# Suppress warnings for a cleaner output
import warnings; warnings.filterwarnings("ignore")

# Main libraries for data, plots, and evaluation
import pandas as pd
import numpy as np
import seaborn as sns
import matplotlib.pyplot as plt
from sklearn.model_selection import train_test_split
from sklearn.linear_model import LogisticRegression
from sklearn.metrics import accuracy_score, precision_score, recall_score, confusion_matrix

# Data setup: Load and peek at the Titanic dataset
titanic = sns.load_dataset("titanic")
print("Titanic shape:", titanic.shape)
titanic.head()

# Let us prepare our data: Drop rows with missing age or embark_town
titanic_clean = titanic.dropna(subset=["age", "embark_town"]).copy()
print("Rows after cleaning:", titanic_clean.shape[0])
titanic_clean.head(3)

# Choose features and the target (who survived)
X = titanic_clean[["pclass", "sex", "age", "fare"]]
y = titanic_clean["survived"]
print("X shape:", X.shape)
X.head(2)

# Convert 'sex' to numeric: male=1, female=0
X["sex"] = X["sex"].map({"male":1, "female":0})
X.head(2)

# Split data: 70% train, 30% test. Always use random_state for reproducibility.
X_train, X_test, y_train, y_test = train_test_split(
    X, y, test_size=0.3, random_state=42
)
print("Train size:", X_train.shape[0], "Test size:", X_test.shape[0])

# Train a logistic regression model
model = LogisticRegression(max_iter=1000)
model.fit(X_train, y_train)
print("Ready to predict!")

# Predict on the test data
y_pred = model.predict(X_test)
print("First 10 predicted survival values:", y_pred[:10])

# Calculate accuracy: What percent of test labels did we get right?
accuracy = accuracy_score(y_test, y_pred)
print("Accuracy:", round(accuracy, 3))

# What about precision and recall?
precision = precision_score(y_test, y_pred)
recall = recall_score(y_test, y_pred)
print("Precision:", round(precision, 3))
print("Recall:", round(recall, 3))

# Show the confusion matrix
cm = confusion_matrix(y_test, y_pred)
print("Confusion matrix:\n", cm)
sns.heatmap(cm, annot=True, fmt="d", cmap="Blues")
plt.xlabel("Predicted")
plt.ylabel("Actual")
plt.title("Confusion Matrix")
plt.show()

# Tune the threshold: Use 0.3 instead of 0.5
y_prob = model.predict_proba(X_test)[:,1]
custom_thresh = 0.3
y_pred_custom = (y_prob > custom_thresh).astype(int)
print("First 10 predictions (threshold 0.3):", y_pred_custom[:10])

# Precision and recall at threshold 0.3
precision_c = precision_score(y_test, y_pred_custom)
recall_c = recall_score(y_test, y_pred_custom)
print("Precision (0.3):", round(precision_c, 3))
print("Recall (0.3):", round(recall_c, 3))

# Your turn! Try your own threshold
answer = input("Type a threshold between 0 and 1 (for example, 0.6): ")
try:
    user_thresh = float(answer)
    y_pred_user = (y_prob > user_thresh).astype(int)
    print("Precision:", round(precision_score(y_test, y_pred_user), 3))
    print("Recall:", round(recall_score(y_test, y_pred_user), 3))
except:
    print("Please enter a valid number.")