improve filename_from_url() performance

Manually extracting the part between the last '/' and '?' instead of
relying on the standard libraries' 'urllib.parse.urlsplit()' increases
performance by ~400%.

urlsplit() : 3.64 secs per 1.000.000 iterations
partition(): 0.87 secs per 1.000.000 iterations

gallery_dl/text.py

@@ -70,7 +70,7 @@ def ensure_http_scheme(url, scheme="https://"):
 def filename_from_url(url):
     """Extract the last part of an URL to use as a filename"""
     try:
-        return urllib.parse.urlsplit(url).path.rpartition("/")[2]
+        return url.partition("?")[0].rpartition("/")[2]
     except (TypeError, AttributeError):
         return ""