mirror of
https://github.com/LonamiWebs/Telethon.git
synced 2024-11-29 12:53:44 +03:00
9965cda968
This way it is easy and cheap to copy the two required values to all incoming updates in case we need to getDifference since the previous pts/date to fetch entities. This is still a work in progress.
439 lines
16 KiB
Python
439 lines
16 KiB
Python
import asyncio
|
|
import itertools
|
|
import random
|
|
import time
|
|
import datetime
|
|
|
|
from .users import UserMethods
|
|
from .. import events, utils, errors
|
|
from ..tl import types, functions
|
|
from ..events.common import EventCommon
|
|
|
|
|
|
class UpdateMethods(UserMethods):
|
|
|
|
# region Public methods
|
|
|
|
async def _run_until_disconnected(self):
|
|
try:
|
|
await self.disconnected
|
|
except KeyboardInterrupt:
|
|
pass
|
|
finally:
|
|
await self.disconnect()
|
|
|
|
def run_until_disconnected(self):
|
|
"""
|
|
Runs the event loop until `disconnect` is called or if an error
|
|
while connecting/sending/receiving occurs in the background. In
|
|
the latter case, said error will ``raise`` so you have a chance
|
|
to ``except`` it on your own code.
|
|
|
|
If the loop is already running, this method returns a coroutine
|
|
that you should await on your own code.
|
|
"""
|
|
if self.loop.is_running():
|
|
return self._run_until_disconnected()
|
|
try:
|
|
return self.loop.run_until_complete(self.disconnected)
|
|
except KeyboardInterrupt:
|
|
pass
|
|
finally:
|
|
# No loop.run_until_complete; it's already syncified
|
|
self.disconnect()
|
|
|
|
def on(self, event):
|
|
"""
|
|
Decorator helper method around `add_event_handler`. Example:
|
|
|
|
>>> from telethon import TelegramClient, events
|
|
>>> client = TelegramClient(...)
|
|
>>>
|
|
>>> @client.on(events.NewMessage)
|
|
... async def handler(event):
|
|
... ...
|
|
...
|
|
>>>
|
|
|
|
Args:
|
|
event (`_EventBuilder` | `type`):
|
|
The event builder class or instance to be used,
|
|
for instance ``events.NewMessage``.
|
|
"""
|
|
def decorator(f):
|
|
self.add_event_handler(f, event)
|
|
return f
|
|
|
|
return decorator
|
|
|
|
def add_event_handler(self, callback, event=None):
|
|
"""
|
|
Registers the given callback to be called on the specified event.
|
|
|
|
Args:
|
|
callback (`callable`):
|
|
The callable function accepting one parameter to be used.
|
|
|
|
Note that if you have used `telethon.events.register` in
|
|
the callback, ``event`` will be ignored, and instead the
|
|
events you previously registered will be used.
|
|
|
|
event (`_EventBuilder` | `type`, optional):
|
|
The event builder class or instance to be used,
|
|
for instance ``events.NewMessage``.
|
|
|
|
If left unspecified, `telethon.events.raw.Raw` (the
|
|
:tl:`Update` objects with no further processing) will
|
|
be passed instead.
|
|
"""
|
|
builders = events._get_handlers(callback)
|
|
if builders is not None:
|
|
for event in builders:
|
|
self._event_builders.append((event, callback))
|
|
return
|
|
|
|
if isinstance(event, type):
|
|
event = event()
|
|
elif not event:
|
|
event = events.Raw()
|
|
|
|
self._event_builders.append((event, callback))
|
|
|
|
def remove_event_handler(self, callback, event=None):
|
|
"""
|
|
Inverse operation of :meth:`add_event_handler`.
|
|
|
|
If no event is given, all events for this callback are removed.
|
|
Returns how many callbacks were removed.
|
|
"""
|
|
found = 0
|
|
if event and not isinstance(event, type):
|
|
event = type(event)
|
|
|
|
i = len(self._event_builders)
|
|
while i:
|
|
i -= 1
|
|
ev, cb = self._event_builders[i]
|
|
if cb == callback and (not event or isinstance(ev, event)):
|
|
del self._event_builders[i]
|
|
found += 1
|
|
|
|
return found
|
|
|
|
def list_event_handlers(self):
|
|
"""
|
|
Lists all added event handlers, returning a list of pairs
|
|
consisting of (callback, event).
|
|
"""
|
|
return [(callback, event) for event, callback in self._event_builders]
|
|
|
|
async def catch_up(self):
|
|
"""
|
|
"Catches up" on the missed updates while the client was offline.
|
|
You should call this method after registering the event handlers
|
|
so that the updates it loads can by processed by your script.
|
|
|
|
This can also be used to forcibly fetch new updates if there are any.
|
|
"""
|
|
# TODO Since which state should we catch up?
|
|
if all(self._new_pts_date):
|
|
pts, date = self._new_pts_date
|
|
elif all(self._old_pts_date):
|
|
pts, date = self._new_pts_date
|
|
else:
|
|
return
|
|
|
|
self.session.catching_up = True
|
|
try:
|
|
while True:
|
|
d = await self(functions.updates.GetDifferenceRequest(
|
|
pts, date, 0
|
|
))
|
|
if isinstance(d, (types.updates.DifferenceSlice,
|
|
types.updates.Difference)):
|
|
if isinstance(d, types.updates.Difference):
|
|
state = d.state
|
|
else:
|
|
state = d.intermediate_state
|
|
|
|
pts, date = state.pts, state.date
|
|
self._handle_update(types.Updates(
|
|
users=d.users,
|
|
chats=d.chats,
|
|
date=state.date,
|
|
seq=state.seq,
|
|
updates=d.other_updates + [
|
|
types.UpdateNewMessage(m, 0, 0)
|
|
for m in d.new_messages
|
|
]
|
|
))
|
|
|
|
# TODO Implement upper limit (max_pts)
|
|
# We don't want to fetch updates we already know about.
|
|
#
|
|
# We may still get duplicates because the Difference
|
|
# contains a lot of updates and presumably only has
|
|
# the state for the last one, but at least we don't
|
|
# unnecessarily fetch too many.
|
|
#
|
|
# updates.getDifference's pts_total_limit seems to mean
|
|
# "how many pts is the request allowed to return", and
|
|
# if there is more than that, it returns "too long" (so
|
|
# there would be duplicate updates since we know about
|
|
# some). This can be used to detect collisions (i.e.
|
|
# it would return an update we have already seen).
|
|
else:
|
|
if isinstance(d, types.updates.DifferenceEmpty):
|
|
date = d.date
|
|
elif isinstance(d, types.updates.DifferenceTooLong):
|
|
pts = d.pts
|
|
break
|
|
except (ConnectionError, asyncio.CancelledError):
|
|
pass
|
|
finally:
|
|
# TODO Save new pts to session
|
|
self._new_pts_date = (pts, date)
|
|
self.session.catching_up = False
|
|
|
|
# endregion
|
|
|
|
# region Private methods
|
|
|
|
# It is important to not make _handle_update async because we rely on
|
|
# the order that the updates arrive in to update the pts and date to
|
|
# be always-increasing. There is also no need to make this async.
|
|
def _handle_update(self, update):
|
|
self.session.process_entities(update)
|
|
self._entity_cache.add(update)
|
|
|
|
if isinstance(update, (types.Updates, types.UpdatesCombined)):
|
|
entities = {utils.get_peer_id(x): x for x in
|
|
itertools.chain(update.users, update.chats)}
|
|
for u in update.updates:
|
|
self._process_update(u, entities)
|
|
|
|
self._new_pts_date = (self._new_pts_date[0], update.date)
|
|
elif isinstance(update, types.UpdateShort):
|
|
self._process_update(update.update)
|
|
self._new_pts_date = (self._new_pts_date[0], update.date)
|
|
else:
|
|
self._process_update(update)
|
|
|
|
# TODO Should this be done before or after?
|
|
self._update_pts_date(update)
|
|
|
|
def _process_update(self, update, entities=None):
|
|
update._entities = entities or {}
|
|
if self._updates_queue is None:
|
|
self._loop.create_task(self._dispatch_update(update))
|
|
else:
|
|
self._updates_queue.put_nowait(update)
|
|
if not self._dispatching_updates_queue.is_set():
|
|
self._dispatching_updates_queue.set()
|
|
self._loop.create_task(self._dispatch_queue_updates())
|
|
|
|
self._update_pts_date(update)
|
|
|
|
def _update_pts_date(self, update):
|
|
pts, date = self._new_pts_date
|
|
if getattr(update, 'pts', None):
|
|
pts = update.pts
|
|
|
|
if getattr(update, 'date', None):
|
|
date = update.date
|
|
|
|
self._new_pts_date = (pts, date)
|
|
|
|
async def _update_loop(self):
|
|
# Pings' ID don't really need to be secure, just "random"
|
|
rnd = lambda: random.randrange(-2**63, 2**63)
|
|
while self.is_connected():
|
|
try:
|
|
await asyncio.wait_for(
|
|
self.disconnected, timeout=60, loop=self._loop
|
|
)
|
|
continue # We actually just want to act upon timeout
|
|
except asyncio.TimeoutError:
|
|
pass
|
|
except asyncio.CancelledError:
|
|
return
|
|
except Exception:
|
|
continue # Any disconnected exception should be ignored
|
|
|
|
# We also don't really care about their result.
|
|
# Just send them periodically.
|
|
try:
|
|
self._sender.send(functions.PingRequest(rnd()))
|
|
except (ConnectionError, asyncio.CancelledError):
|
|
return
|
|
|
|
# Entities and cached files are not saved when they are
|
|
# inserted because this is a rather expensive operation
|
|
# (default's sqlite3 takes ~0.1s to commit changes). Do
|
|
# it every minute instead. No-op if there's nothing new.
|
|
self.session.save()
|
|
|
|
# We need to send some content-related request at least hourly
|
|
# for Telegram to keep delivering updates, otherwise they will
|
|
# just stop even if we're connected. Do so every 30 minutes.
|
|
#
|
|
# TODO Call getDifference instead since it's more relevant
|
|
if time.time() - self._last_request > 30 * 60:
|
|
if not await self.is_user_authorized():
|
|
# What can be the user doing for so
|
|
# long without being logged in...?
|
|
continue
|
|
|
|
try:
|
|
await self(functions.updates.GetStateRequest())
|
|
except (ConnectionError, asyncio.CancelledError):
|
|
return
|
|
|
|
async def _dispatch_queue_updates(self):
|
|
while not self._updates_queue.empty():
|
|
await self._dispatch_update(self._updates_queue.get_nowait())
|
|
|
|
self._dispatching_updates_queue.clear()
|
|
|
|
async def _dispatch_update(self, update):
|
|
built = EventBuilderDict(self, update)
|
|
if self._conversations:
|
|
for conv in self._conversations.values():
|
|
if await built.get(events.NewMessage):
|
|
conv._on_new_message(built[events.NewMessage])
|
|
if await built.get(events.MessageEdited):
|
|
conv._on_edit(built[events.MessageEdited])
|
|
if await built.get(events.MessageRead):
|
|
conv._on_read(built[events.MessageRead])
|
|
if conv._custom:
|
|
await conv._check_custom(built)
|
|
|
|
for builder, callback in self._event_builders:
|
|
event = await built.get(type(builder))
|
|
if not event:
|
|
continue
|
|
|
|
if not builder.resolved:
|
|
await builder.resolve(self)
|
|
|
|
if not builder.filter(event):
|
|
continue
|
|
|
|
try:
|
|
await callback(event)
|
|
except errors.AlreadyInConversationError:
|
|
name = getattr(callback, '__name__', repr(callback))
|
|
self._log[__name__].debug(
|
|
'Event handler "%s" already has an open conversation, '
|
|
'ignoring new one', name)
|
|
except events.StopPropagation:
|
|
name = getattr(callback, '__name__', repr(callback))
|
|
self._log[__name__].debug(
|
|
'Event handler "%s" stopped chain of propagation '
|
|
'for event %s.', name, type(event).__name__
|
|
)
|
|
break
|
|
except Exception:
|
|
name = getattr(callback, '__name__', repr(callback))
|
|
self._log[__name__].exception('Unhandled exception on %s',
|
|
name)
|
|
|
|
async def _handle_auto_reconnect(self):
|
|
# Upon reconnection, we want to send getState
|
|
# for Telegram to keep sending us updates.
|
|
try:
|
|
self._log[__name__].info(
|
|
'Asking for the current state after reconnect...')
|
|
|
|
# TODO consider:
|
|
# If there aren't many updates while the client is disconnected
|
|
# (I tried with up to 20), Telegram seems to send them without
|
|
# asking for them (via updates.getDifference).
|
|
#
|
|
# On disconnection, the library should probably set a "need
|
|
# difference" or "catching up" flag so that any new updates are
|
|
# ignored, and then the library should call updates.getDifference
|
|
# itself to fetch them.
|
|
#
|
|
# In any case (either there are too many updates and Telegram
|
|
# didn't send them, or there isn't a lot and Telegram sent them
|
|
# but we dropped them), we fetch the new difference to get all
|
|
# missed updates. I feel like this would be the best solution.
|
|
|
|
# If a disconnection occurs, the old known state will be
|
|
# the latest one we were aware of, so we can catch up since
|
|
# the most recent state we were aware of.
|
|
await self.catch_up()
|
|
|
|
self._log[__name__].info('Successfully fetched missed updates')
|
|
except errors.RPCError as e:
|
|
self._log[__name__].warning('Failed to get missed updates after '
|
|
'reconnect: %r', e)
|
|
except Exception:
|
|
self._log[__name__].exception('Unhandled exception while getting '
|
|
'update difference after reconnect')
|
|
|
|
# endregion
|
|
|
|
|
|
class EventBuilderDict:
|
|
"""
|
|
Helper "dictionary" to return events from types and cache them.
|
|
"""
|
|
def __init__(self, client, update):
|
|
self.client = client
|
|
self.update = update
|
|
|
|
def __getitem__(self, builder):
|
|
return self.__dict__[builder]
|
|
|
|
async def get(self, builder):
|
|
try:
|
|
return self.__dict__[builder]
|
|
except KeyError:
|
|
event = self.__dict__[builder] = builder.build(self.update)
|
|
if isinstance(event, EventCommon):
|
|
event.original_update = self.update
|
|
event._set_client(self.client)
|
|
if not event._load_entities():
|
|
await self.get_difference()
|
|
if not event._load_entities():
|
|
self.client._log[__name__].info(
|
|
'Could not find all entities for update.pts = %s',
|
|
getattr(self.update, 'pts', None)
|
|
)
|
|
elif event:
|
|
# Actually a :tl:`Update`, not much processing to do
|
|
event._client = self.client
|
|
|
|
return event
|
|
|
|
async def get_difference(self):
|
|
"""
|
|
Calls :tl:`updates.getDifference`, which fills the entities cache
|
|
(always done by `__call__`) and lets us know about the full entities.
|
|
"""
|
|
pts = getattr(self.update, 'pts', None)
|
|
if not pts:
|
|
return
|
|
|
|
date = getattr(self.update, 'date', None)
|
|
if date:
|
|
# Get the difference from one second ago to now
|
|
date -= datetime.timedelta(seconds=1)
|
|
else:
|
|
# No date known, 1 is the earliest date that works
|
|
date = 1
|
|
|
|
self.client._log[__name__].debug('Getting difference for entities')
|
|
result = await self.client(functions.updates.GetDifferenceRequest(
|
|
pts - 1, date, 0
|
|
))
|
|
|
|
if isinstance(result, (types.updates.Difference,
|
|
types.updates.DifferenceSlice)):
|
|
self.update._entities.update({
|
|
utils.get_peer_id(x): x for x in
|
|
itertools.chain(result.users, result.chats)
|
|
})
|