Begin work on a data request API

docs/framework_commands.rst

@@ -13,6 +13,14 @@ extend functionalities used throughout the bot, as outlined below.
 
 .. autofunction:: redbot.core.commands.group
 
+.. autoclass:: redbot.core.commands.Cog
+
+    .. automethod:: format_help_for_context
+
+    .. automethod:: red_get_data_for_user
+
+    .. automethod:: red_delete_data_for_user
+
 .. autoclass:: redbot.core.commands.Command
     :members:
     :inherited-members: format_help_for_context

docs/guide_cog_creation.rst

@@ -98,6 +98,7 @@ Open :code:`__init__.py`. In that file, place the following:
 
     from .mycog import Mycog
 
+
     def setup(bot):
         bot.add_cog(Mycog())
 
@@ -238,3 +239,20 @@ Not all of these are strict requirements (some are) but are all generally advisa
     but a cog which takes actions based on messages should not.
 
 15. Respect settings when treating non command messages as commands.
+
+16. Handle user data responsibly
+
+  - Don't do unexpected things with user data.
+  - Don't expose user data to additional audiences without permission.
+  - Don't collect data your cogs don't need.
+  - Don't store data in unexpected locations.
+    Utilize the cog data path, Config, or if you need something more
+    prompt the owner to provide it.
+
+17. Utilize the data deletion and statement APIs
+
+  - See `redbot.core.commands.Cog.red_delete_data_for_user`
+  - Make a statement about what data your cogs use with the module level
+    variable ``__red_end_user_data_statement__``.
+    This should be a string containing a user friendly explanation of what data
+    your cog stores and why.

docs/index.rst

@@ -31,6 +31,7 @@ Welcome to Red - Discord Bot's documentation!
     :caption: User guides:
 
     getting_started
+    red_core_data_statement
 
 .. toctree::
     :maxdepth: 2

docs/red_core_data_statement.rst

@@ -0,0 +1,87 @@
+.. Red Core Data Statement
+
+=====================
+Red and End User Data
+=====================
+
+Notes for everyone
+******************
+
+What data Red collects
+----------------------
+
+Red and the cogs included with it collect some amount of data
+about users for the bot's normal operations. 
+
+In particular the bot will keep track of a short history of usernames/nicknames
+which actions refer to your Discord account (such as creating a playlist)
+as well as the content of specific messages used directly as commands for the bot
+(such as reports sent to servers).
+
+By default, Red will not collect any more data than it needs, and will not use it
+for anything other than the portion of the Red's functionality that necessitated it.
+
+3rd party extensions may store additional data beyond what Red does by default.
+You can use the command ``[p]mydata 3rdparty``
+to view statements about how extensions use your data made by the authors of 
+the specific 3rd party extensions an instance of Red has installed.
+
+How can I delete data Red has about me?
+---------------------------------------
+
+The command ``[p]mydata forgetme`` provides a way for users to remove
+large portions of their own data from the bot. This command will not
+remove operational data, such as a record that your
+Discord account was the target of a moderation action.
+
+3rd party extensions to Red are able to delete data when this command
+is used as well, but this is something each extension must implement.
+If a loaded extension does not implement this, the user will be informed.
+
+Additional Notes for Bot Owners and Hosts
+*****************************************
+
+How to comply with a request from Discord Trust & Safety
+--------------------------------------------------------
+
+There are a handful of these available to bot owners in the command group
+``[p]mydata ownermanagement``.
+
+The most pertinent one if asked to delete data by a member of Trust & Safety
+is
+
+``[p]mydata ownermanagement processdiscordrequest`` 
+
+This will cause the bot to get rid of or disassociate all data
+from the specified user ID. 
+
+.. warning::
+
+    You should not use this unless
+    Discord has specifically requested this with regard to a deleted user.
+    This will remove the user from various anti-abuse measures.
+    If you are processing a manual request from a user, read the next section
+
+
+How to process deletion requests from users
+-------------------------------------------
+
+You can point users to the command ``[p]mydata forgetme`` as a first step.
+
+If users cannot use that for some reason, the command
+
+``[p]mydata ownermanagement deleteforuser``
+
+exists as a way to handle this as if the user had done it themselves.
+
+Be careful about using the other owner level deletion options on behalf of users,
+as this may also result in losing operational data such as data used to prevent spam.
+
+What owners and hosts are responsible for
+-----------------------------------------
+
+Owners and hosts must comply both with Discord's terms of service and any applicable laws.
+Owners and hosts are responsible for all actions their bot takes.
+
+We cannot give specific guidance on this, but recommend that if there are any issues
+you be forthright with users, own up to any mistakes, and do your best to handle it.

redbot/cogs/admin/admin.py

@@ -87,6 +87,10 @@ def __init__(self):
     async def cog_before_invoke(self, ctx: commands.Context):
         await self._ready.wait()
 
+    async def red_delete_data_for_user(self, **kwargs):
+        """ Nothing to delete """
+        return
+
     async def handle_migrations(self):
 
         lock = self.config.get_guilds_lock()

redbot/cogs/alias/__init__.py

@@ -4,5 +4,5 @@
 
 async def setup(bot: Red):
     cog = Alias(bot)
-    await cog.initialize()
     bot.add_cog(cog)
+    cog.sync_init()

redbot/cogs/alias/alias.py

@@ -1,7 +1,9 @@
+import asyncio
+import logging
 from copy import copy
 from re import search
 from string import Formatter
-from typing import Dict, List
+from typing import Dict, List, Literal
 
 import discord
 from redbot.core import Config, commands, checks
@@ -14,6 +16,8 @@
 
 _ = Translator("Alias", __file__)
 
+log = logging.getLogger("red.cogs.alias")
+
 
 class _TrackingFormatter(Formatter):
     def __init__(self):
@@ -38,24 +42,107 @@ class Alias(commands.Cog):
     and append them to the stored alias.
     """
 
-    default_global_settings: Dict[str, list] = {"entries": []}
-
-    default_guild_settings: Dict[str, list] = {"entries": []}  # Going to be a list of dicts
-
     def __init__(self, bot: Red):
         super().__init__()
         self.bot = bot
         self.config = Config.get_conf(self, 8927348724)
 
-        self.config.register_global(**self.default_global_settings)
-        self.config.register_guild(**self.default_guild_settings)
+        self.config.register_global(entries=[], handled_string_creator=False)
+        self.config.register_guild(entries=[])
         self._aliases: AliasCache = AliasCache(config=self.config, cache_enabled=True)
+        self._ready_event = asyncio.Event()
+
+    async def red_delete_data_for_user(
+        self,
+        *,
+        requester: Literal["discord_deleted_user", "owner", "user", "user_strict"],
+        user_id: int,
+    ):
+        if requester != "discord_deleted_user":
+            return
+
+        await self._ready_event.wait()
+        await self._aliases.anonymize_aliases(user_id)
+
+    async def cog_before_invoke(self, ctx):
+        await self._ready_event.wait()
+
+    async def _maybe_handle_string_keys(self):
+        # This isn't a normal schema migration because it's being added
+        # after the fact for GH-3788
+        if await self.config.handled_string_creator():
+            return
+
+        async with self.config.entries() as alias_list:
+            bad_aliases = []
+            for a in alias_list:
+                for keyname in ("creator", "guild"):
+                    if isinstance((val := a.get(keyname)), str):
+                        try:
+                            a[keyname] = int(val)
+                        except ValueError:
+                            # Because migrations weren't created as changes were made,
+                            # and the prior form was a string of an ID,
+                            # if this fails, there's nothing to go back to
+                            bad_aliases.append(a)
+                            break
+
+            for a in bad_aliases:
+                alias_list.remove(a)
+
+        # if this was using a custom group of (guild_id, aliasname) it would be better but...
+        all_guild_aliases = await self.config.all_guilds()
+
+        for guild_id, guild_data in all_guild_aliases.items():
+
+            to_set = []
+            modified = False
+
+            for a in guild_data.get("entries", []):
+
+                for keyname in ("creator", "guild"):
+                    if isinstance((val := a.get(keyname)), str):
+                        try:
+                            a[keyname] = int(val)
+                        except ValueError:
+                            break
+                        finally:
+                            modified = True
+                else:
+                    to_set.append(a)
+
+            if modified:
+                await self.config.guild_from_id(guild_id).entries.set(to_set)
+
+            await asyncio.sleep(0)
+            # control yielded per loop since this is most likely to happen
+            # at bot startup, where this is most likely to have a performance
+            # hit.
+
+        await self.config.handled_string_creator.set(True)
+
+    def sync_init(self):
+        t = asyncio.create_task(self._initialize())
+
+        def done_callback(fut: asyncio.Future):
+            try:
+                t.result()
+            except Exception as exc:
+                log.exception("Failed to load alias cog", exc_info=exc)
+                # Maybe schedule extension unloading with message to owner in future
+
+        t.add_done_callback(done_callback)
+
+    async def _initialize(self):
+        """ Should only ever be a task """
+
+        await self._maybe_handle_string_keys()
 
-    async def initialize(self):
-        # This can be where we set the cache_enabled attribute later
         if not self._aliases._loaded:
             await self._aliases.load_aliases()
 
+        self._ready_event.set()
+
     def is_command(self, alias_name: str) -> bool:
         """
         The logic here is that if this returns true, the name should not be used for an alias
@@ -327,6 +414,8 @@ async def _list_global_alias(self, ctx: commands.Context):
     @commands.Cog.listener()
     async def on_message_without_command(self, message: discord.Message):
 
+        await self._ready_event.wait()
+
         if message.guild is not None:
             if await self.bot.cog_disabled_in_guild(self, message.guild):
                 return

redbot/cogs/alias/alias_entry.py

@@ -90,6 +90,30 @@ def __init__(self, config: Config, cache_enabled: bool = True):
         self._loaded = False
         self._aliases: Dict[Optional[int], Dict[str, AliasEntry]] = {None: {}}
 
+    async def anonymize_aliases(self, user_id: int):
+
+        async with self.config.entries() as global_aliases:
+            for a in global_aliases:
+                if a.get("creator", 0) == user_id:
+                    a["creator"] = 0xDE1
+                    if self._cache_enabled:
+                        self._aliases[None][a["name"]] = AliasEntry.from_json(a)
+
+        all_guilds = await self.config.all_guilds()
+        async for guild_id, guild_data in AsyncIter(all_guilds.items(), steps=100):
+            for a in guild_data["entries"]:
+                if a.get("creator", 0) == user_id:
+                    break
+            else:
+                continue
+            # basically, don't build a context manager wihout a need.
+            async with self.config.guild_from_id(guild_id).entries() as entry_list:
+                for a in entry_list:
+                    if a.get("creator", 0) == user_id:
+                        a["creator"] = 0xDE1
+                        if self._cache_enabled:
+                            self._aliases[guild_id][a["name"]] = AliasEntry.from_json(a)
+
     async def load_aliases(self):
         if not self._cache_enabled:
             self._loaded = True

redbot/cogs/audio/apis/playlist_wrapper.py

@@ -27,6 +27,7 @@
     PRAGMA_SET_read_uncommitted,
     PRAGMA_SET_temp_store,
     PRAGMA_SET_user_version,
+    HANDLE_DISCORD_DATA_DELETION_QUERY,
 )
 from ..utils import PlaylistScope
 from .api_utils import PlaylistFetchResult
@@ -58,6 +59,8 @@ def __init__(self, bot: Red, config: Config, conn: APSWConnectionWrapper):
         self.statement.get_all_with_filter = PLAYLIST_FETCH_ALL_WITH_FILTER
         self.statement.get_all_converter = PLAYLIST_FETCH_ALL_CONVERTER
 
+        self.statement.drop_user_playlists = HANDLE_DISCORD_DATA_DELETION_QUERY
+
     async def init(self) -> None:
         """Initialize the Playlist table"""
         with concurrent.futures.ThreadPoolExecutor(max_workers=1) as executor:
@@ -247,3 +250,11 @@ async def upsert(
                     "tracks": json.dumps(tracks),
                 },
             )
+
+    async def handle_playlist_user_id_deletion(self, user_id: int):
+        with concurrent.futures.ThreadPoolExecutor(max_workers=1) as executor:
+            executor.submit(
+                self.database.cursor().execute,
+                self.statement.drop_user_playlists,
+                {"user_id": user_id},
+            )