import streamlit as st
import numpy as np
import matplotlib.pyplot as plt
from sklearn.neighbors import KNeighborsClassifier
from sklearn.metrics import accuracy_score,f1_score
from sklearn.model_selection import train_test_split
from sklearn.datasets import make_classification, make_circles, make_blobs, make_moons
from mlxtend.plotting import plot_decision_regions, plot_learning_curves

# Streamlit UI
st.set_page_config(page_title="KNN Classifier App", layout="wide")
st.title("🔍 KNN Classifier Interactive App")
st.write("Select a dataset, modify parameters, and tune KNN hyperparameters.")

# Sidebar options
dataset_choice = st.sidebar.selectbox("📂 Select a Dataset", ["Classification", "Circles", "Blobs", "Moons"])

# Dataset parameters
if dataset_choice == "Classification":
    n_samples = st.sidebar.slider("Samples", 1000, 5000, 2000)
    class_sep = st.sidebar.slider("Class Separation", 0.5, 5.0, 1.0)
    X, y = make_classification(n_samples=n_samples, n_features=2, n_redundant=0, 
                               n_clusters_per_class=1, class_sep=class_sep, random_state=23)

elif dataset_choice == "Circles":
    n_samples = st.sidebar.slider("Samples", 500, 5000, 2000)
    noise = st.sidebar.slider("Noise", 0.0, 0.5, 0.2)
    factor = st.sidebar.slider("Factor", 0.1, 0.9, 0.2)
    X, y = make_circles(n_samples=n_samples, factor=factor, noise=noise, random_state=23)

elif dataset_choice == "Blobs":
    n_samples = st.sidebar.slider("Samples", 500, 5000, 2000)
    clusters = st.sidebar.slider("Clusters", 2, 5, 3)
    X, y = make_blobs(n_samples=n_samples, centers=clusters, n_features=2, random_state=23)

else:  # Moons
    n_samples = st.sidebar.slider("Samples", 500, 5000, 2000)
    noise = st.sidebar.slider("Noise", 0.0, 0.5, 0.2)
    X, y = make_moons(n_samples=n_samples, noise=noise, random_state=23)

# Split dataset
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=23)

# KNN Parameters
st.sidebar.subheader("⚙️ KNN Parameters")
n_neighbors = st.sidebar.slider("Neighbors (k)", 1, 15, 3)
weights = st.sidebar.selectbox("Weights", ["uniform", "distance"])
p = st.sidebar.selectbox("p (Minkowski)", [1, 2])

# Train KNN
knn = KNeighborsClassifier(n_neighbors=n_neighbors, weights=weights, p=p, metric='minkowski')
knn.fit(X_train, y_train)
y_pred = knn.predict(X_test)
accuracy = accuracy_score(y_test, y_pred)
f1=f1_score(y_test,y_pred,average="weighted")

# Display results
st.subheader("📊 Model Accuracy")
st.write(f"**Accuracy Score: {accuracy:.4f}**")

st.subheader("📊 Model F1-Score")
st.write(f"**F1 Score: {f1:.4f}**")

fig, ax = plt.subplots(figsize=(6, 4))
plot_learning_curves(X_train, y_train, X_test, y_test, knn, scoring='accuracy')
st.subheader("📈 Learning Curve")
st.pyplot(fig)

fig, ax = plt.subplots(figsize=(6, 4))
knn.fit(X, y)
plot_decision_regions(X, y, knn)
st.subheader("🔷 Decision Boundary")
st.pyplot(fig)