import warnings; warnings.filterwarnings("ignore")  # Turn off warnings to keep our notebook clean.

# Let us make sure pandas and matplotlib are available for our data exploration.
import pandas as pd
import matplotlib.pyplot as plt

# Data setup
# We will use a real dataset: Daily minimum temperatures in Melbourne, Australia.
url = "https://raw.githubusercontent.com/jbrownlee/Datasets/master/daily-min-temperatures.csv"

temps = pd.read_csv(url)
print("Shape of the data:", temps.shape)
temps.head()

# Let us look at a simple line plot of the temperature values over time.
plt.figure(figsize=(10,4))
plt.plot(pd.to_datetime(temps['Date']), temps['Temp'])
plt.title('Daily Minimum Temperatures Over Time')
plt.xlabel('Date')
plt.ylabel('Temperature (Celsius)')
plt.show()

# Let us look at basic dataset info: types, missing values, quick summary.
temps.info()

print("\nMissing values per column:")
print(temps.isnull().sum())

print("\nSummary stats:")
print(temps.describe())

# Convert the 'Date' to a datetime object, so pandas treats it as a timeline.
temps['Date'] = pd.to_datetime(temps['Date'])
temps = temps.set_index('Date')

# Look up the temperature on January 3, 1990.
date_to_find = "1990-01-03"
if date_to_find in temps.index:
    print("Temperature on", date_to_find, ":", temps.loc[date_to_find, 'Temp'], "Celsius")
else:
    print("Date not found!")
    

# What happens if we look for a date not in the data? Let us try.
try:
    print("Temperature on 2040-01-01:", temps.loc['2040-01-01', 'Temp'])
except KeyError:
    print("Date not found - be careful when picking your date!")
    

# Slice a few days to see a small time window.
sample_days = temps.loc['1990-01-01':'1990-01-05']
print(sample_days)

# Fixing a single bad value. Suppose January 2, 1990's temperature was recorded wrong.
old_value = temps.loc['1990-01-02', 'Temp']
print("Before fix: ", old_value)
temps.loc['1990-01-02', 'Temp'] = 15.0
print("After fix: ", temps.loc['1990-01-02', 'Temp'])

# Add a new date at the end of the data.
from datetime import timedelta
last_date = temps.index[-1]
next_date = last_date + timedelta(days=1)
temps.loc[next_date] = 10.0
print(f"Added {next_date.date()} with temperature 10.0.")

# Remove a day: maybe something was entered by mistake.
temps = temps.drop(next_date)
print(f"Removed {next_date.date()} from the dataset.")

# Calculate average temperature for a month.
monthly_mean = temps['Temp'].resample('M').mean()
print(monthly_mean.head())

# Plot monthly average temperatures for a better visual trend.
plt.figure(figsize=(10,4))
plt.plot(monthly_mean.index, monthly_mean.values, marker='o')
plt.title('Monthly Average Temperatures')
plt.xlabel('Month')
plt.ylabel('Avg Temperature (Celsius)')
plt.show()

# Filter for hot days: show all days above 20C.
hot_days = temps[temps['Temp'] > 20]
print(hot_days.head())

# Find coldest temperature and the date it happened.
min_temp = temps['Temp'].min()
coldest_day = temps['Temp'].idxmin()
print("Coldest day:", coldest_day.date(), "with", min_temp, "Celsius")

# Time series split: training vs. test data for modeling.
total_days = temps.shape[0]
train_size = int(total_days * 0.8)
train = temps.iloc[:train_size]
test = temps.iloc[train_size:]
print("Train shape:", train.shape, ", Test shape:", test.shape)

# Mini-project part 1: summarize the hottest week in the dataset.
weekly_mean = temps['Temp'].resample('W').mean()
hottest_week = weekly_mean.idxmax()
print("Hottest week starts on:", hottest_week.date())

# Show all days in that week.
hot_week_days = temps.loc[hottest_week : hottest_week + pd.Timedelta(days=6)]
print(hot_week_days)

# Mini-project part 2: plot the hottest week as a zoomed-in line chart.
plt.figure(figsize=(8,3))
plt.plot(hot_week_days.index, hot_week_days['Temp'], marker='s', color='red')
plt.title('Temperatures During the Hottest Week')
plt.xlabel('Date')
plt.ylabel('Temperature (Celsius)')
plt.grid()
plt.show()

# Troubleshooting: check for gaps in time (missing dates).
all_days = pd.date_range(start=temps.index.min(), end=temps.index.max(), freq='D')
missing_dates = all_days.difference(temps.index)
if len(missing_dates) > 0:
    print('Missing dates in data:')
    print(missing_dates)
else:
    print('No missing dates  our time series is complete!')
    

# Extra tip: convert to yearly averages for very long-term trends.
yearly_mean = temps['Temp'].resample('A').mean()
plt.figure(figsize=(8,3))
plt.plot(yearly_mean.index.year, yearly_mean.values, '-o', color='green')
plt.title('Yearly Average Temperatures')
plt.xlabel('Year')
plt.ylabel('Avg Temp (Celsius)')
plt.show()

# Challenge: input a year and print its monthly average temperatures.
year = input("Enter a year (e.g., 1990): ")
year = int(year)
subset = temps.loc[temps.index.year == year]
monthly_avg = subset['Temp'].resample('M').mean()
print(f"Monthly averages for {year}:")
print(monthly_avg)