C2W1_L1_Building the vocabulary.ipynb

Kernel: Python 3

NLP Course 2 Week 1 Lesson : Building The Model - Lecture Exercise 01

Estimated Time: 10 minutes

Vocabulary Creation

Create a tiny vocabulary from a tiny corpus
It's time to start small !

Imports and Data

# imports
import re # regular expression library; for tokenization of words
from collections import Counter # collections library; counter: dict subclass for counting hashable objects
import matplotlib.pyplot as plt # for data visualization

# the tiny corpus of text ! 
text = 'red pink pink blue blue yellow ORANGE BLUE BLUE PINK' # 🌈
print(text)
print('string length : ',len(text))

red pink pink blue blue yellow ORANGE BLUE BLUE PINK
string length :  52

Preprocessing

# convert all letters to lower case
text_lowercase = text.lower()
print(text_lowercase)
print('string length : ',len(text_lowercase))

red pink pink blue blue yellow orange blue blue pink
string length :  52

# some regex to tokenize the string to words and return them in a list
words = re.findall(r'\w+', text_lowercase)
print(words)
print('count : ',len(words))

['red', 'pink', 'pink', 'blue', 'blue', 'yellow', 'orange', 'blue', 'blue', 'pink']
count :  10

Create Vocabulary

Option 1 : A set of distinct words from the text

# create vocab
vocab = set(words)
print(vocab)
print('count : ',len(vocab))

{'orange', 'blue', 'red', 'yellow', 'pink'}
count :  5

Add Information with Word Counts

Option 2 : Two alternatives for including the word count as well

# create vocab including word count
counts_a = dict()
for w in words:
    counts_a[w] = counts_a.get(w,0)+1
print(counts_a)
print('count : ',len(counts_a))

{'red': 1, 'pink': 3, 'blue': 4, 'yellow': 1, 'orange': 1}
count :  5

# create vocab inlcuding word count using collections.Counter
counts_b = dict()
counts_b = Counter(words)
print(counts_b)
print('count : ',len(counts_b))

Counter({'blue': 4, 'pink': 3, 'red': 1, 'yellow': 1, 'orange': 1})
count :  5

# barchart of sorted word counts
d = {'blue': counts_b['blue'], 'pink': counts_b['pink'], 'red': counts_b['red'], 'yellow': counts_b['yellow'], 'orange': counts_b['orange']}
plt.bar(range(len(d)), list(d.values()), align='center', color=d.keys())
_ = plt.xticks(range(len(d)), list(d.keys()))

Ungraded Exercise

Note that counts_b, above, returned by collections.Counter is sorted by word count

Can you modify the tiny corpus of text so that a new color appears between pink and red in counts_b ?

Do you need to run all the cells again, or just specific ones ?

print('counts_b : ', counts_b)
print('count : ', len(counts_a))

counts_b :  Counter({'blue': 4, 'pink': 3, 'red': 1, 'yellow': 1, 'orange': 1})
count :  5

Expected Outcome:

counts_b : Counter({'blue': 4, 'pink': 3, 'your_new_color_here': 2, red': 1, 'yellow': 1, 'orange': 1})
count : 6

Summary

This is a tiny example but the methodology scales very well.
In the assignment you will create a large vocabulary of thousands of words, from a corpus
of tens of thousands or words! But the mechanics are exactly the same.
The only extra things to pay attention to will be; run time, memory management and the vocab data structure.
So the choice of approach used in code blocks counts_a vs counts_b, above, will be important.