# Import local package import text_analyzer # Sum word_counts using sum_counters from text_analyzer word_count_totals = text_analyzer.sum_counters(word_counts) # Plot word_count_totals using plot_counter from text_analyzer text_analyzer.plot_counter(word_count_totals)
# Import custom text_analyzer package import text_analyzer # Create a SocialMedia instance with datacamp_tweets dc_tweets = text_analyzer.SocialMedia(text=datacamp_tweets) # Print the top five most most mentioned users print(dc_tweets.mention_counts.most_common(5)) # Plot the most used hashtags text_analyzer.plot_counter(dc_tweets.hashtag_counts)
self.hashtag_counts = self._count_hashtags()
self.mention_counts = self._count_mentions()
def _count_hashtags(self):
# Filter attribute so only words starting with '#' remain
return filter_word_counts(self.word_counts, first_char='#')
def _count_mentions(self):
# Filter attribute so only words starting with '@' remain
return filter_word_counts(self.word_counts, first_char='@')
Instructions
100 XP
import your text_analyzer custom package.
Define dc_tweets as an instance of SocialMedia with the preloaded datacamp_tweets object as the text.
print the 5 most_common mentioned users in the data using the appropriate dc_tweets attribute.
Use text_analyzer's plot_counter() method to plot the most used hashtags in the data using the appropriate dc_tweets attribute.
'''
SOLUTION
# Import custom text_analyzer package
import text_analyzer
# Create a SocialMedia instance with datacamp_tweets
dc_tweets = text_analyzer.SocialMedia(text=datacamp_tweets)
# Print the top five most most mentioned users
print(dc_tweets.mention_counts.most_common(5))
# Plot the most used hashtags
text_analyzer.plot_counter(dc_tweets.word_counts)
class Document:
"""A class for text analysis
:param text: string of text to be analyzed
:ivar text: string of text to be analyzed; set by `text` parameter
"""
# Method to create a new instance of MyClass
def __init__(self, text):
# Store text parameter to the text attribute
self.text = text
# Tokenize the document with non-public tokenize method
self.tokens = self._tokenize()
# Perform word count with non-public count_words method
self.word_counts = self._count_words
# Assumes tokenize and Counter from other packages are imported
def _tokenize(self):
return tokenize(self.text)
# Non-public method to tally document's word counts with Counter
def _count_words(self):
return Counter(self.tokens)
# Can be imported with
from .document import Document
# Using the functionality
# Import custom text_analyzer package
import text_analyzer
# Create an instance of Document with datacamp_tweet
my_document = text_analyzer.Document(text=datacamp_tweet)
# Print the text attribute of the Document instance
print(my_document.text)
# Print the first 5 tokens from datacamp_doc
print(datacamp_doc.tokens[:5])
# Print the top 5 most used words in datacamp_doc
print(datacamp_doc.word_counts.most_common(5))
# Class inheritance ----------------------------------------------------------------------------------------------------
# Child inherits all attributes from the parant, plus can have some extensions
# Define a SocialMedia class that is a child of the `Document class`
class SocialMedia(Document):
def __init__(self, text):
Document.__init__(self, text)
self.hashtag_counts = self._count_hashtags()
self.mention_counts = self._count_mentions()
def _count_hashtags(self):
# Filter attribute so only words starting with '#' remain
return filter_word_counts(self.word_counts, first_char='#')
def _count_mentions(self):
# Filter attribute so only words starting with '@' remain
return filter_word_counts(self.word_counts, first_char='@')
# Using child class
# Import custom text_analyzer package
import text_analyzer
# Create a SocialMedia instance with datacamp_tweets
dc_tweets = text_analyzer.SocialMedia(text=datacamp_tweets)
# Print the top five most most mentioned users
print(dc_tweets.mention_counts.most_common(5))
# Plot the most used hashtags
text_analyzer.plot_counter(dc_tweets.hashtag_counts)
# Multilevel inheritance -----------------------------------------------------------------------------------------------
# One child can inherit from multiple parents
# Define a Tweet class that inherits from SocialMedia
class Tweets(SocialMedia):
def __init__(self, text):
# Call parent's __init__ with super()
super().__init__(self, text)
# Define retweets attribute with non-public method
self.retweets = self._process_retweets()
def _process_retweets(self):
# Filter tweet text to only include retweets
retweet_text = filter_lines(self.text, first_chars='RT')
# Return retweet_text as a SocialMedia object
return SocialMedia(retweet_text)
# Using the grandchild
# Import needed package
import text_analyzer
# Create instance of Tweets
my_tweets = text_analyzer.Tweets(datacamp_tweets)
# Plot the most used hashtags in the tweets
my_tweets.plot_counts('hashtag_counts')
# Plot the most used hashtags in the retweets
my_tweets.retweets.plot_counts('hashtag_counts')
# Docstrings ----------------------------------------------------------------------------------------------------------
def tokenize(text, regex=r'[a-zA-z]+'):
"""Split text into tokens using a regular expression
:param text: text to be tokenized
:param regex: regular expression used to match tokens using re.findall
:return: a list of resulting tokens
>>> tokenize('the rain in spain')
['the', 'rain', 'in', 'spain']
"""
return re.findall(regex, text, flags=re.IGNORECASE)
# Print the docstring
help(tokenize)
# Google docstring style
"""Google style.
The Google style tends to result in
wider docstrings with fewer lines of code.
Section 1:
Item 1: Item descriptions don't need line breaks.
"""
# Numpy docstring style
"""Numpy style.
The Numpy style tends to results in
narrower docstrings with more lines of code.
Section 1
---------
Item 1
Item descriptions are indented on a new line.
"""
# Building docstring from multiple strings in parenthesis
def get_matches(word_list: List[str], query:str) -> List[str]:
("Find lines containing the query string.\nExamples:\n\t"
# Complete the docstring example below
">>> get_matches(['a', 'list', 'of', 'words'], 's')\n\t"
# Fill in the expected result of the function call
"['list', 'words']")
return [line for line in word_list if query in line]
# The Zen of Python ---------------------------------------------------------------------------------------------------
import this
# Testing with doctest ------------------------------------------------------------------------------------------------
def sum_counters(counters):
"""Aggregate collections.Counter objects by summing counts
:param counters: list/tuple of counters to sum
:return: aggregated counters with counts summed
>>> d1 = text_analyzer.Document('1 2 fizz 4 buzz fizz 7 8')
>>> d2 = text_analyzer.Document('fizz buzz 11 fizz 13 14')
>>> sum_counters([d1.word_counts, d2.word_counts])
Counter({'buzz': 2, 'fizz': 4})
"""
return sum(counters, Counter())
doctest.testmod()
# Testing with pytest -------------------------------------------------------------------------------------------------
# working_dir
# ├── text_analyzer
# │ ├── __init__.py
# │ ├── counter_utils.py
# │ ├── document.py
# ├── setup.py
# ├── requirements.py
# └── tests
# └── test_unit.py
from collections import Counter
from text_analyzer import SocialMedia
# Create an instance of SocialMedia for testing
test_post = 'learning #python & #rstats is awesome! thanks @datacamp!'
sm_post = SocialMedia(test_post)
# Test hashtag counts are created properly
def test_social_media_hashtags():
expected_hashtag_counts = Counter({'#python': 1, '#rstats': 1})
assert sm_post.hashtag_counts == expected_hashtag_counts
# Run from command line
$ pytest
# Parametrizing test
@pytest.mark.parametrize("inputs", ["intro.md", "plot.py", "discussion.md"])
def test_nbuild(inputs):
assert nbuild([inputs]).cells[0].source == Path(inputs).read_text()
# Check whether error is raised
@pytest.mark.parametrize("not_exporters", ["htm", "ipython", "markup"])
def test_nbconv(not_exporters):
with pytest.raises(ValueError):
nbconv(nb_name="mynotebook.ipynb", exporter=not_exporters)
# Classmethods --------------------------------------------------------------------------------------------------------
# Due to the classmethod decorator, the first argument in the function it decorates is not the class' instance (self),
# but the class itself. This allows to instatiate multiple instances at ones.
class TextFile:
instances = []
def__init__(self, file):
self.text = Path(file).read_text()
self.__class__.instances.append(file)
@classmethod
def instantiate(cls, filenames):
return map(cls, filenames)
'manipulate': 1,
'p': 1,
'pandas': 1,
'series': 1,
'time': 1,
'to': 1,
'with': 1
})
]
# Sum word_counts using sum_counters from text_analyzer
word_count_totals = text_analyzer.sum_counters(word_counts)
print(word_count_totals)
# Plot word_count_totals using plot_counter from text_analyzer
top_5_items = text_analyzer.plot_counter(word_count_totals)
print(top_5_items)
# Create an instance of Document with datacamp_tweet
datacamp_tweet = 'Basic linear regression example. #DataCamp #DataScience #Python #sklearn'
my_document = text_analyzer.Document(text=datacamp_tweet)
# Print the text attribute of the Document instance
print(my_document.text)
# create a new document instance from datacamp_tweets
datacamp_tweets = '[DataCamp] Introduction to H2O AutoML --> In this tutorial, you will learn about H2O and have a glimpse of its auto…\n[DataCamp] Stocks, Significance Testing & p-Hacking --> Learn how to manipulate time series data with pandas and co…\nRT @cbismuth: Linear regression example with most significant features detection. #DataCamp #DataScience #Python #sklearn …\nLinear regression example with most significant features detection. #DataCamp #DataScience #Python #sklearn\nBasic linear regression example. #DataCamp #DataScience #Python #sklearn\nRT @David_Makinde_: I just completed Introduction to Python for Data Science \n#Datacamp\n#DataScience \n#Python\n[DataCamp] Enter the #DataFramedChallenge for a chance to be on an upcoming podcast segment. --> DataCamp has a pod…\n[DataCamp] Introduction to Python Metaclasses --> In this tutorial, you\'ll learn about metaclasses in Python. by De…\nI just completed Introduction to Python for Data Science \n#Datacamp\n#DataScience \n#Python\nRT @cbismuth: My pretty first classifier! #DataCamp #Python #sklearn\nMy pretty first classifier! #DataCamp #Python #sklearn\nRT @ascentt: The different #DataScience roles on the job market.'
datacamp_doc = text_analyzer.Document(datacamp_tweets)
# print the first 5 tokens from datacamp_doc
print(datacamp_doc.text)