Add anti-raid levels to automod. Large refactor of spam detection. Add member_join and member_join_spam triggers.

Anti-raid levels don't by themselves do anything, but they can be
used in overrides to activate specific automod items.

Spam detection should now be more reliable and also combine further
spam messages after the initial detection into the archive.

Messages deleted by automod no longer create the normal deletion log
entry. Instead, the AUTOMOD_ACTION log entry contains the deleted
message or an archive if there are multiple (i.e. spam).

backend/src/plugins/Automod/info.ts

@@ -0,0 +1,101 @@
+import { PluginInfo, trimPluginDescription } from "../ZeppelinPlugin";
+
+export const pluginInfo: PluginInfo = {
+  prettyName: "Automod",
+  description: trimPluginDescription(`
+      Allows specifying automated actions in response to triggers. Example use cases include word filtering and spam prevention.
+    `),
+  configurationGuide: trimPluginDescription(`
+      The automod plugin is very customizable. For a full list of available triggers, actions, and their options, see Config schema at the bottom of this page.    
+    
+      ### Simple word filter
+      Removes any messages that contain the word 'banana' and sends a warning to the user.
+      Moderators (level >= 50) are ignored by the filter based on the override.
+      
+      ~~~yml
+      automod:
+        config:
+          rules:
+            my_filter:
+              triggers:
+              - match_words:
+                  words: ['banana']
+                  case_sensitive: false
+                  only_full_words: true
+              actions:
+                clean: true
+                warn:
+                  reason: 'Do not talk about bananas!'
+        overrides:
+        - level: '>=50'
+          config:
+            rules:
+              my_filter:
+                enabled: false
+      ~~~
+      
+      ### Spam detection
+      This example includes 2 filters:
+      
+      - The first one is triggered if a user sends 5 messages within 10 seconds OR 3 attachments within 60 seconds.
+        The messages are deleted and the user is muted for 5 minutes.
+      - The second filter is triggered if a user sends more than 2 emoji within 5 seconds.
+        The messages are deleted but the user is not muted.
+      
+      Moderators are ignored by both filters based on the override.
+      
+      ~~~yml
+      automod:
+        config:
+          rules:
+            my_spam_filter:
+              triggers:
+              - message_spam:
+                  amount: 5
+                  within: 10s
+              - attachment_spam:
+                  amount: 3
+                  within: 60s
+              actions:
+                clean: true
+                mute:
+                  duration: 5m
+                  reason: 'Auto-muted for spam'
+            my_second_filter:
+              triggers:
+              - message_spam:
+                  amount: 5
+                  within: 10s
+              actions:
+                clean: true
+        overrides:
+        - level: '>=50'
+          config:
+            rules:
+              my_spam_filter:
+                enabled: false
+              my_second_filter:
+                enabled: false
+      ~~~
+      
+      ### Custom status alerts
+      This example sends an alert any time a user with a matching custom status sends a message.
+      
+      ~~~yml
+      automod:
+        config:
+          rules:
+            bad_custom_statuses:
+              triggers:
+              - match_words:
+                  words: ['banana']
+                  match_custom_status: true
+              actions:
+                alert:
+                  channel: "473087035574321152"
+                  text: |-
+                    Bad custom status on user <@!{user.id}>:
+                    {matchSummary}
+      ~~~
+    `),
+};