Refactored regex

frandier · Oct 11, 2022 · 95c9cea · 95c9cea
1 parent 2d95686
commit 95c9cea
Show file tree

Hide file tree

Showing 2 changed files with 7 additions and 7 deletions.
diff --git a/cleaner/cleaner.py b/cleaner/cleaner.py
@@ -3,10 +3,9 @@
 class Cleaner:
     def clean_tweet(self, tweet):
         r = tweet.lower()
-        r = re.sub("@[A-Za-z0-9_]+","", r)
-        r = re.sub("#[A-Za-z0-9_]+","", r)
-        r = re.sub(r'http\S+', '', r)
-        r = re.sub('[()!?]', ' ', r)
-        r = re.sub('\[.*?\]',' ', r)
-        r = re.sub("[^a-z0-9]"," ", r)
+        r = re.sub("(@[A-Za-z0-9_]+)", ' ', r)
+        r = re.sub(r'((?<=[A-Za-z])(?=[A-Z][a-z]))',' ', r)
+        r = re.sub("([^A-Za-z0-9äÄëËïÏöÖüÜáéíóúáéíóúÁÉÍÓÚÂÊÎÔÛâêîôûàèìòùÀÈÌÒÙñÑ])",' ', r)
+        r = re.sub("(\w+:\/\/\S+)",' ', r)
+        r = ' '.join(r.split())
         return r
diff --git a/main.py b/main.py
@@ -17,7 +17,8 @@ def main():
 
     try:
         query = str(input("Enter a search query: "))
-        tweets = client.get_tweets(query, 1, "4.570868,-74.297333,100km")
+        total = int(input("Enter the number of tweets to collect: "))
+        tweets = client.get_tweets(query, total, "4.570868,-74.297333,100km")
     except:
         print("Could not get tweets")
         sys.exit(1)