# Setup: Turn off unnecessary warnings to keep things tidy
import warnings
warnings.filterwarnings("ignore")

# Data setup: Let us get some real sales data
import pandas as pd

url = "https://raw.githubusercontent.com/jbrownlee/Datasets/master/shampoo.csv"
data = pd.read_csv(url)

print("Data shape:", data.shape)
print(data.head())

# Let us plot the sales data to visualize patterns
import matplotlib.pyplot as plt

plt.figure(figsize=(8,4))
plt.plot(data["Sales"], label="Monthly Sales")
plt.title("Shampoo Sales Over Time")
plt.xlabel("Month Number")
plt.ylabel("Sales Volume")
plt.legend()
plt.show()

# Prepare the data for forecasting: use numbers instead of dates
import numpy as np

X = np.arange(len(data)).reshape(-1, 1)
y = data["Sales"].values

print("X shape:", X.shape)
print("y shape:", y.shape)

# Split our data into training and testing sets
from sklearn.model_selection import train_test_split

X_train, X_test, y_train, y_test = train_test_split(
    X, y, test_size=0.25, random_state=42
)

print("Training size:", X_train.shape[0])
print("Testing size:", X_test.shape[0])

# Build and train a basic linear regression model
from sklearn.linear_model import LinearRegression

lr = LinearRegression()
lr.fit(X_train, y_train)

print("Model learned.")

# Use the model to predict sales for test months
y_pred = lr.predict(X_test)

print("First five real sales:", y_test[:5])
print("First five predicted sales:", y_pred[:5])

# Plot the actual vs predicted sales
plt.figure(figsize=(8,4))
plt.scatter(X_test, y_test, color="blue", label="Real Sales")
plt.scatter(X_test, y_pred, color="red", label="Predicted Sales")
plt.xlabel("Month Number")
plt.ylabel("Sales Volume")
plt.title("Real vs Predicted Sales")
plt.legend()
plt.show()

# Check how well our model did: mean squared error
from sklearn.metrics import mean_squared_error

mse = mean_squared_error(y_test, y_pred)
print(f"Mean Squared Error: {mse:.2f}")

# Use Ridge regression (regularized)
from sklearn.linear_model import Ridge

ridge = Ridge(alpha=1.0)
ridge.fit(X_train, y_train)
y_ridge = ridge.predict(X_test)
mse_ridge = mean_squared_error(y_test, y_ridge)

print(f"Ridge MSE: {mse_ridge:.2f}")

# Try Lasso regression for comparison
from sklearn.linear_model import Lasso

lasso = Lasso(alpha=0.5)
lasso.fit(X_train, y_train)
y_lasso = lasso.predict(X_test)
mse_lasso = mean_squared_error(y_test, y_lasso)

print(f"Lasso MSE: {mse_lasso:.2f}")

# Ask the user to change alpha values themselves
user_alpha = float(input("Type a new alpha value for Ridge regularization: "))
ridge2 = Ridge(alpha=user_alpha)
ridge2.fit(X_train, y_train)
y_ridge2 = ridge2.predict(X_test)
mse_ridge2 = mean_squared_error(y_test, y_ridge2)
print(f"With alpha = {user_alpha}, Ridge MSE: {mse_ridge2:.2f}")

# Mini-project: Forecast next month's sales using the best model
last_month = X.max() + 1
best_model = ridge if mse_ridge < mse and mse_ridge < mse_lasso else lr
next_pred = best_model.predict([[last_month]])
print(f"Predicted sales for next month: {next_pred[0]:.2f}")

# Practice: Ask the user to input a future month number
month_input = int(input("Type which future month you want to predict (e.g., 38): "))
future_pred = best_model.predict([[month_input]])
print(f"Predicted sales for month {month_input}: {future_pred[0]:.2f}")

# If you get an error: Check for missing or weird data
print("Any missing values?", data.isnull().any().any())
print("Minimum sales:", data["Sales"].min())
print("Maximum sales:", data["Sales"].max())

# Extra: Try using more than one feature (advanced)
# For now, we just use month number, but you could add weather, holidays, or promotions if you had them.
print("You can try extra features for fun in your own projects!")