# print() function
print('Python is great!')

# Remember: run cells using play button or Shift-Return / Shift-Enter

# Creating variables
day_of_month = 24         # int (integer)
print(day_of_month)
hour_of_day = 9 + 45/60   # float (floating-point, or decimal, number)
print(hour_of_day)
month = 'January'         # string (can use single or double quotes)
print(month)
date = 'January' + ' ' + '24th'   # multiple strings concatenated together
print(date)
cold = True               # boolean (True or False)
print(cold)

# Remember: write variable names that are meaningful

# Note that variables persist in Colab's memory across cells

# Modifying variables
day_of_month = day_of_month + 1    # adds 1 to the existing value
print(day_of_month)
day_of_month += 1                  # a more compact way of writing the line above
print(day_of_month)

# Another useful type of variable: datetime object
from datetime import datetime     # first, import the package
now = datetime(2023,1,24,9,45,0)  # format: (year, month, day, hour, minute, second)
print(now)
print(now.month)   # extract month from datetime

# An unfortunately common object: NaN (not-a-number)
import numpy as np
today_data = np.NaN
print(today_data)
print(np.isnan(today_data))

# Creating lists
months = ['Jan','Feb','Mar','Apr','May','Jun','Jul','Aug','Sep','Oct','Nov','Dec']
mon_nums = [1,2,3,4,5,6,7,8,9,10,11,12]
print(months)
print(mon_nums)

# Check length of a list using len()
print(len(months))

# Remember: lists can have any type of object in them, including different types of objects
# (e.g., numbers and strings, and even lists inside the list)

# List indexing
print(months[0])       # the 1st element
print(months[11])      # the 12th element
print(months[-2])      # the second-to-last element
print(months[0:3])     # the first 3 elements
print(months[5::2])    # every 2nd element starting at the 6th element (Jun)

# Modifying a list
months[11] = 'December'   # change the 11th element to 'December'
print(months)

# Append to a list
months.append('January of the next year')
print(months)

# Which months in Seattle have comfortable weather?
# Let's Google "Seattle temperature by month" and transcribe the table of high temps
# ASK: try to write a loop to tell us which months in Seattle are at least 66°F

months = ['Jan','Feb','Mar','Apr','May','Jun','Jul','Aug','Sep','Oct','Nov','Dec']
temp = [47,49,52,57,63,66,72,72,67,59,51,46]

# Let's start by reviewing for-loops
for mon_idx in range(12):
  print(mon_idx, months[mon_idx])

# Here we add an if-statement to check each month's temp against the threshold
temp_threshold = 66
for mon_idx in range(len(months)):
  if temp[mon_idx] >= temp_threshold:
    print('This month is comfortable: ' + months[mon_idx])

# What about a loop to check which month is exactly 57°F?
for mon_idx in range(len(months)):
  if temp[mon_idx] == 57:
    print("This is the month we're looking for: " + months[mon_idx])

# Add an else statement for months that don't meet the criteria
for mon_idx in range(len(months)):
  if temp[mon_idx] == 57:
    print("This is the month we're looking for: " + months[mon_idx])
  else:
    print("This is not the month we're looking for: " + months[mon_idx])

# Now let's convert the Seattle temperatures from Fahrenheit to Celsius
# Formula: subtract 32 and multiple by 5/9

# This won't work, because you can't do math with lists
# new_temp = (temp - 32) * (5/9)

# Instead, we have to import NumPy and use NumPy arrays
import numpy as np
example_array = np.array([1,2,3,4,5,6])  # array() is a function in the np package
print(4 * example_array)
example_list = [1,2,3,4,5,6]             # for comparison
print(4 * example_list)

# Now the math will apply to each NumPy array element
temp = np.array(temp)           # convert temp from list to array
new_temp = (temp - 32) * (5/9)
print(new_temp)

# ASK: try to do the same temp threshold exercise, but WITHOUT a loop
# (Hint: use boolean indexing!)

months = np.array(months)                  # convert from list to array

comfy_temp_mask = temp >= temp_threshold   # this creates a Boolean mask
print(comfy_temp_mask)

print(months[comfy_temp_mask])             # index into months using the Boolean mask

# Some common NumPy functions you should know...

print('The average temp is:',np.mean(temp))
print('The median temp is:',np.median(temp))
print('The highest temp is:',np.max(temp))
print('The lowest temp is:',np.min(temp))
print('The sum of all the temps is:',np.sum(temp))
print('The standard deviation of all the temps is:',np.std(temp))

from google.colab import drive, auth
drive.mount('/content/drive',force_remount=True)

import numpy as np
import pandas as pd
from datetime import datetime

filepath = '/content/sst0n170w_mon.txt'

# Load data into the variable "data"
# See the API link above for function argument info
data = pd.read_csv(filepath,delimiter=' ',index_col=0,parse_dates=['Date'])
display(data)

# Select the index column (dates)
print(data.index)                # display all items
print(data.index.values[:10])    # display the first 10 elements

# Note: if you ever need to convert between Datetime and datetime64 objects,
#       use this chart: https://stackoverflow.com/a/21916253

# Select one column of data using its label
data['SST']

# Extract just the values as a NumPy array
data['SST'].values[:10]   # just the first 10 elements

# Select a single row
data.loc['2010-01-16']

# Select multiple rows
data.loc['2010-01-16':'2010-04-16']

# Select one year of data
data.loc['2010']

# Select a single value
data.loc['2010-01-16','SST']   # use: .loc[row label, column label]

# You can also select values using NumPy-style integer indexing
data.iloc[99,1]    # the 100th row and 2nd column

# Apply NumPy functions such as mean() directly to the Pandas object
print(data['SST'].mean())    # average of all SST data
print()
print(data['SST'].loc['2010'].mean())  # average of all SST data from 2010
print()
print(data.mean(axis=0))     # average of all rows

# If needed, you can convert from Pandas into a NumPy 2D array
data_as_array = data.values
print(data_as_array)

# When indexing into NumPy 2D arrays, use [row, column] order
print('Row #1, Column #2 =',data_as_array[0,1])

# Check shape of 2D array
print('Array shape in (rows, columns):',data_as_array.shape)

# With a Pandas Series, plotting is as simple as specifying a column of data
plt.plot(data['SST'])

# But generally, you should specify the (x, y) data
plt.plot(data.index.values,data['SST'].values)

# Now let's create a nicely formatted plot and save it
plt.figure(figsize=(10,4))
plt.scatter(data.index.values,data['SST'].values,
            c='k',s=5)
plt.plot(data.index.values,data['SST'].values,
         c='k',lw=1.0,label='SST')
plt.xlabel('Year')
plt.ylabel('SST (°C)')
plt.title('Sea surface temperatures from TAO mooring at 0°N, 170°W')
plt.grid(alpha=0.5)
plt.legend()
plt.savefig('/content/sst_tao_monthly.pdf')

# Add a second line to the plot
# You can download Niño 3.4 index data here: https://tinyurl.com/OCEAN443-data-nino
nino = pd.read_csv('/content/nino34.csv',index_col=0,parse_dates=True)

plt.figure(figsize=(10,4))
plt.scatter(data.index.values,data['SST'].values,
            c='k',s=5)
plt.plot(data.index.values,data['SST'].values,
         c='k',lw=1.0,label='SST')
plt.plot(nino.index,nino.values,c='r',lw=1.0,label='El Niño index')
plt.xlim([datetime(1990,1,1),datetime(2023,1,1)])
plt.xlabel('Year')
plt.ylabel('SST (°C)')
plt.title('Sea surface temperatures from TAO mooring at 0°N, 170°W')
plt.grid(alpha=0.5)
plt.legend()
plt.savefig('/content/sst_tao_monthly_with_nino.pdf')

nino

import numpy as np
import xarray as xr
from datetime import datetime

filepath = '/content/sst.mon.ltm.1991-2020.nc'

# Load data into the variable "data"
# See the API link above for function argument info
data = xr.open_dataset(filepath)
display(data)

# Select a single variable from the netCDF dataset
data['sst']

# The metadata attributes are a dict (dictionary) that can be accessed using:
display(data['sst'].attrs)
print()
print(data['sst'].attrs['units'])   # select just units

# To subset the data, use .sel() or .isel() for selection by label or by indices
data['sst'].sel(lat=slice(-10,10),lon=179.625,time='0001-06')

# Note that the line above uses slice(lower,upper) to select a range of values

data['sst'].isel(time=5)

# Similar to Pandas, you can apply NumPy functions such as mean() to Xarray objects

# Calculate the average SST in June at 179.625°E between 10°S-10°N
display(data['sst'].sel(lat=slice(-10,10),lon=179.625,time='0001-06').mean())

# Extract the value from Xarray into NumPy
data['sst'].sel(lat=slice(-10,10),lon=179.625,time='0001-06').mean().values

# Load and plot CSV data in just 2 lines!
data = pd.read_csv('/content/sst0n170w_mon.txt',delimiter=' ',index_col=0,parse_dates=['Date'])
data['SST'].plot()

# Load and plot netCDF data in just 2 lines!
data = xr.open_dataset('/content/sst.mon.ltm.1991-2020.nc')
data['sst'].sel(time='0001-01').plot()

# Create a pseudocolor plot using Matplotlib

# First, subset data to the region and time of interest
data_subset = data['sst'].sel(lat=slice(-25,25),lon=slice(120,280),time='0001-03')

# "Squeeze" out the non-existent time dimension
data_subset = data_subset.squeeze()

plt.figure(figsize=(10,4))
plt.pcolormesh(data_subset['lon'],data_subset['lat'],data_subset,
               cmap='RdBu_r')
plt.colorbar(label='SST (°C)')
plt.xlabel('Longitude (°E)')
plt.ylabel('Latitude (°N)')
plt.title('OISST climatology for March');

# Add scatter points for Honolulu and Fiji
plt.scatter(180+180-157.9,21.3,marker='o',s=100,color='g',label='Honolulu')
plt.scatter(178.1,-17.7,marker='o',s=100,color='gold',label='Fiji')
plt.legend(loc='lower right');

fig = plt.figure(figsize=(14,4))
ax = plt.axes(projection=ccrs.PlateCarree(central_longitude=180.0))

pcm = ax.pcolormesh(data_subset['lon'],data_subset['lat'],data_subset,
                    transform=ccrs.PlateCarree())
contours = ax.contour(data_subset['lon'],data_subset['lat'],data_subset,
                      levels=12,colors='w',transform=ccrs.PlateCarree())
ax.clabel(contours,levels=contours.levels[::2],colors='w',fontsize=8)
c = plt.colorbar(pcm,ax=ax,label='SST (°C)')

ax.add_feature(cfeature.LAND,color='k',alpha=0.8)
gl = ax.gridlines(crs=ccrs.PlateCarree(),draw_labels=True,
                  linewidth=0.5,color='w',alpha=0.7,linestyle='--')
gl.xlabels_top = False
gl.ylabels_right = False
gl.xformatter = LONGITUDE_FORMATTER
gl.yformatter = LATITUDE_FORMATTER