python/aiogram
1
0
Fork 0
mirror of https://github.com/aiogram/aiogram.git synced 2026-04-08 16:37:47 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

Update migration guide

Browse source
This commit is contained in:
Alex Root Junior 2023-03-12 03:03:02 +02:00
parent d4233025b1
commit d42d4e97f4
No known key found for this signature in database
GPG key ID: 074C1D455EBEA4AC
5 changed files with 311 additions and 319 deletions
Download patch file Download diff file Expand all files Collapse all files

44
docs/api/methods/set_sticker_set_thumb.rst
View file

@ -1,44 +0,0 @@
##################
setStickerSetThumb
##################
Returns: :obj:`bool`
.. automodule:: aiogram.methods.set_sticker_set_thumb
:members:
:member-order: bysource
:undoc-members: True
Usage
=====
As bot method
-------------
.. code-block::
result: bool = await bot.set_sticker_set_thumb(...)
Method as object
----------------
Imports:
- :code:`from aiogram.methods.set_sticker_set_thumb import SetStickerSetThumb`
- alias: :code:`from aiogram.methods import SetStickerSetThumb`
With specific bot
~~~~~~~~~~~~~~~~~
.. code-block:: python
result: bool = await bot(SetStickerSetThumb(...))
As reply into Webhook in handler
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
.. code-block:: python
return SetStickerSetThumb(...)

2
docs/dispatcher/filters/callback_data.rst
View file

@ -1,3 +1,5 @@
.. _callback-data-factory
==============================
Callback Data Factory & Filter
==============================

2
docs/dispatcher/middlewares.rst
View file

@ -1,3 +1,5 @@
.. _middlewares:
===========
Middlewares
===========

581
docs/migration_2_to_3.rst
View file

@ -2,346 +2,377 @@
Migration FAQ (2.x -> 3.0)
==========================
aiogram v3 introduced many breaking changes. This guide will help you to migrate from your existing v2.x code to new v3.0
.. danger::
This guide is still in progress.
This version introduces much many breaking changes and architectural improvements,
helping to reduce global variables count in your code, provides mechanisms useful mechanisms
to separate your code to modules or just make sharable modules via packages on the PyPi,
makes middlewares and filters more controllable and others.
On this page you can read about points that changed corresponding to last stable 2.x version.
Dispatcher
==========
- :class:`Dispatcher` class no longer accepts the `Bot` instance into the initializer, it should be passed to dispatcher only for starting polling or handling event from webhook.
- :class:`Dispatcher` now can be extended with another Dispatcher-like thing named :class:`Router` (:ref:`Read more » <Nested routers>`)
- Removed the **_handler** suffix from all event handler decorators and registering methods. (:ref:`Read more » <Event observers>`)
- :class:`Dispatcher` class no longer accepts the `Bot` instance into the initializer,
it should be passed to dispatcher only for starting polling or handling event from webhook.
Also this way adds possibility to use multiple bot instances at the same time ("multibot")
- :class:`Dispatcher` now can be extended with another Dispatcher-like
thing named :class:`Router` (:ref:`Read more » <Nested routers>`).
With routes you can easily separate your code to multiple modules
and maybe share this modules between projects.
- Removed the **_handler** suffix from all event handler decorators and registering methods.
(:ref:`Read more » <Event observers>`)
- Executor entirely removed, now you can use Dispatcher directly to start polling or webhook.
Filtering events
================
- Keyword filters can no more be used, use filters explicitly. (`Read more » <https://github.com/aiogram/aiogram/issues/942>`_)
- In due to keyword filters was removed all enabled by default filters (state and content_type now is not enabled)
- Most of common filters is replaced by "magic filter". (:ref:`Read more » <magic-filters>`)
- Now by default message handler receives any content type, if you want specific one just add the filters (Magic or any other)
- State filter is not enabled by default, that's mean if you using :code:`state="*"` in v2
then you should not pass any state filter in v3, and vice versa,
if the state in v2 is not specified now you should specify the state.
- Added possibility to register per-router global filters, that helps to reduces
the number of repetitions in the code and makes easily way to control
for what each router will be used.
Bot API Methods
===============
- Now all API methods is classes with validation (via `pydantic <https://docs.pydantic.dev/>`_)
(all API calls is also available as methods in the Bot class)
- Increased pre-defined Enums count and moved into `aiogram.enums` sub-package.
- Separated HTTP client session into container that can be reused between different
Bot instances in the application.
Middlewares
===========
(:ref:`Read more » <magic-filters>`)
- Middlewares can now control a execution context, e.g. using context managers (:ref:`Read more » <middlewares>`)
- All contextual data now is shared between middlewares, filters and handlers to end-to-end use.
For example now you can easily pass some data into context inside middleware and
get it in the filters layer as the same way as in the handlers via keyword arguments.
- Added mechanism named **flags**, that helps to customize handler behavior
in conjunction with middlewares.
- Middlewares is now can control a execution context
-
Keyboard Markup
===============
Markups
=======
- Now :class:`aiogram.types.inline_keyboard_markup.InlineKeyboardMarkup`
and :class:`aiogram.types.reply_keyboard_markup.ReplyKeyboardMarkup` has no methods to extend it,
instead you have to use markup builders :class:`aiogram.utils.keyboard.ReplyKeyboardBuilder`
and :class:`aiogram.utils.keyboard.KeyboardBuilder` respectively
(:ref:`Read more » <keyboard-builder>`)
-
Callbacks data
==============
-
- Callback data factory now is strictly typed via `pydantic <https://docs.pydantic.dev/>`_ models
(:ref:`Read more » <callback-data-factory>`)
Finite State machine
====================
- State filter will no more added to all handlers, you will need to specify state if you want
Sending Files
=============
-
- From now you should wrap sending files into InputFile object before send instead of passing
IO object directly to the API method. (:ref:`Read more » <sending-files>`)
Webhook
=======
-
- Simplified aiohttp web app configuration
- By default added possibility to upload files when you use reply into webhook
..
This part of document will be removed
Draft
=====
.. code-block:: markdown
This part of document will be removed
## Keyboards and filters (part 1)
.. code-block:: markdown
- `(Reply|Inline)KeyboardMarkup` is no longer used for building keyboards via `add`/`insert`/`row`, use `(Reply|Inline)KeyboardBuilder` and `button` instead.
- `CallbackData` is now a base class, not a factory.
- Integrate `[magic-filter](https://pypi.org/project/magic-filter/)` into aiogram.
- Code for 2.x
## Keyboards and filters (part 1)
```python
from secrets import token_urlsafe
- `(Reply|Inline)KeyboardMarkup` is no longer used for building keyboards via `add`/`insert`/`row`, use `(Reply|Inline)KeyboardBuilder` and `button` instead.
- `CallbackData` is now a base class, not a factory.
- Integrate `[magic-filter](https://pypi.org/project/magic-filter/)` into aiogram.
- Code for 2.x
```python
from secrets import token_urlsafe
from aiogram import Bot, Dispatcher
from aiogram.types import (
CallbackQuery,
InlineKeyboardButton,
InlineKeyboardMarkup,
Message,
)
from aiogram.utils.callback_data import CallbackData
dp = Dispatcher(Bot(TOKEN))
vote_cb = CallbackData("vote", "action", "id", sep="_")
votes = {} # For demo purposes only! Use database in real code!
@dp.message_handler(commands="start")
async def post(message: Message) -> None:
vote_id = token_urlsafe(8) # Lazy way to generate a random string
kb = (
InlineKeyboardMarkup(row_width=2)
.insert(InlineKeyboardButton(text="+1", callback_data=vote_cb.new(action="up", id=vote_id)))
.insert(InlineKeyboardButton(text="-1", callback_data=vote_cb.new(action="down", id=vote_id)))
.insert(InlineKeyboardButton(text="?", callback_data=vote_cb.new(action="count", id=vote_id)))
from aiogram import Bot, Dispatcher
from aiogram.types import (
CallbackQuery,
InlineKeyboardButton,
InlineKeyboardMarkup,
Message,
)
await message.reply("Vote on this post", reply_markup=kb)
from aiogram.utils.callback_data import CallbackData
@dp.callback_query_handler(vote_cb.filter(action="count"))
async def show_voters_count(query: CallbackQuery, callback_data: dict) -> None:
vote_id = int(callback_data["id"])
votes[vote_id] = votes.setdefault(vote_id, 0) + 1
await query.answer(votes[vote_id], cache_time=1)
dp = Dispatcher(Bot(TOKEN))
@dp.callback_query_handler(vote_cb.filter()) # all other actions
async def vote(query: CallbackQuery, callback_data: dict) -> None:
if (action := callback_data["action"]) == "up":
d = 1
elif action == "down":
d = -1
else:
raise AssertionError(f"action action!r} is not implemented")
votes[int(callback_data["id"])] += d
await query.answer(f"{action.capitalize()}voted!")
```
vote_cb = CallbackData("vote", "action", "id", sep="_")
votes = {} # For demo purposes only! Use database in real code!
- Code for 3.0
@dp.message_handler(commands="start")
async def post(message: Message) -> None:
vote_id = token_urlsafe(8) # Lazy way to generate a random string
kb = (
InlineKeyboardMarkup(row_width=2)
.insert(InlineKeyboardButton(text="+1", callback_data=vote_cb.new(action="up", id=vote_id)))
.insert(InlineKeyboardButton(text="-1", callback_data=vote_cb.new(action="down", id=vote_id)))
.insert(InlineKeyboardButton(text="?", callback_data=vote_cb.new(action="count", id=vote_id)))
)
await message.reply("Vote on this post", reply_markup=kb)
```python
from enum import Enum
from secrets import token_urlsafe
@dp.callback_query_handler(vote_cb.filter(action="count"))
async def show_voters_count(query: CallbackQuery, callback_data: dict) -> None:
vote_id = int(callback_data["id"])
votes[vote_id] = votes.setdefault(vote_id, 0) + 1
await query.answer(votes[vote_id], cache_time=1)
from aiogram import Dispatcher, F
from aiogram.types import CallbackQuery, Message
from aiogram.dispatcher.filters.callback_data import CallbackData
from aiogram.utils.keyboard import InlineKeyboardBuilder
dp = Dispatcher()
votes = {} # For demo purposes only! Use database in real code!
class VoteAction(Enum):
UP = "up"
DOWN = "down"
COUNT = "count"
class VoteCallback(CallbackData, prefix="vote", sep="_"):
action: VoteAction # Yes, it also supports `Enum`s
id: str
@dp.message(commands="start")
async def post(message: Message) -> None:
vote_id = token_urlsafe(8) # Lazy way to generate a random string
kb = (
InlineKeyboardBuilder()
.button(text="+1", callback_data=VoteCallback(action=VoteAction.UP, id=vote_id))
.button(text="-1", callback_data=VoteCallback(action=VoteAction.DOWN, id=vote_id))
.button(text="?", callback_data=VoteCallback(action=VoteAction.COUNT, id=vote_id))
.adjust(2) # row_width=2
)
await message.reply("Vote on this post", reply_markup=kb.as_markup())
# `F` is a `MagicFilter` instance, see docs for `magic-filter` for more info
@dp.callback_query(VoteCallback.filter(F.action == VoteAction.COUNT))
async def show_voters_count(
query: CallbackQuery,
callback_data: VoteCallback, # Now it is the class itself, not a mysterious `dict`
) -> None:
vote_id = callback_data.id
votes[vote_id] = votes.setdefault(vote_id, 0) + 1
await query.answer(votes[vote_id], cache_time=1)
@dp.callback_query(VoteCallback.filter()) # all other actions
async def vote(query: CallbackQuery, callback_data: VoteCallback) -> None:
if callback_data.action == VoteAction.UP:
d = 1
elif callback_data.action == VoteAction.DOWN:
d = -1
else:
raise AssertionError(f"action {callback_data.action!r} is not implemented")
votes[callback_data.id] += d
await query.answer(f"{action.capitalize()}voted!")
```
## Code style
- Allow the code to be split into several files in a convenient way with `Router`s.
- Make `Dispatcher` a router with some special abilities.
- Remove `<event>_handler` in favor of `<event>` (e.g. `dp.message()` instead of `dp.message_handler()`)
- Code for 2.x (one of possible ways)
```python
from aiogram import Bot, Dispatcher, executor
from mybot import handlers
dp = Dispatcher(Bot(TOKEN))
handlers.hello.setup(dp)
...
executor.start_polling(dp, ...)
```
```python
from aiogram import Dispatcher
from aiogram.types import Message
# No way to use decorators :(
async def hello(message: Message) -> None:
await message.reply("Hello!")
async def goodbye(message: Message) -> None:
await message.reply("Bye!")
def setup(dp: Dispatcher) -> None:
dp.register_message_handler(hello, commands=["hello", "hi"])
dp.register_message_handler(goodbye, commands=["goodbye", "bye"])
# This list can become huge in a time, may be inconvenient
```
- Code for 3.0
```python
from aiogram import Bot, Dispatcher
from mybot import handlers
dp = Dispatcher()
# Any router can include a sub-router
dp.include_router(handlers.hello.router) # `Dispatcher` is a `Router` too
...
dp.run_polling(Bot(TOKEN)) # But it's special, e.g. it can `run_polling`
```
```python
from aiogram import Router
from aiogram.types import Message
router = Router()
# Event handler decorator is an event type itself without `_handler` suffix
@router.message(commands=["hello", "hi"]) # Yay, decorators!
async def hello(message: Message) -> None:
await message.reply("Hello!")
async def goodbye(message: Message) -> None:
await message.reply("Bye!")
# If you still prefer registering handlers without decorators, use this
router.message.register(goodbye, commands=["goodbye", "bye"])
```
## Webhooks and API methods
- All methods are classes now.
- Allow using Reply into webhook with polling. *Check whether it worked in 2.x*
- Webhook setup is more flexible ~~and complicated xd~~
## Exceptions
- No more specific exceptions, only by status code.
- Code for 2.x [todo]
```python
from asyncio import sleep
from aiogram import Bot, Dispatcher
from aiogram.dispatcher.filters import Command
from aiogram.types import Message
from aiogram.utils.exceptions import (
BadRequest,
BotBlocked,
RestartingTelegram,
RetryAfter,
)
dp = Dispatcher(Bot(TOKEN))
chats = set()
async def broadcaster(bot: Bot, chat: int, text: str) -> bool:
"""Broadcasts a message and returns whether it was sent"""
while True:
try:
await bot.send_message(chat, text)
except BotBlocked:
chats.discard(chat)
log.warning("Remove chat %d because bot was blocked", chat)
return False
except RetryAfter as e:
log.info("Sleeping %d due to flood wait", e.retry_after)
await sleep(e.retry_after)
continue
except RestartingTelegram:
log.info("Telegram is restarting, sleeping for 1 sec")
await sleep(1)
continue
except BadRequest as e:
log.warning("Remove chat %d because of bad request", chat)
chats.discard(chat)
return False
@dp.callback_query_handler(vote_cb.filter()) # all other actions
async def vote(query: CallbackQuery, callback_data: dict) -> None:
if (action := callback_data["action"]) == "up":
d = 1
elif action == "down":
d = -1
else:
return True
raise AssertionError(f"action action!r} is not implemented")
votes[int(callback_data["id"])] += d
await query.answer(f"{action.capitalize()}voted!")
```
@dp.message_handler(commands="broadcast")
async def broadcast(message: Message, command: Command.CommandObj) -> None:
# TODO ...
```
- Code for 3.0
- Code for 3.x [todo]
```python
from enum import Enum
from secrets import token_urlsafe
```python
...
```
from aiogram import Dispatcher, F
from aiogram.types import CallbackQuery, Message
from aiogram.dispatcher.filters.callback_data import CallbackData
from aiogram.utils.keyboard import InlineKeyboardBuilder
dp = Dispatcher()
votes = {} # For demo purposes only! Use database in real code!
class VoteAction(Enum):
UP = "up"
DOWN = "down"
COUNT = "count"
class VoteCallback(CallbackData, prefix="vote", sep="_"):
action: VoteAction # Yes, it also supports `Enum`s
id: str
@dp.message(commands="start")
async def post(message: Message) -> None:
vote_id = token_urlsafe(8) # Lazy way to generate a random string
kb = (
InlineKeyboardBuilder()
.button(text="+1", callback_data=VoteCallback(action=VoteAction.UP, id=vote_id))
.button(text="-1", callback_data=VoteCallback(action=VoteAction.DOWN, id=vote_id))
.button(text="?", callback_data=VoteCallback(action=VoteAction.COUNT, id=vote_id))
.adjust(2) # row_width=2
)
await message.reply("Vote on this post", reply_markup=kb.as_markup())
# `F` is a `MagicFilter` instance, see docs for `magic-filter` for more info
@dp.callback_query(VoteCallback.filter(F.action == VoteAction.COUNT))
async def show_voters_count(
query: CallbackQuery,
callback_data: VoteCallback, # Now it is the class itself, not a mysterious `dict`
) -> None:
vote_id = callback_data.id
votes[vote_id] = votes.setdefault(vote_id, 0) + 1
await query.answer(votes[vote_id], cache_time=1)
@dp.callback_query(VoteCallback.filter()) # all other actions
async def vote(query: CallbackQuery, callback_data: VoteCallback) -> None:
if callback_data.action == VoteAction.UP:
d = 1
elif callback_data.action == VoteAction.DOWN:
d = -1
else:
raise AssertionError(f"action {callback_data.action!r} is not implemented")
votes[callback_data.id] += d
await query.answer(f"{action.capitalize()}voted!")
```
## Filters (part 2)
## Code style
- Remove the majority of filters in favor of `MagicFilter` (aka `F`).
- Deprecate usage of bound filters in favor of classes, functions and `F`.
- Message handler defaults to any content type.
- Per-router filters.
- Allow the code to be split into several files in a convenient way with `Router`s.
- Make `Dispatcher` a router with some special abilities.
- Remove `<event>_handler` in favor of `<event>` (e.g. `dp.message()` instead of `dp.message_handler()`)
- Code for 2.x (one of possible ways)
## Middlewares and app state
```python
from aiogram import Bot, Dispatcher, executor
from mybot import handlers
- Rework middleware logic.
- Pass `**kwargs` from `start_polling` to handlers and filters.
- No more global `bot` and `message.bot`.
- `bot["foo"]` → `dp["foo"]`.
dp = Dispatcher(Bot(TOKEN))
## FSM
handlers.hello.setup(dp)
...
- FSMStrategy.
- Default to any state.
- States are also callable filters.
- No more `next` and `proxy`.
- No state filtering is done by default:
executor.start_polling(dp, ...)
```
[Default state is not None · Issue #954 · aiogram/aiogram](https://github.com/aiogram/aiogram/issues/954#issuecomment-1172967490)
```python
from aiogram import Dispatcher
from aiogram.types import Message
# No way to use decorators :(
async def hello(message: Message) -> None:
await message.reply("Hello!")
async def goodbye(message: Message) -> None:
await message.reply("Bye!")
def setup(dp: Dispatcher) -> None:
dp.register_message_handler(hello, commands=["hello", "hi"])
dp.register_message_handler(goodbye, commands=["goodbye", "bye"])
# This list can become huge in a time, may be inconvenient
```
- Code for 3.0
```python
from aiogram import Bot, Dispatcher
from mybot import handlers
dp = Dispatcher()
# Any router can include a sub-router
dp.include_router(handlers.hello.router) # `Dispatcher` is a `Router` too
...
dp.run_polling(Bot(TOKEN)) # But it's special, e.g. it can `run_polling`
```
```python
from aiogram import Router
from aiogram.types import Message
router = Router()
# Event handler decorator is an event type itself without `_handler` suffix
@router.message(commands=["hello", "hi"]) # Yay, decorators!
async def hello(message: Message) -> None:
await message.reply("Hello!")
async def goodbye(message: Message) -> None:
await message.reply("Bye!")
# If you still prefer registering handlers without decorators, use this
router.message.register(goodbye, commands=["goodbye", "bye"])
```
## Misc
## Webhooks and API methods
- No more unrelated attributes and methods in types.
- `get_args()`
- `get_(full_)command()`
- …?
- Add handler flags.
- ???
- All methods are classes now.
- Allow using Reply into webhook with polling. *Check whether it worked in 2.x*
- Webhook setup is more flexible ~~and complicated xd~~
## Exceptions
- No more specific exceptions, only by status code.
- Code for 2.x [todo]
```python
from asyncio import sleep
from aiogram import Bot, Dispatcher
from aiogram.dispatcher.filters import Command
from aiogram.types import Message
from aiogram.utils.exceptions import (
BadRequest,
BotBlocked,
RestartingTelegram,
RetryAfter,
)
dp = Dispatcher(Bot(TOKEN))
chats = set()
async def broadcaster(bot: Bot, chat: int, text: str) -> bool:
"""Broadcasts a message and returns whether it was sent"""
while True:
try:
await bot.send_message(chat, text)
except BotBlocked:
chats.discard(chat)
log.warning("Remove chat %d because bot was blocked", chat)
return False
except RetryAfter as e:
log.info("Sleeping %d due to flood wait", e.retry_after)
await sleep(e.retry_after)
continue
except RestartingTelegram:
log.info("Telegram is restarting, sleeping for 1 sec")
await sleep(1)
continue
except BadRequest as e:
log.warning("Remove chat %d because of bad request", chat)
chats.discard(chat)
return False
else:
return True
@dp.message_handler(commands="broadcast")
async def broadcast(message: Message, command: Command.CommandObj) -> None:
# TODO ...
```
- Code for 3.x [todo]
```python
...
```
## Filters (part 2)
- Remove the majority of filters in favor of `MagicFilter` (aka `F`).
- Deprecate usage of bound filters in favor of classes, functions and `F`.
- Message handler defaults to any content type.
- Per-router filters.
## Middlewares and app state
- Rework middleware logic.
- Pass `**kwargs` from `start_polling` to handlers and filters.
- No more global `bot` and `message.bot`.
- `bot["foo"]` → `dp["foo"]`.
## FSM
- FSMStrategy.
- Default to any state.
- States are also callable filters.
- No more `next` and `proxy`.
- No state filtering is done by default:
[Default state is not None · Issue #954 · aiogram/aiogram](https://github.com/aiogram/aiogram/issues/954#issuecomment-1172967490)
## Misc
- No more unrelated attributes and methods in types.
- `get_args()`
- `get_(full_)command()`
- …?
- Add handler flags.
- ???

1
docs/utils/keyboard.rst
View file

@ -1,3 +1,4 @@
.. _keyboard-builder
================
Keyboard builder
================