Python cheatsheet

Variables and common literals

x = 5            # int
r = 0.1          # float
s = "text"       # str
flag = True      # bool
nothing = None

Lists

xs = [1, 2, 3]
xs.append(4)
xs[0]; xs[-1]; len(xs)

Tuples

cell = (4, 0)
row, col = cell

Dictionaries

d = {"a": 1}
d["b"] = 2
d["a"]; "a" in d; d.items()

Sets

blocked = {(1, 1), (2, 2)}
(1, 1) in blocked

Indexing and slicing

xs[0]; xs[-1]
xs[1:3]; xs[:2]; xs[2:]

Conditionals

if x > 0:
    ...
elif x < 0:
    ...
else:
    ...

Loops

for v in xs: ...
for i in range(5): ...
for i, v in enumerate(xs): ...
while not done: ...   # break / continue

Functions and multiple returns

def f(a, b=1):
    return x, y, z

x, y, z = f(2)

Imports

import numpy as np
import random
from math import sqrt

f-strings

f"episode {ep}: {total:.2f}"

List comprehensions

[v * v for v in range(5)]
[v for v in xs if v % 2 == 0]

Reading a text file

with open("corpus.txt", encoding="utf-8") as f:
    text = f.read()      # encoding="utf-8" keeps non-English text intact

with open("corpus.txt", encoding="utf-8") as f:
    for line in f:
        ...

Strings and adjacent pairs

for ch in s: ...           # iterate characters
"".join(list_of_chars)      # characters back to a string
zip(seq, seq[1:])          # adjacent pairs: (s0,s1), (s1,s2), ...
max(d, key=d.get)          # dict key with the largest value
from collections import Counter
Counter(zip(seq, seq[1:])).most_common(1)   # most frequent pair

numpy patterns

import numpy as np
np.zeros((5, 4)); np.ones(3)
a.shape; a.reshape(2, 10)
a[0, 1]; a[0]; a[1:3]
a + b; a * 2
a.sum(); np.max(a); np.argmax(a)
a @ b; np.dot(a, b)        # dot product / matmul
np.linalg.norm(v)          # vector length
M.mean(axis=0)             # column means
M @ q                      # (N,D) matrix times (D,) vector -> (N,) scores
np.argsort(scores)         # sort indices (ascending)
X, Y = np.meshgrid(xs, ys) # 2D coordinate grid from two 1D arrays

matplotlib

import matplotlib.pyplot as plt
plt.scatter(xs, ys); plt.plot(xs, ys)
plt.contour(X, Y, Z, levels=30)   # contour lines (plt.contourf to fill)
plt.colorbar()                    # optional scale
plt.xlabel("x"); plt.ylabel("y"); plt.title("t")
plt.savefig("plot.png"); plt.show()

Debugging reminders

• Print the value: print("Q row:", Q[s]).
• Print the shape: print(a.shape).
• Read the bottom line of a traceback first; it names the error.
• Check you are in the right folder before running.
• Check your imports succeeded (a ModuleNotFoundError means the package is not installed in this Python).