def test_find_token_no_matches(self, token_re):
"""None should be returned if the regex matches no tokens in a message."""
token_re.finditer.return_value = ()
return_value = TokenRemover.find_token_in_message(self.msg)
self.assertIsNone(return_value)
token_re.finditer.assert_called_once_with(self.msg.content)
def test_find_token_ignores_bot_messages(self, token_re):
"""The token finder should ignore messages authored by bots."""
self.msg.author.bot = True
return_value = TokenRemover.find_token_in_message(self.msg)
self.assertIsNone(return_value)
token_re.finditer.assert_not_called()
def test_find_token_invalid_matches(self, token_re, token_cls, is_valid_id,
is_valid_timestamp):
"""None should be returned if no matches have valid user IDs or timestamps."""
token_re.finditer.return_value = [
mock.create_autospec(Match, spec_set=True, instance=True)
]
token_cls.return_value = mock.create_autospec(Token,
spec_set=True,
instance=True)
is_valid_id.return_value = False
is_valid_timestamp.return_value = False
return_value = TokenRemover.find_token_in_message(self.msg)
self.assertIsNone(return_value)
token_re.finditer.assert_called_once_with(self.msg.content)
def test_find_token_valid_match(self, token_re, token_cls, is_valid_id,
is_valid_timestamp):
"""The first match with a valid user ID and timestamp should be returned as a `Token`."""
matches = [
mock.create_autospec(Match, spec_set=True, instance=True),
mock.create_autospec(Match, spec_set=True, instance=True),
]
tokens = [
mock.create_autospec(Token, spec_set=True, instance=True),
mock.create_autospec(Token, spec_set=True, instance=True),
]
token_re.finditer.return_value = matches
token_cls.side_effect = tokens
is_valid_id.side_effect = (
False, True) # The 1st match will be invalid, 2nd one valid.
is_valid_timestamp.return_value = True
return_value = TokenRemover.find_token_in_message(self.msg)
self.assertEqual(tokens[1], return_value)
token_re.finditer.assert_called_once_with(self.msg.content)
async def relay_message(self, msg: Message) -> None:
"""Relays the message to the relevant watch channel."""
limit = BigBrotherConfig.header_message_limit
if (msg.author.id != self.message_history.last_author
or msg.channel.id != self.message_history.last_channel
or self.message_history.message_count >= limit):
self.message_history = MessageHistory(last_author=msg.author.id,
last_channel=msg.channel.id)
await self.send_header(msg)
if TokenRemover.find_token_in_message(msg) or WEBHOOK_URL_RE.search(
msg.content):
cleaned_content = "Content is censored because it contains a bot or webhook token."
elif cleaned_content := msg.clean_content:
# Put all non-media URLs in a code block to prevent embeds
media_urls = {
embed.url
for embed in msg.embeds if embed.type in ("image", "video")
}
for url in URL_RE.findall(cleaned_content):
if url not in media_urls:
cleaned_content = cleaned_content.replace(url, f"`{url}`")
async def on_message(self, msg: Message) -> None:
"""
Detect poorly formatted Python code in new messages.
If poorly formatted code is detected, send the user a helpful message explaining how to do
properly formatted Python syntax highlighting codeblocks.
"""
is_help_channel = (getattr(msg.channel, "category", None)
and msg.channel.category.id
in (Categories.help_available,
Categories.help_in_use))
parse_codeblock = ((is_help_channel
or msg.channel.id in self.channel_cooldowns
or msg.channel.id in self.channel_whitelist)
and not msg.author.bot
and len(msg.content.splitlines()) > 3
and not TokenRemover.find_token_in_message(msg))
if parse_codeblock: # no token in the msg
on_cooldown = (time.time() -
self.channel_cooldowns.get(msg.channel.id, 0)) < 300
if not on_cooldown or DEBUG_MODE:
try:
if self.has_bad_ticks(msg):
ticks = msg.content[:3]
content = self.codeblock_stripping(
f"```{msg.content[3:-3]}```", True)
if content is None:
return
content, repl_code = content
if len(content) == 2:
content = content[1]
else:
content = content[0]
space_left = 204
if len(content) >= space_left:
current_length = 0
lines_walked = 0
for line in content.splitlines(keepends=True):
if current_length + len(
line
) > space_left or lines_walked == 10:
break
current_length += len(line)
lines_walked += 1
content = content[:current_length] + "#..."
content_escaped_markdown = RE_MARKDOWN.sub(
r'\\\1', content)
howto = (
"It looks like you are trying to paste code into this channel.\n\n"
"You seem to be using the wrong symbols to indicate where the codeblock should start. "
f"The correct symbols would be \\`\\`\\`, not `{ticks}`.\n\n"
"**Here is an example of how it should look:**\n"
f"\\`\\`\\`python\n{content_escaped_markdown}\n\\`\\`\\`\n\n"
"**This will result in the following:**\n"
f"```python\n{content}\n```")
else:
howto = ""
content = self.codeblock_stripping(msg.content, False)
if content is None:
return
content, repl_code = content
# Attempts to parse the message into an AST node.
# Invalid Python code will raise a SyntaxError.
tree = ast.parse(content[0])
# Multiple lines of single words could be interpreted as expressions.
# This check is to avoid all nodes being parsed as expressions.
# (e.g. words over multiple lines)
if not all(
isinstance(node, ast.Expr)
for node in tree.body) or repl_code:
# Shorten the code to 10 lines and/or 204 characters.
space_left = 204
if content and repl_code:
content = content[1]
else:
content = content[0]
if len(content) >= space_left:
current_length = 0
lines_walked = 0
for line in content.splitlines(keepends=True):
if current_length + len(
line
) > space_left or lines_walked == 10:
break
current_length += len(line)
lines_walked += 1
content = content[:current_length] + "#..."
content_escaped_markdown = RE_MARKDOWN.sub(
r'\\\1', content)
howto += (
"It looks like you're trying to paste code into this channel.\n\n"
"Discord has support for Markdown, which allows you to post code with full "
"syntax highlighting. Please use these whenever you paste code, as this "
"helps improve the legibility and makes it easier for us to help you.\n\n"
f"**To do this, use the following method:**\n"
f"\\`\\`\\`python\n{content_escaped_markdown}\n\\`\\`\\`\n\n"
"**This will result in the following:**\n"
f"```python\n{content}\n```")
log.debug(
f"{msg.author} posted something that needed to be put inside python code "
"blocks. Sending the user some instructions.")
else:
log.trace(
"The code consists only of expressions, not sending instructions"
)
if howto != "":
howto_embed = Embed(description=howto)
bot_message = await msg.channel.send(
f"Hey {msg.author.mention}!", embed=howto_embed)
self.codeblock_message_ids[msg.id] = bot_message.id
self.bot.loop.create_task(
wait_for_deletion(bot_message,
user_ids=(msg.author.id, ),
client=self.bot))
else:
return
if msg.channel.id not in self.channel_whitelist:
self.channel_cooldowns[msg.channel.id] = time.time()
except SyntaxError:
log.trace(
f"{msg.author} posted in a help channel, and when we tried to parse it as Python code, "
"ast.parse raised a SyntaxError. This probably just means it wasn't Python code. "
f"The message that was posted was:\n\n{msg.content}\n\n"
)