import streamlit as st
import pandas as pd
import numpy as np
from sklearn.datasets import load_iris
from sklearn.model_selection import train_test_split
from sklearn.tree import DecisionTreeClassifier, plot_tree
from sklearn.preprocessing import StandardScaler
from sklearn.metrics import classification_report, accuracy_score
import matplotlib.pyplot as plt
import seaborn as sns

st.set_page_config(page_title="Explore Decision Tree Algorithm", layout="wide")
st.title("🌳 Decision Tree Classifier Explained")

# Tabs for better UX
tab1, tab2, tab3 = st.tabs(["📘 About Decision Tree", "🧪 Try It Out", "📈 Visualization"])

with tab1:
    st.markdown("""
    ## 🌳 What is a Decision Tree?

    A **Decision Tree** is a flowchart-like structure used in supervised learning for both **classification** and **regression** tasks. 
    It splits the dataset into smaller and smaller subsets while at the same time an associated decision tree is incrementally developed.

    ### 🧠 Core Concepts:

    - **Root Node**: The first decision point (split) based on the most important feature.
    - **Internal Nodes**: Decision points that split data based on feature values.
    - **Leaf Nodes**: The final output class (for classification) or value (for regression).
    - **Branches**: Represent outcomes of a decision or test.

    ---
    ### 📊 How it Works:
    1. The algorithm chooses a feature and a split point that best separates the data.
    2. Splits are made recursively until a stopping criterion is met (like max depth or pure leaves).
    3. At each node, it chooses the split that **maximizes information gain** or **minimizes impurity**.

    ---
    ### ⚖️ Splitting Criteria:

    - **Gini Impurity**:
        - Measures how often a randomly chosen element would be incorrectly labeled.
        - Best when you want a fast computation.
        - Formula: $Gini(t) = 1 - \\sum p(i)^2$

    - **Entropy (Information Gain)**:
        - Measures the level of disorder or uncertainty.
        - Based on the concept from information theory.
        - Formula: $Entropy(t) = - \\sum p(i) \\log_2(p(i))$

    ---

    ### 🔍 Visual Intuition:

    Imagine a dataset of flowers. You might first split by **petal length**. One branch contains flowers with short petals, another with long petals. 
    Then, each of those groups is further split based on **petal width**, and so on — until the flowers are grouped into species.

    ---
    ### 📦 Use Cases:

    - Medical Diagnosis (e.g., "Is this tumor malignant or benign?")
    - Loan Approval ("Should this applicant get a loan?")
    - Customer Churn Prediction
    - Email Spam Detection
    - Any tabular data with clear patterns and labeled outcomes

    ---
    ### ✅ Advantages:

    - Simple to understand and interpret (like flowcharts)
    - Requires little data preprocessing (no scaling or normalization)
    - Can handle both numerical and categorical data
    - Easily visualized

    ### ⚠️ Limitations:

    - Prone to overfitting if not pruned or constrained
    - Small changes in data can lead to very different trees
    - Not good with very large datasets or high-dimensional data alone (better with ensembles like Random Forest)

    ---
    🎯 **Tip**: Use pruning, set a maximum depth, or use ensemble methods (like Random Forest or Gradient Boosted Trees) to enhance performance and reduce overfitting.

    """)

with tab2:
    st.subheader("🌸 Train a Decision Tree on the Iris Dataset")
    iris = load_iris()
    df = pd.DataFrame(iris.data, columns=iris.feature_names)
    df['target'] = iris.target
    st.dataframe(df.head(), use_container_width=True)

    criterion = st.radio("Select Splitting Criterion", ["gini", "entropy"])
    max_depth = st.slider("Select Max Depth", 1, 10, value=3)

    X = df.drop('target', axis=1)
    y = df['target']

    X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)

    model = DecisionTreeClassifier(criterion=criterion, max_depth=max_depth, random_state=42)
    model.fit(X_train, y_train)
    y_pred = model.predict(X_test)

    acc = accuracy_score(y_test, y_pred)
    st.success(f"✅ Model Accuracy: {acc*100:.2f}%")
    
    st.markdown("### 📊 Classification Report")
    st.text(classification_report(y_test, y_pred, target_names=iris.target_names))

with tab3:
    st.subheader("📉 Visualize the Decision Tree")
    
    fig, ax = plt.subplots(figsize=(12, 6))
    plot_tree(model, feature_names=iris.feature_names, class_names=iris.target_names, filled=True)
    st.pyplot(fig)

    st.subheader("📌 Feature Importance")
    importance_df = pd.DataFrame({
        'Feature': iris.feature_names,
        'Importance': model.feature_importances_
    }).sort_values(by="Importance", ascending=False)

    st.dataframe(importance_df, use_container_width=True)

    fig2, ax2 = plt.subplots()
    sns.barplot(x='Importance', y='Feature', data=importance_df, palette='viridis')
    st.pyplot(fig2)