import warnings
warnings.filterwarnings("ignore")

# Let us set up the Flights dataset, which is great for learning time handling.
import seaborn as sns
import numpy as np
import pandas as pd
np.random.seed(42)
df = sns.load_dataset('flights')
print(df.shape)
print(df.head(3))

print(df.columns)
print(df.dtypes)

# Creating a single column with both year and month as a string.
df['date_str'] = df['year'].astype(str) + '-' + df['month'].astype(str)
print(df[['year', 'month', 'date_str']].head())

# Try to sort by date_str.
print(df.sort_values('date_str').head(3))

df['date'] = pd.to_datetime(df['date_str'], format='%Y-%b')
print(df[['date_str', 'date']].head())

# Check the type of the new column.
print(type(df.loc[0, 'date']))

df['year_extracted'] = df['date'].dt.year
df['month_number'] = df['date'].dt.month
df['weekday'] = df['date'].dt.weekday
print(df[['date', 'year_extracted', 'month_number', 'weekday']].head())

# You can get the weekday as the name, too.
df['weekday_name'] = df['date'].dt.day_name()
print(df[['date', 'weekday_name']].head())

# Challenge: What date had the most passengers?
max_idx = df['passengers'].idxmax()
print(df.loc[max_idx, ['date', 'passengers']])

# Example: Count total passengers by weekday name.
weekday_totals = df.groupby('weekday_name')['passengers'].sum().sort_values(ascending=False)
print(weekday_totals)

# You can also filter the DataFrame for weekends.
weekends = df[df['weekday_name'].isin(['Saturday', 'Sunday'])]
print(weekends.head(3))

# Visualize passengers over time using the datetime index.
import matplotlib.pyplot as plt
plt.figure(figsize=(10, 4))
plt.plot(df['date'], df['passengers'], marker='o')
plt.title('Monthly Airline Passengers Over Time')
plt.xlabel('Date')
plt.ylabel('Passengers')
plt.grid(True)
plt.tight_layout()
plt.show()

# Try: Select all months in 1952.
in_1952 = df[df['date'].dt.year == 1952]
print(in_1952.head())

# Next: What is the average number of passengers in July flights?
july_avg = df[df['month'] == 'July']['passengers'].mean()
print(f"Average passengers in July: {july_avg:.2f}")

# Sometimes real datasets have multiple datetime columns or tricky formats.
example_dates = pd.Series(['2022/03/01', '01-04-2023', '2023.05.01'])
parsed = pd.to_datetime(example_dates, dayfirst=False, errors='coerce')
print(pd.DataFrame({'original': example_dates, 'parsed': parsed}))

# Slicing and resampling: What was the total number of passengers in 1950?
total_1950 = df[df['date'].dt.year == 1950]['passengers'].sum()
print(f"Total passengers in 1950: {total_1950}")

# Using input to enter your own date string to convert. Try entering '2024-05-15' (YYYY-MM-DD).
user_date_str = input("Type a date (YYYY-MM-DD): ")
user_date = pd.to_datetime(user_date_str)
print("You entered:", user_date)
print("Year:", user_date.year, "Month:", user_date.month, "Day:", user_date.day)