Merge branch 'master' into metrics

cogs/emote.py

@@ -76,7 +76,7 @@ class Emotes(commands.Cog):
 					+ self.bot.http.user_agent
 			})
 
-		self.emote_client = EmoteClient(token=self.bot.config['tokens']['discord'])
+		self.emote_client = EmoteClient(self.bot)
 
 		with open('data/ec-emotes-final.json') as f:
 			self.ec_emotes = json.load(f)
@@ -444,7 +444,7 @@ class Emotes(commands.Cog):
 		image_data = await utils.image.resize_in_subprocess(image_data)
 		if reason is None:
 			reason = 'Created by ' + utils.format_user(self.bot, author_id)
-		return await self.emote_client.create(guild_id=guild.id, name=name, image=image_data, reason=reason)
+		return await self.emote_client.create(guild=guild, name=name, image=image_data, reason=reason)
 
 	@commands.command(aliases=('delete', 'delet', 'rm'))
 	async def remove(self, context, emote, *emotes):

requirements.txt

@@ -1,4 +1,3 @@
-aioec>=0.6.0
 aiohttp_socks
 bot_bin>=1.5.0,<2.0.0
 discord.py>=1.5.0,<2.0.0

utils/emote_client.py

@@ -53,22 +53,18 @@ class EmoteClient:
 		HTTPStatus.SERVICE_UNAVAILABLE: DiscordServerError,
 	}
 
-	def __init__(self, *, token):
+	def __init__(self, bot):
 		self.guild_rls: Dict[GuildId, GuildRetryTimes] = {}
-		user_agent = (
-			'EmoteManager-EmoteClient; '
-			f'aiohttp/{aiohttp.__version__}; '
-			f'{platform.python_implementation()}/{".".join(map(str, sys.version_info))}'
-		)
 		self.http = aiohttp.ClientSession(headers={
-			'User-Agent': user_agent,
-			'Authorization': 'Bot ' + token,
+			'User-Agent': bot.config['user_agent'] + ' ' + bot.http.user_agent,
+			'Authorization': 'Bot ' + bot.config['tokens']['discord'],
 			'X-Ratelimit-Precision': 'millisecond',
 		})
+		# internals 🤫
+		self._connection_state = bot._connection
 
 	async def request(self, method, path, guild_id, **kwargs):
 		self._check_rl(method, guild_id)
-		print('post check rl')
 
 		headers = {}
 		# Emote Manager shouldn't use walrus op until Debian adopts 3.8 :(
@@ -80,7 +76,6 @@ class EmoteClient:
 		# TODO handle OSError and 500/502, like dpy does
 		async with self.http.request(method, self.BASE_URL + path, **kwargs) as resp:
 			if resp.status == HTTPStatus.TOO_MANY_REQUESTS:
-				print('handling rl')
 				return await self._handle_rl(resp, method, path, guild_id, **kwargs)
 
 			data = await json_or_text(resp)
@@ -94,17 +89,14 @@ class EmoteClient:
 		try:
 			rls = self.guild_rls[guild_id]
 		except KeyError:
-			print('guild not found', repr(guild_id))
 			return
 
 		if not rls.validate():
-			print('guild rls invalid')
 			del self.guild_rls[guild_id]
 			return
 
 		retry_at = getattr(rls, method, None)
 		if retry_at:
-			print('retry later')
 			raise RateLimitedError(retry_at)
 
 	async def _handle_rl(self, resp, method, path, guild_id, **kwargs):
@@ -134,13 +126,17 @@ class EmoteClient:
 	def check_delete(self, guild_id):
 		self._check_rl('DELETE', guild_id)
 
-	async def create(self, *, guild_id, name, image: bytes, role_ids=(), reason=None):
-		return await self.request(
-			'POST', f'/guilds/{guild_id}/emojis',
-			guild_id,
+	async def create(self, *, guild, name, image: bytes, role_ids=(), reason=None):
+		data = await self.request(
+			'POST', f'/guilds/{guild.id}/emojis',
+			guild.id,
 			json=dict(name=name, image=image_utils.image_to_base64_url(image), roles=role_ids),
 			reason=reason,
 		)
+		# internals 🤫
+		# this is A) so we can return a bona-fide, authentic, Emoji object,
+		# and B) because it's what dpy does to keep the emoji cache up to date for s
+		return self._connection_state.store_emoji(guild, data)
 
 	async def delete(self, *, guild_id, emote_id, reason=None):
 		return await self.request('DELETE', f'/guilds/{guild_id}/emojis/{emote_id}', guild_id, reason=reason)