Add rich embed filter

4 files changed, 62 insertions, 19 deletions

diff --git a/bot/exts/filtering/_filter_context.py b/bot/exts/filtering/_filter_context.py
index da7ba0c77..bcbafe393 100644
--- a/bot/exts/filtering/_filter_context.py
+++ b/bot/exts/filtering/_filter_context.py
@@ -1,6 +1,6 @@
 from __future__ import annotations
 
-from collections.abc import Coroutine
+from collections.abc import Callable, Coroutine
 from dataclasses import dataclass, field, replace
 from enum import Enum, auto
 from typing import Optional, Union
@@ -35,7 +35,8 @@ class FilterContext:
     action_descriptions: list = field(default_factory=list)  # What actions were taken
     matches: list = field(default_factory=list)  # What exactly was found
     notification_domain: str = field(default_factory=str)  # A domain to send the user for context
-    additional_actions: list[Coroutine] = field(default_factory=list)  # Additional actions to perform
+    # Additional actions to perform
+    additional_actions: list[Callable[[FilterContext], Coroutine]] = field(default_factory=list)
 
     def replace(self, **changes) -> FilterContext:
         """Return a new context object assigning new values to the specified fields."""
diff --git a/bot/exts/filtering/_filters/unique/rich_embed.py b/bot/exts/filtering/_filters/unique/rich_embed.py
new file mode 100644
index 000000000..75f578d3e
--- /dev/null
+++ b/bot/exts/filtering/_filters/unique/rich_embed.py
@@ -0,0 +1,43 @@
+import re
+
+from botcore.utils.logging import get_logger
+
+from bot.exts.filtering._filter_context import Event, FilterContext
+from bot.exts.filtering._filters.filter import UniqueFilter
+from bot.utils.helpers import remove_subdomain_from_url
+
+log = get_logger(__name__)
+
+URL_RE = re.compile(r"(https?://\S+)", flags=re.IGNORECASE)
+
+
+class RichEmbedFilter(UniqueFilter):
+    """Filter messages which contain rich embeds not auto-generated from a URL."""
+
+    name = "rich_embed"
+    events = (Event.MESSAGE, Event.MESSAGE_EDIT)
+
+    def triggered_on(self, ctx: FilterContext) -> bool:
+        """Determine if `msg` contains any rich embeds not auto-generated from a URL."""
+        if ctx.message.embeds:
+            for embed in ctx.message.embeds:
+                if embed.type == "rich":
+                    urls = URL_RE.findall(ctx.content)
+                    final_urls = set(urls)
+                    # This is due to the way discord renders relative urls in Embeds
+                    # if the following url is sent: https://mobile.twitter.com/something
+                    # Discord renders it as https://twitter.com/something
+                    for url in urls:
+                        final_urls.add(remove_subdomain_from_url(url))
+                    if not embed.url or embed.url not in final_urls:
+                        # If `embed.url` does not exist or if `embed.url` is not part of the content
+                        # of the message, it's unlikely to be an auto-generated embed by Discord.
+                        ctx.alert_embeds.extend(ctx.message.embeds)
+                        return True
+                    else:
+                        log.trace(
+                            "Found a rich embed sent by a regular user account, "
+                            "but it was likely just an automatic URL embed."
+                        )
+
+        return False
diff --git a/bot/exts/filtering/_filters/unique/webhook.py b/bot/exts/filtering/_filters/unique/webhook.py
index 10f27a922..ee6b7135e 100644
--- a/bot/exts/filtering/_filters/unique/webhook.py
+++ b/bot/exts/filtering/_filters/unique/webhook.py
@@ -1,4 +1,5 @@
 import re
+from collections.abc import Callable, Coroutine
 
 from botcore.utils.logging import get_logger
 
@@ -41,18 +42,22 @@ class WebhookFilter(UniqueFilter):
         for i, match in enumerate(matches, start=1):
             extra = "" if len(matches) == 1 else f" ({i})"
             # Queue the webhook for deletion.
-            ctx.additional_actions.append(self._delete_webhook(ctx, match[0], extra))
+            ctx.additional_actions.append(self._delete_webhook_wrapper(match[0], extra))
             # Don't show the full webhook in places such as the mod alert.
             ctx.content = ctx.content.replace(match[0], match[1] + "xxx")
 
         return True
 
     @staticmethod
-    async def _delete_webhook(ctx: FilterContext, webhook_url: str, extra_message: str) -> None:
-        """Delete the given webhook and update the filter context."""
-        async with bot.instance.http_session.delete(webhook_url) as resp:
-            # The Discord API Returns a 204 NO CONTENT response on success.
-            if resp.status == 204:
-                ctx.action_descriptions.append("webhook deleted" + extra_message)
-            else:
-                ctx.action_descriptions.append("failed to delete webhook" + extra_message)
+    def _delete_webhook_wrapper(webhook_url: str, extra_message: str) -> Callable[[FilterContext], Coroutine]:
+        """Create the action to perform when a webhook should be deleted."""
+        async def _delete_webhook(ctx: FilterContext) -> None:
+            """Delete the given webhook and update the filter context."""
+            async with bot.instance.http_session.delete(webhook_url) as resp:
+                # The Discord API Returns a 204 NO CONTENT response on success.
+                if resp.status == 204:
+                    ctx.action_descriptions.append("webhook deleted" + extra_message)
+                else:
+                    ctx.action_descriptions.append("failed to delete webhook" + extra_message)
+
+        return _delete_webhook
diff --git a/bot/exts/filtering/_settings.py b/bot/exts/filtering/_settings.py
index d53334c1c..85a6f3d2b 100644
--- a/bot/exts/filtering/_settings.py
+++ b/bot/exts/filtering/_settings.py
@@ -197,14 +197,8 @@ class ActionSettings(Settings[ActionEntry]):
         for entry in self.values():
             await entry.action(ctx)
 
-        _i = len(ctx.additional_actions)
-        try:
-            for _i, action in enumerate(ctx.additional_actions):
-                await action
-        except Exception:
-            for action in ctx.additional_actions[_i+1:]:
-                action.close()
-            raise
+        for action in ctx.additional_actions:
+            await action(ctx)
 
     def fallback_to(self, fallback: ActionSettings) -> ActionSettings:
         """Fill in missing entries from `fallback`."""