# lecture 2, numpy-module

# little Python warmup:

# pep8 is the standard usually followed
# when it comes to Python coding style
numbers = [3, 5, 6, 7, 8, 6, 5, 4, 6, 7]

print(numbers)

print(numbers[5])

print(numbers[0:3])

for n in numbers:
    print(n)


# a list of lists, or in numpy => Matrix

day_1 = [20, 24, 18, 23]
day_2 = [15, 13, 12, 10]
day_3 = [30, 32, 27, 28]

temperatures = [day_1, day_2, day_3]

print()
print(temperatures[1][0])
print()

for day in temperatures:
    print("NEW DAY!")
    for t in day:
        print(t)

# NEW FILE!

import numpy as np

# if you have existing Python lists, or lists of lists
# you can easily convert these into numpy arrays!
numbers = [1, 2, 3, 400, 3]

data = np.array(numbers)

list1 = [1, 2, 3, 4]
list2 = [8, 9, 10, 11]
list3 = [15, 16, 17, 18]

combined = [list1, list2, list3]

matrix = np.array(combined)

# NEW FILE!

import numpy as np

# first parameter is included in the list
# second parameter is excluded in the list
# so last number is 20 - 1 = 19
data = np.arange(0, 20)

data = np.arange(5, 101, 5)

ones = np.ones(10)
zeroes = np.zeros(15)

# matrix of ones, 5 by 5
one_matrix = np.ones((5, 5))

# 10 by 10 matrix = 10 * 10 = 100
zero_matrix = np.zeros((10, 10))

# if you ever need an identity matrix quickly,
# this is the way!
id_matrix = np.eye(8)

# a list of evenly distributed values
lin_data = np.linspace(0, 150, 11)

# NEW FILE!

import numpy as np

# vector, 30 values
data = np.random.rand(30)

# 10 by 10 = 100 values
data = np.random.rand(10, 10)

standard = np.random.randn(10)

number = np.random.randint(0, 100)
numbers = np.random.randint(1, 1000, 50)

# reshape-function!

data = np.arange(1, 501)
data = data.reshape(50, 10)

# the recommended way, chained functions.
# first create a large vector
# and immediately reshape it to matrix
data = np.arange(1, 101).reshape(10, 10)

numbers = np.random.randint(1, 1000, 50).reshape(10, 5)

# if you ever need to know the dimensions of the data,
# use the shape
print(numbers.shape)

# these work mostly on vectors
max_val = numbers.max()
min_val = numbers.min()

max_val_pos = numbers.argmax()
min_val_pos = numbers.argmin()

# NEW FILE!

import numpy as np

# basic index selections are identical to plain Python
data = np.arange(50, 100)

single_value = data[15]

sublist1 = data[1:10]

sublist2 = data[:25]

sublist3 = data[25:]

matrix = np.arange(5, 46, 5).reshape(3, 3)

# first row and column
value = matrix[0][0]

# second row, third column
value = matrix[1,2]

# third row
row = matrix[2]

# everything BEFORE second row,
# everything AFTER second column
section = matrix[:1, 1:]

matrix = np.arange(1, 26).reshape(5,5)

section = matrix[:3, 2:]

# taking the fifth column
column = matrix[:, 4]

# NEW FILE!

import numpy as np

numbers = np.random.randint(1, 1000, 20)

big_numbers = numbers[numbers >= 500]
odd_numbers = numbers[numbers % 2 != 0]

# broadcasting
# remember to make a copy of your data
# if you don't want operations to
# affect original data as well!
# this example shows what happens:
# compare data and data_copy
data = np.arange(1, 101)
data_copy = data[0:10]
data_copy[:] = -1

# version that uses copy()
data = np.arange(1, 101)
data_copy = data[0:10].copy()
data_copy[:] = -1

# arithmetic operations
data = np.arange(1, 101)
combined = data + data
reduced = data - data
multiplied = data * data

increased = data + 100
data = np.ones(10) * 25

# NEW FILE!

import numpy as np

data = np.arange(0, 11)

# 0/0 => NaN => Not a Number
result = data / data

# 1 / 0 => inf => infinite
other = 1 / data

# sum and standard deviation
numbers = np.random.randint(1, 1000, 5)
random_total = np.sum(numbers)

# low standard deviation means the values are close to the average value
# high standard deviation menas the values are far apart from each other
random_deviation = np.std(numbers)