Better URL detection regex

rocketbot/cogs/urlspamcog.py

@@ -228,12 +228,24 @@ class URLSpamCog(BaseCog, name='URL Spam'):
 
 	def is_url(self, s: str):
 		'Tests if a string is strictly a URL'
-		pattern = r'http[s]?://(?:[a-zA-Z]|[0-9]|[$-_@.&+]|[!*\(\),]|(?:%[0-9a-fA-F][0-9a-fA-F]))+'
+		ipv6_host_pattern = r'\[[0-9a-fA-F:]+\]'
+		ipv4_host_pattern = r'[0-9\.]+'
+		hostname_pattern = r'[a-zA-Z0-9-]+\.[a-zA-Z0-9-\.]+'
+		host_pattern = r'(?:' + ipv6_host_pattern + '|' + ipv4_host_pattern + '|' + hostname_pattern + ')'
+		port_pattern = '(?::[0-9]+)?'
+		path_pattern = r'(?:/[^ \]\)]*)?'
+		pattern = r'^http[s]?://' + host_pattern + port_pattern + path_pattern + '$'
 		return re.match(pattern, s, re.IGNORECASE) != None
 
 	def is_casual_url(self, s: str):
 		'Tests if a string is a "casual URL" with no scheme included'
-		pattern = r'(?:[a-zA-Z]|[0-9]|[$-_@.&+]|[!*\(\),]|(?:%[0-9a-fA-F][0-9a-fA-F]))+'
+		ipv6_host_pattern = r'\[[0-9a-fA-F:]+\]'
+		ipv4_host_pattern = r'[0-9\.]+'
+		hostname_pattern = r'[a-zA-Z0-9-]+\.[a-zA-Z0-9-\.]+'
+		host_pattern = r'(?:' + ipv6_host_pattern + '|' + ipv4_host_pattern + '|' + hostname_pattern + ')'
+		port_pattern = '(?::[0-9]+)?'
+		path_pattern = r'(?:/[^ \]\)]*)?'
+		pattern = r'^' + host_pattern + port_pattern + path_pattern + '$'
 		return re.match(pattern, s, re.IGNORECASE) != None
 
 	async def on_mod_react(self,