Shrikrishna commited on
Commit
7217837
·
verified ·
1 Parent(s): 9337873

Update app.py

Browse files
Files changed (1) hide show
  1. app.py +16 -0
app.py CHANGED
@@ -41,6 +41,22 @@ def remove_stopwords(text, language='english'):
41
  Returns:
42
  filtered_text (str): Text without stopwords.
43
  """
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
44
  words = text.split()
45
  filtered_words = [word for word in words if word.lower() not in stop_words]
46
  filtered_text = ' '.join(filtered_words)
 
41
  Returns:
42
  filtered_text (str): Text without stopwords.
43
  """
44
+ stop_words = set([
45
+ "i", "me", "my", "myself", "we", "our", "ours", "ourselves", "you", "your",
46
+ "yours", "yourself", "yourselves", "he", "him", "his", "himself", "she",
47
+ "her", "hers", "herself", "it", "its", "itself", "they", "them", "their",
48
+ "theirs", "themselves", "what", "which", "who", "whom", "this", "that",
49
+ "these", "those", "am", "is", "are", "was", "were", "be", "been", "being",
50
+ "have", "has", "had", "having", "do", "does", "did", "doing", "a", "an",
51
+ "the", "and", "but", "if", "or", "because", "as", "until", "while", "of",
52
+ "at", "by", "for", "with", "about", "against", "between", "into", "through",
53
+ "during", "before", "after", "above", "below", "to", "from", "up", "down",
54
+ "in", "out", "on", "off", "over", "under", "again", "further", "then",
55
+ "once", "here", "there", "when", "where", "why", "how", "all", "any",
56
+ "both", "each", "few", "more", "most", "other", "some", "such", "no",
57
+ "nor", "not", "only", "own", "same", "so", "than", "too", "very", "s",
58
+ "t", "can", "will", "just", "don", "should", "now"
59
+ ])
60
  words = text.split()
61
  filtered_words = [word for word in words if word.lower() not in stop_words]
62
  filtered_text = ' '.join(filtered_words)