def test_tokenise_no(norwegian_language_model, text, expectedoutput):
# arange - not needed
# act
list = tokenise(norwegian_language_model, text)
# assert
assert list == expectedoutput
def test_tokenise_en(english_language_model, text, expectedoutput):
# arange - not needed
# act
list = tokenise(english_language_model, text)
# assert
assert list == expectedoutput
def test_retainspaces(english_language_model, text, expectedoutput):
# arange - not needed
# act
tokens = tokenise(english_language_model, text)
raw_str = ' '.join(tokens)
new_text = retain_spaces(raw_str)
# assert
assert new_text == expectedoutput
distribution of the words in the text.
"""
import enlp.visualisation.freq_distribution as viz
import enlp.understanding.distributions as dists
import enlp.processing.stdtools as nlp
import spacy
import matplotlib.pyplot as plt
plt.close(
'all'
) # very important for read the docs to avoid it crashing due to memory
###############################################################################
# Load Spacy's Norwegian language model and the example text
langmodel = spacy.load('nb_dep_ud_sm')
with open("example_data/no_den_stygge_andungen.txt", "r") as file:
text = file.read()
###############################################################################
# Make strings into list of tokens and count them
word_list = nlp.tokenise(langmodel, text)
counts = dists.freq_dist(word_list)
###############################################################################
# Visualise
fig, (ax1, ax2) = plt.subplots(1, 2, figsize=(10, 5))
viz.dist_plot_detailed(counts[:25],
ax=ax1) # Detailed distribution of top 25 tokens
viz.dist_plot(counts, log=True, ax=ax2) # Full distribution of corpus
plt.show()
# Get a list of stopwords to be removed from the text
# Get stopwords
all_stopwords, stopwords_nb, stopwords_en = get_stopwords()
###############################################################################
# Using NLP pipeline class to create a processing workflow. The pipeline shall involve:
# - remove punctuation
# - remove stopwords
# - stem remaining words
# Initialise object
processed_text = NLPPipeline(langmodel, text)
# Run processing as a pipeline
processed_text.rm_punctuation().rm_stopwords(
stopwords=all_stopwords).nltk_stem_no()
#############################################################################
# Compare text strings of firs 80 characters of the original and processed.
print('Original: ', text[:80], '...')
print('Processed: ', processed_text.text[:80], '...')
#############################################################################
# Wordcloud comparison between most common words
fig, (ax1, ax2) = plt.subplots(1, 2, figsize=(10, 5))
wordcloud_plot(tokenise(langmodel, text), ax=ax1)
wordcloud_plot(processed_text.tokenise().tokens, ax=ax2)
plt.tight_layout()
print (text)
# Load spacy language model
langmodel = spacy.load('en_core_web_md')
###############################################################################
# Get stopwords: the stopwords function will get all english and norwegian stopwords
stopwords_func, stopwords_nb_func, stopwords_en_func = get_stopwords()
print (stopwords_en_func[:5])
###############################################################################
# Remove english stopwords
processed_text = rm_stopwords(langmodel, text, stopwords_en_func)
print (processed_text)
###############################################################################
# Now that the stopwords have been removed then the sentences no longer make much sense. Therefore, an alternative to
# viewing the text output is to look at the distribution of the remaining text
orig_top10 = pd.DataFrame(dists.freq_dist(tokenise(langmodel, text)), columns=['token', 'count'])
pr_top10 = pd.DataFrame(dists.freq_dist(tokenise(langmodel, processed_text)), columns=['token', 'count'])
print ("ORIGINAL - top 10 words")
print (orig_top10.head(10))
print (" ")
print ("PROCESSED - top 10 words")
print (pr_top10.head(10))