-
Notifications
You must be signed in to change notification settings - Fork 0
/
vectorization.py
46 lines (33 loc) · 1.58 KB
/
vectorization.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
from sklearn.feature_extraction.text import CountVectorizer
from sklearn.feature_extraction.text import TfidfVectorizer
import nltk
nltk.download('stopwords')
nltk.download('punkt')
def bagOfWords(text: str, ngram_range=(1, 1)):
"""
Bag Of Words implementation
"""
CountVec = CountVectorizer(ngram_range=ngram_range) # to use bigrams ngram_range=(2,2)
Count_data = CountVec.fit_transform([text])
# return python dict instead of df
# cv_dataframe = pd.DataFrame(Count_data.toarray(), columns=CountVec.get_feature_names_out())
res = list(map(lambda row: dict(zip(CountVec.get_feature_names_out(), row)), Count_data.toarray()))
return res[0]
def TfIdf(text: str, ngram_range=(1, 1), min_df: int = 3):
"""
TFIDF implementation. corpus are all texts
"""
tfIdfVectorizer = TfidfVectorizer(ngram_range=ngram_range)
Count_data = tfIdfVectorizer.fit_transform([text])
# return python dict instead of df
# cv_dataframe = pd.DataFrame(Count_data.toarray(), columns=CountVec.get_feature_names_out())
res = list(map(lambda row: dict(zip(tfIdfVectorizer.get_feature_names_out(), row)), Count_data.toarray()))
return res[0]
def getVectorizer(vecType, ngram_range=(1, 3), max_df=0.3, min_df=7):
"""
returns the vectorizer depending on the vectorization type
"""
if vecType.value == 2: # bag of words
return CountVectorizer(ngram_range=ngram_range, max_df=max_df, min_df=min_df)
elif vecType.value == 4: # tfidf
return TfidfVectorizer(ngram_range=ngram_range, norm="l2", max_df=max_df, min_df=min_df)