.1, 0.5, 0.2, 0.05)
            random_state = st.sidebar.number_input("Random State:", 42)
            cv_folds = st.sidebar.slider("Cross-Validation Folds:", 3, 10, 5)
            
            if st.sidebar.button("🚀 Train Models"):
                if not feature_cols:
                    st.error("Please select at least one feature!")
                    return
                
                # Prepare data
                X = df[feature_cols]
                y = df[target_col]
                
                # Handle categorical variables
                X = pd.get_dummies(X)
                
                # Split data
                X_train, X_test, y_train, y_test = train_test_split(
                    X, y, test_size=test_size, random_state=random_state
                )
                
                # Train and evaluate models
                results = {}
                
                for model_name in selected_models:
                    with st.spinner(f"Training {model_name}..."):
                        model = models[model_name]
                        
                        # Train model
                        model.fit(X_train, y_train)
                        
                        # Evaluate model
                        if problem_type == "Classification":
                            accuracy, report, cm, y_pred = evaluate_model(
                                model, X_test, y_test, "classification"
                            )
                            
                            # Cross-validation
                            cv_scores = cross_val_score(model, X, y, cv=cv_folds)
                            
                            results[model_name] = {
                                'type': 'classification',
                                'accuracy': accuracy,
                                'report': report,
                                'cm': cm,
                                'cv_scores': cv_scores,
                                'model': model
                            }
                        else:
                            mse, r2, y_pred = evaluate_model(
                                model, X_test, y_test, "regression"
                            )
                            
                            cv_scores = cross_val_score(model, X, y, cv=cv_folds, 
                                                      scoring='r2')
                            
                            results[model_name] = {
                                'type': 'regression',
                                'mse': mse,
                                'r2': r2,
                                'cv_scores': cv_scores,
                                'model': model,
                                'predictions': y_pred
                            }
                
                # Display results
                st.header("📊 Model Results")
                
                # Metrics comparison
                if problem_type == "Classification":
                    metrics_df = pd.DataFrame({
                        'Model': list(results.keys()),
                        'Accuracy': [results[model]['accuracy'] for model in results],
                        'CV Mean Score': [results[model]['cv_scores'].mean() for model in results],
                        'CV Std': [results[model]['cv_scores'].std() for model in results]
                    })
                else:
                    metrics_df = pd.DataFrame({
                        'Model': list(results.keys()),
                        'R² Score': [results[model]['r2'] for model in results],
                        'MSE': [results[model]['mse'] for model in results],
                        'CV Mean Score': [results[model]['cv_scores'].mean() for model in results],
                        'CV Std': [results[model]['cv_scores'].std() for model in results]
                    })
                
                st.subheader("Model Comparison")
                st.dataframe(metrics_df.style.highlight_max(axis=0), 
                           use_container_width=True)
                
                # Detailed results for each model
                for model_name, result in results.items():
                    with st.expander(f"Detailed Results - {model_name}"):
                        col1, col2 = st.columns(2)
                        
                        with col1:
                            st.metric("Accuracy" if problem_type == "classification" else "R² Score", 
                                    f"{result['accuracy' if problem_type == 'classification' else 'r2']:.3f}")
                            st.metric("CV Mean Score", f"{result['cv_scores'].mean():.3f}")
                        
                        with col2:
                            st.metric("CV Std", f"{result['cv_scores'].std():.3f}")
                            if problem_type == "regression":
                                st.metric("MSE", f"{result['mse']:.3f}")
                        
                        # Visualizations
                        if problem_type == "classification":
                            # Confusion matrix
                            fig_cm = px.imshow(result['cm'], 
                                             title=f"Confusion Matrix - {model_name}",
                                             labels=dict(x="Predicted", y="Actual"),
                                             color_continuous_scale="Blues")
                            st.plotly_chart(fig_cm, use_container_width=True)
                        else:
                            # Regression plot
                            fig_reg = px.scatter(x=y_test, y=result['predictions'],
                                              title=f"Actual vs Predicted - {model_name}",
                                              labels={'x': 'Actual', 'y': 'Predicted'})
                            fig_reg.add_shape(type='line', line=dict(dash='dash'),
                                            x0=y_test.min(), y0=y_test.min(),
                                            x1=y_test.max(), y1=y_test.max())
                            st.plotly_chart(fig_reg, use_container_width=True)
                
                # Model saving
                st.header("💾 Model Export")
                best_model_name = metrics_df.iloc[metrics_df['Accuracy' if problem_type == 'classification' else 'R² Score'].idxmax()]['Model']
                best_model = results[best_model_name]['model']
                
                # Save model
                model_buffer = io.BytesIO()
                joblib.dump(best_model, model_buffer)
                model_buffer.seek(0)
                
                st.download_button(
                    label=f"Download Best Model ({best_model_name})",
                    data=model_buffer,
                    file_name=f"best_model_{best_model_name}.pkl",
                    mime="application/octet-stream"
                )
    
    else:
        st.info("👆 Please upload your dataset to begin model training")
        st.markdown("""
        ### Supported Features:
        - **Classification & Regression**: Multiple algorithm support
        - **Model Comparison**: Side-by-side performance comparison
        - **Cross-Validation**: Robust model evaluation
        - **Visual Analytics**: Confusion matrices, regression plots
        - **Model Export**: Download trained models
        - **Hyperparameter Tuning**: Basic parameter configuration
        """)

if __name__ == "__main__":
    main()