# Let us start by importing Python's built-in warning filter
import warnings
warnings.filterwarnings('ignore')

# Let us create our first variable
age = 29

# Variables can store different types of values
name = "Sam"
has_heart_disease = False

# This dictionary stores one patient's data
patient = {
    "name": "Sam",
    "age": 29,
    "has_heart_disease": False
}

# Access information from a dictionary
print("Name:", patient["name"])
print("Heart Disease:", patient["has_heart_disease"])

# Let us handle a missing key carefully
if "cholesterol" in patient:
    print(patient["cholesterol"])
else:
    print("No cholesterol data")
    

# Data setup
import pandas as pd
url = "https://raw.githubusercontent.com/Breezercoder/Heart-disease-prediction/main/heart_disease_dataset.csv"
heart_data = pd.read_csv(url)
print("Rows and columns:", heart_data.shape)
heart_data.head()

# What does the target look like?
print(heart_data["target"].value_counts())

# Let us explore some basic statistics
heart_data.describe()

# Split into train and test sets
from sklearn.model_selection import train_test_split
train, test = train_test_split(heart_data, test_size=0.2, random_state=42)
print("Train size:", len(train), "Test size:", len(test))

# Everyone predicted as healthy (target 0)
test["prediction"] = 0
accuracy = (test["prediction"] == test["target"]).mean()
print("Accuracy if we guess all are healthy:", accuracy)

# Let us try a smarter rule based on a real feature
test["prediction"] = test["age"].apply(lambda x: 1 if x > 50 else 0)
accuracy = (test["prediction"] == test["target"]).mean()
print("Accuracy guessing heart disease if age > 50:", accuracy)

# Decision tree classifier example
from sklearn.tree import DecisionTreeClassifier
features = [col for col in heart_data.columns if col != "target"]
clf = DecisionTreeClassifier(random_state=42)
clf.fit(train[features], train["target"])
test["prediction"] = clf.predict(test[features])
accuracy = (test["prediction"] == test["target"]).mean()
print("Decision tree accuracy:", accuracy)

# Try out your own patient!
user_age = int(input("Enter your age: "))
user_sex = int(input("Enter 1 for male, 0 for female: "))
user_cp = int(input("Chest pain type (0-3): "))
user_trestbps = int(input("Resting blood pressure: "))
user_chol = int(input("Serum cholesterol: "))
user_fbs = int(input("Fasting blood sugar > 120mg/dl? (1 = yes, 0 = no): "))
user_restecg = int(input("Rest ECG result (0-2): "))
user_thalach = int(input("Max heart rate achieved: "))
user_exang = int(input("Exercise induced angina (1 = yes, 0 = no): "))
user_oldpeak = float(input("ST depression: "))
user_slope = int(input("Slope of ST segment (0-2): "))
user_ca = int(input("Number of vessels colored (0-3): "))
user_thal = int(input("Thalassemia (1 = normal, 2 = fixed defect, 3 = reversible defect): "))
user_row = [[user_age, user_sex, user_cp, user_trestbps, user_chol, user_fbs, user_restecg, user_thalach, user_exang, user_oldpeak, user_slope, user_ca, user_thal]]
user_pred = clf.predict(user_row)[0]
if user_pred == 1:
    print("Prediction: Risk of Heart Disease")
else:
    print("Prediction: No Heart Disease Detected")
    

# Best practices: check your work
missing = heart_data.isnull().sum()
print("Missing values in each column:\n", missing)

# Common mistakes: wrong data shapes
try:
    clf.predict([[29]])
except Exception as e:
    print("Error message:", e)
    

# Challenge: Change the tree depth and see the result!
new_clf = DecisionTreeClassifier(max_depth=3, random_state=42)
new_clf.fit(train[features], train["target"])
score = new_clf.score(test[features], test["target"])
print("Tree with max_depth=3 has accuracy:", score)

# Challenge: Predict a patient using only age and cholesterol
short_features = ["age", "chol"]
clf_small = DecisionTreeClassifier(random_state=42)
clf_small.fit(train[short_features], train["target"])
small_pred = clf_small.predict(test[short_features])
print("Accuracy with just age and cholesterol:", (small_pred == test["target"]).mean())