Spaces:
Running
Running
aliasgerovs
commited on
Commit
•
932cfaa
1
Parent(s):
aae711a
Update utils.py
Browse files
utils.py
CHANGED
@@ -31,10 +31,10 @@ def remove_accents(input_str):
|
|
31 |
|
32 |
|
33 |
def remove_special_characters(text):
|
|
|
34 |
text = remove_accents(text)
|
35 |
pattern = r'[^\w\s\d.,!?\'"()-;]+'
|
36 |
text = re.sub(pattern, "", text)
|
37 |
-
text = text.replace("<s>", "").replace("</s>", "")
|
38 |
return text
|
39 |
|
40 |
|
|
|
31 |
|
32 |
|
33 |
def remove_special_characters(text):
|
34 |
+
text = text.replace("<s>", "").replace("</s>", "")
|
35 |
text = remove_accents(text)
|
36 |
pattern = r'[^\w\s\d.,!?\'"()-;]+'
|
37 |
text = re.sub(pattern, "", text)
|
|
|
38 |
return text
|
39 |
|
40 |
|