SimpleText 1.0.1

A package to manage textual data in a simple fashion.

SimpleText

---

A package to manage textual data in a simple fashion.

Install with:

pip install SimpleText

1) The preprocess function

This function takes a string as an input and outputs a list of tokens. There are several parameters in the function to help quickly pre-process a string.

Parameters:

text (string): a string of text

n_grams (tuple, default = (1,1)): specifies the number of ngrams e.g. (1,2) would be unigrams and bigram, (2,2) would be just bigrams

remove_accents (boolean, default = False): removes accents

lower (boolean, default = False): lowercases text

remove_less_than (int, default = 0): removes words less than X letters

remove_more_than (int, default = 20): removes words more than X letters

remove_punct (boolean, default = False): removes punctuation

remove_alpha (boolean, default = False): removes non-alphabetic tokens

remove_stopwords (boolean, default = False): removes stopwords

remove_custom_stopwords (list, default = [ ]): removes custom stopwords

lemma (boolean, default = False): lemmantises tokens (via the Word Net Lemmantizer algorithm)

stem (boolean, default = False): stems tokens (via the Porter Stemming algorithm)

In the example below we preprocess the string by:

• lowercasing letters
• removing punctuation
• removing stop words
• removing words with more than 15 letters and less than 1 letter

from SimpleText.preprocessor import preprocess

text = 'Last week, I went to the shops.'

preprocess(text, n_grams=(1, 1), remove_accents=False, lower=True, remove_less_than=1,
           remove_more_than=15, remove_punct=True, remove_alpha=False, remove_stopwords=True,
           remove_custom_stopwords=[], lemma=False, stem=False, remove_url=False)

The output would be:

['last', 'went', 'shops', 'week']

In this second example we process the string by:

• generating unigrams and bigrams
• stemming
• removing the url
• removing accents
• lowercasing letters

from SimpleText.preprocessor import preprocess

text = "I'm loving the weather this year in españa! https://en.tutiempo.net/spain.html"

preprocess(text, n_grams=(1, 2), remove_accents=True, lower=True, remove_less_than=0, 
           remove_more_than=20, remove_punct=False, remove_alpha=False, remove_stopwords=False,remove_custom_stopwords=[], lemma=False, stem=True, remove_url=True)

This outputs:

["i'm",'love','the','weather','thi','year','in','espana!',("i'm", 'loving'),('loving', 'the'),('the', weather',
 ('weather', 'this'),('this', 'year'),('year', 'in'),('in', 'espana!')]

2) Individually preprocessing text

Alternatively, one can also individually apply a preprocessing step without having to use the whole preprocess function. The functions available are:

from SimpleText.preprocessor import lowercase, strip_accents, strip_punctuation, strip_url, 
tokenise, strip_alpha_numeric_characters, strip_stopwords, lemantization, stemming, get_ngrams

lowercase("Hi again") # outputs "hi again"

strip_accents("Hi ágain") # outputs "Hi again"

strip_punctuation("Hi again!") # outputs "Hi again"

strip_url("Hi again https//example.example.com/example/example") # outputs "Hi again"

tokenise("Hi again") # outputs ["Hi", "again"]

strip_alpha_numeric_characters(["Hi", "again", "@", "#", "*"]) # outputs ["Hi", "again"]

strip_stopwords(["Hi", "again"], ["Hi"]) # outputs ["again"]

lemantization(["bats", "feet"]) # outputs ["bat", "foot"]

stemming(["consult", "consultant", "consulting"]) # outputs ["consult", "consult", "consult"]

get_ngrams("hi all I'm", (1,3)) # outputs [('hi', 'all'), ('all', "I'm"), ('hi', 'all', "I'm")]