Python

• Remove special characters from a string:

re.sub('[^A-Za-z0-9-_ ]+', '', inp_string)

• Update a list while iterating on it (using slice assignment)

b = a
a[:] = [col.lower() for col in a]
print(b)

• All combinations of numbers

from itertools import combinations

list(combinations(range(4), 2)) # combinations(iterable, r) # n C r
# output [(0, 1), (0, 2), (0, 3), (1, 2), (1, 3), (2, 3)]

• Max of an iterable (list, tuple). key specifies a function of one argument that is used to extract a comparison key from each list element (for example, key=str.lower). The key corresponding to each item in the list is calculated once and then used for the entire sorting process. The default value of None means that list items are sorted directly without calculating a separate key value.

max(X_train, key=len) # key is a one argument function

• Sorting elements of an iterable (list, tuple) or instances of an object (class) on multiple levels and descending order. https://docs.python.org/3/howto/sorting.html#key-functions

student_tuples = [
     ('john', 'A', 15),
     ('jane', 'B', 12),
     ('dave', 'B', 10),
]

sorted(student_tuples, key=lambda student: student[2])   # sort by age
# output: [('dave', 'B', 10), ('jane', 'B', 12), ('john', 'A', 15)]

# using operator module in python to extend sorting features
from operator import itemgetter, attrgetter # attrgetter can be used to access named attributes of an instance (e.g. of a class)

sorted(student_tuples, key=itemgetter(2)) # sorting by the item at index 2
# output: [('dave', 'B', 10), ('jane', 'B', 12), ('john', 'A', 15)]

sorted(student_tuples, key=itemgetter(1,2)) # sorting by item at index 1 and then index 2
# output: [('john', 'A', 15), ('dave', 'B', 10), ('jane', 'B', 12)]

sorted(student_tuples, key=itemgetter(2), reverse=True)
# output: [('john', 'A', 15), ('jane', 'B', 12), ('dave', 'B', 10)]

NumPy

• Views vs Copy. https://numpy.org/doc/stable/user/basics.copies.html
  • Views are a way to access the data buffer (contiguous block of memory) of an numpy array, changing its stride or dtype. Any changes made to the view are reflected in the view.
    • Reshape creates a view if possible, but can create a copy. Basic indexing a[1:3] creates view
  • However, Copy duplicates the data buffer, so any changes made to the copy are not reflected back in the original array.
    • Flatten creates a copy. Advanced indexing a[[1, 2]] creates a copy

x = np.arange(9)

x
# array([0, 1, 2, 3, 4, 5, 6, 7, 8])

y = x.reshape(3, 3)

y
# array([[0, 1, 2],
#        [3, 4, 5],
#        [6, 7, 8]])

y.base  # .reshape() creates a view
# array([0, 1, 2, 3, 4, 5, 6, 7, 8])

z = y[[2, 1]]

z
# array([[6, 7, 8],
#       [3, 4, 5]])

z.base is None  # advanced indexing creates a copy
# True

• Reshape array. Useful to convert 1D array into a 2D array

  a = np.arange(5)
  print("Current shape {}".format(a.shape))
  print(a.reshape(-1, 1).shape) # -1 is the unknown dimension, it means as many rows as there are elements in the array
  print(a.reshape(1, -1).shape) # 1 row, as many columns as there are elements in the array
  

• Check if two arrays are element-wise equal with a tolerance

# np.allcose in an assert statement
assert np.allclose(avg, [0.1666667, 2.0]), "Debug the last example"

• Random number generator

  • From standard normal distribution (mean=0, std dev=1)

  # generates array of random numbers with shape (d0, d1, d2) from a normal distribution
  np.random.randn(d0, d1, d2)
  
  # generates a single floating point
  np.random.randn()