# Let's start with a quick setup.
import warnings
warnings.filterwarnings("ignore")

print("Ready to get started!")

# Let's try some basic math first!
a = 9
b = 16
result = (a + b) / 2
print("Average:", result)

# Now let's use a list.
numbers = [10, 20, 30, 40, 50]
print("First number:", numbers[0])
print("Last number:", numbers[-1])

# Let's measure distance by hand with two numbers.
x1 = 7
x2 = 17
distance = abs(x1 - x2)
print("Distance:", distance)

# Now let's use two features, like height and weight.
person_A = [170, 65]
person_B = [160, 60]
diff_height = person_A[0] - person_B[0]
diff_weight = person_A[1] - person_B[1]
euclidean_distance = (diff_height ** 2 + diff_weight ** 2) ** 0.5
print("Distance between A and B:", euclidean_distance)

# Data setup
from sklearn.datasets import load_iris
import pandas as pd

iris = load_iris(as_frame=True)
df = iris['frame']
print('Shape:', df.shape)
df.head()

# Check the target distribution
df['target'].value_counts()

# Split the data into a training set and a test set
from sklearn.model_selection import train_test_split

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)
print('Training samples:', len(X_train))
print('Test samples:', len(X_test))

# Train a KNN model
from sklearn.neighbors import KNeighborsClassifier

knn = KNeighborsClassifier(n_neighbors=3)
knn.fit(X_train, y_train)

print('KNN model trained!')

# Predict on test data
y_pred = knn.predict(X_test)
print('Predicted labels:', y_pred[:10])

# Measure accuracy
from sklearn.metrics import accuracy_score

accuracy = accuracy_score(y_test, y_pred)
print("Accuracy:", accuracy)

# Try predicting a new flower using input()
values = []
feature_names = list(X.columns)
for name in feature_names:
    val = float(input(f"Enter {name}: "))
    values.append(val)
prediction = knn.predict([values])[0]
print("Predicted species:", iris['target_names'][prediction])

# What if we pick a bad value for k?
knn_bad = KNeighborsClassifier(n_neighbors=100)
knn_bad.fit(X_train, y_train)
bad_pred = knn_bad.predict(X_test)
bad_acc = accuracy_score(y_test, bad_pred)
print('Accuracy with k=100:', bad_acc)

# Find the best k for our data
scores = []
for k in range(1, 16):
    model = KNeighborsClassifier(n_neighbors=k)
    model.fit(X_train, y_train)
    y_pred = model.predict(X_test)
    scores.append(accuracy_score(y_test, y_pred))
print('K values:', list(range(1, 16)))
print('Accuracies:', scores)

# Mini project: Let's classify unknown flowers
new_flowers = [[5, 3.2, 1.1, 0.1], [6, 2.9, 4.5, 1.5], [7, 3, 6.3, 2.5]]
results = knn.predict(new_flowers)
for i, res in enumerate(results):
    print(f"Flower {i+1} is predicted as:", iris['target_names'][res])
    

# Best practices: Scale features for KNN
from sklearn.preprocessing import StandardScaler

scaler = StandardScaler()
X_train_scaled = scaler.fit_transform(X_train)
X_test_scaled = scaler.transform(X_test)

knn_scaled = KNeighborsClassifier(n_neighbors=3)
knn_scaled.fit(X_train_scaled, y_train)
scaled_acc = accuracy_score(y_test, knn_scaled.predict(X_test_scaled))
print('Accuracy (with scaling):', scaled_acc)

# Troubleshooting: What if we get an error?
try:
    test_pred = knn.predict([[5, 3]])  # only two features instead of four!
except Exception as e:
    print('Oops! Error:', str(e))
    

# Extra tip: Use kneighbors to look up the closest data points.
distances, indices = knn.kneighbors([X_test.iloc[0]])
print('Distances:', distances)
print('Indices:', indices)

# Challenge: Try changing the dataset or k value.
print('Try using Wine or Breast Cancer datasets with K N N.')
print('Or try new values for k, like 5 or 7!')