
import random
import time
import timeit
import matplotlib.pyplot as plt
import numpy as np


# utility functions

def get_words(n, wordsfile='/usr/share/dict/words'):
    """ Return list of n words """
    # Note that the lines when read in end in newlines; strip() gets rid of that.
    # Ignore words ending in 's or fewer than three letters or Capitalized.
    # Google "python random" for docs on random.sample.
    words = []
    for line in open(wordsfile):
        word = line.strip()
        if word.islower() and word[-2:] != "'s" and len(word) > 2:
            words.append(word)
    return random.sample(words, n)
    
def elapsed_sec(function, toofast=0.05):
    """ Return number of seconds it takes to run a given function. """
    #
    # The approach here is
    #   (a) run it n times and divide by n if it runs too fast to measure,
    #   (b) pick a reasonable value for n by starting small and doubling
    # See also https://docs.python.org/3/library/timeit.html .
    #
    # If you have a function like f(x) and you want to run that with x=10,
    # then pass in lambda:f(10) as the function argument;
    # see https://docs.python.org/3/tutorial/controlflow.html#lambda-expressions
    #
    n = 1
    while True:
        start_time = time.time()              # start timer
        for i in range(n):
            function()
        seconds = time.time() - start_time    # stop timer
        if seconds > toofast:
            return seconds/n
        else:
            n *= 2


# An example of Python's lambda function notation : 
f = lambda x: 2*x+1
f(10)

21


# Examples of get_words() and elapsed_sec() : 
print(f"4 words : {get_words(4)}")
print(f"time to get 100 words : {elapsed_sec(lambda:get_words(100)):.4f} sec.")

4 words : ['peopled', 'castoffs', 'pliable', 'bibliographical']
time to get 100 words : 0.0757 sec.


# Let's visualize how long that takes for different values of n.
ns = [100, 300, 1000, 3000, 10000, 30000]
secs = [elapsed_sec(lambda:get_words(n)) for n in ns]
print(f"secs = {secs}")

secs = [0.07015085220336914, 0.058068037033081055, 0.05502915382385254, 0.05725407600402832, 0.06629180908203125, 0.09157991409301758]


# And now let's make a plot of that.
# Note that some possible plot commands are 
#  (axis.semilogx, axis.semilogy, axis.loglog, axis.plot)
# depending on whether you want a log scale on x, y, both, or neither.

figure = plt.figure(dpi=220, figsize=(3, 2))   # dots_per_inch and (width, height) in inches
axis = figure.add_subplot(111)                 # 111 indicates (rows,cols,number) i.e. 1x1 #1 .
axis.set(xlabel="n", ylabel="seconds", title="time to get n words")
axis.set_ylim((0,0.2))
axis.semilogx(ns, secs, marker="+", color="blue", linestyle="none")
plt.show()


# Here's the same plot with a horizontal line added, showing that this is O(1).
x = np.linspace(ns[0], ns[-1], 100)  # 100 points
y = np.zeros(len(x)) + 0.06          # 100 points, all 0.05 (eyeballing an OK line.)

figure = plt.figure(dpi=220, figsize=(3, 2))   # dots_per_inch and (width, height) in inches
axis = figure.add_subplot(111)                 # 111 indicates (rows,cols,number) i.e. 1x1 #1 .
axis.set(xlabel="n", ylabel="seconds", title="time to get n words")
axis.set_ylim((0,0.2))
axis.semilogx(ns, secs, marker="+", color="blue", linestyle="none")

axis.semilogx(x, y, color='red')  # Here's the added line.

plt.show()

word search¶

TODO in class¶

Notebook review/reminder :¶

How to turn in one of these notebooks as homework :¶

linear search¶

binary search¶

hash table¶