Community-based extensions for the python-telegram-bot library.

Hello again! In Pyrogram there's a nice feature that helps with some tasks like editing posts before publications: https://docs.pyrogram.org/api/bound-methods/Message.forward https://github.com/pyrogram/pyrogram/blob/develop/pyrogram/client/methods/messages/forward_messages.py

It just get the whole message and resends it as if it's written by bot, without a forwarding header.I have a solution in rather ugly manner, but it's not good with photo albums (many pics with caption viewed as single message in the client) cause it force 'em to be send as many single photo-messages not a whole one.

 
    def determine_msg_type_then_send( ... , msg, ... ): 
        if msg.photo:   
            ...
            send_a_photo_function( <someargs> ) 
        elif msg.document:  
            ...
            send_a_document_function( <someargs> ) 
        elif msg.sticker:  
            ...
            send_a_sticker_function( <someargs> )  
        ...
        else:  
            ... 
            send_a_message_function( <someargs> ) 
 

I still haven't added tests and completed the README. I'll fix those when the code is complete. If you think of a better name than chat_to_link and get_chat_link, kindly suggest it.

Reworking of ptb/#1789 for ptbcontrib.

Currently will fail, as the persistence setup needs ptb/#2212 to be fixed (hence marking as WIP)

On the fly edits some stuff in conftest.py allowing to actually run the tests. Overlooked that on the initial repo setup b/c extract_urls doesn't use dispatcher/bot

Right now, chat.link only returns the username in a link form. We could also return the chat_link there if it exists, or potentially (with a flag?) go even further and call export link and then return the link.

Might also be a nice fit for the contrib repo if this feature is a bit over the top.

Module this question is about

ptb_sqlalchemy_jobstore

Issue I am facing

How do I adapt this for python-telegram-bot 20.0a4?

And if so, how should I change the code:

updater = Updater("TOKEN")
dispatcher = updater.dispatcher
dispatcher.job_queue.scheduler.add_jobstore(
   PTBSQLAlchemyJobStore(
       dispatcher=dispatcher, url=DB_URI,
    ),
)

Traceback to the issue

Traceback (most recent call last):
  File "E:\YandexDisk\github\namaz_bot\main.py", line 26, in <module>
    from ptbcontrib.ptb_sqlalchemy_jobstore import PTBSQLAlchemyJobStore
  File "E:\YandexDisk\github\namaz_bot\venv\lib\site-packages\ptbcontrib\ptb_sqlalchemy_jobstore\__init__.py", line 18, in <module>
    from .jobstore import PTBSQLAlchemyJobStore
  File "E:\YandexDisk\github\namaz_bot\venv\lib\site-packages\ptbcontrib\ptb_sqlalchemy_jobstore\jobstore.py", line 27, in <module>
    from telegram.ext import CallbackContext, Dispatcher
ImportError: cannot import name 'Dispatcher' from 'telegram.ext' (E:\YandexDisk\github\namaz_bot\venv\lib\site-packages\telegram\ext\__init__.py)

Related part of your code

Python 3.9, python-telegram-bot==20.0a4

I started out with just updating to v20, but ended up also

• refactoring the internals to better leverage the parent class DictPersistence and
• adding support for callback_data

Unfortunately I'm not familiar for postgres (& databases in general), so I didn't try to actually test this with a real database (instead of just mocking as in the tests). If would be great if someone could cover that part :) CCing @starry69 for completeness 🙆‍♂️

It's not necessary that this is merged into v20-updates before #47 is merged. Just chose that as target branch so that the differences to main are more clearly visible.

so far:

• ReplyToMessageFilter
• get_chat_link
• send_by_kwargs (also closes #39)
• roles - here I removed the Roles.chat_admins/owner thingies, which made api requests in the check_update (rather: filter) step. Don't know why I even thought that were a good idea …

Based on #46. Selected that branch as target so that it's easier to see the changes. Should be merged into baster once #46 is merged.

• Added url parameter, now it's not necessary to manually build scoped session. though it can still be passed instead or url if needed.
• Added on_flush parameter when set True persistence will only save data on shutdown else it'll update the database on every update.
• Improved data saving query now it'll update the existing data in table instead of deleting and adding new data.

Hi @Poolitzer,

@harshil21 recently suggested me UsernameToChatAPI solution.

I tried to use example with the telegram public api like presented below however get error:

wrapper = UsernameToChatAPI("https://api.telegram.org/", "AnimusBoom", bot)

try:
    chat = wrapper.resolve("@julien59247787")
    print("Found")
except error.RetryAfter as e:
    time.sleep(e.retry_after)
    # both variants work
    chat = wrapper.resolve("julien59247787")
    print("found 2")

chat = wrapper.resolve("@julien59247787")
raise error.TelegramError(result["description"])
telegram.error.TelegramError: Not Found

Could you give me please some hints on what could be the issues?

Thank you for your time and answer

Animus

Currently we install extra-requirements of contributions during runtime in the tests, which has culprits (see https://github.com/python-telegram-bot/ptbcontrib/pull/27#issuecomment-904366995).

I suggest to instead run pytest several times and tell it to just consider one specific file each time. We probably could try to make each of that a standalone workflow, but this gives (number of python versions) × (number of contributions) suits and this seems a waste of resources to me. We can maybe write a simple script (bash? python?) that iterates over the contributions & calls pytest. Then we can run that script in the GH workflow and even locally as well.

This feature should introduce a give_me_a_link_to_the_chat_no_matter_what(chat: Chat) -> str function. This means, it should check if a username exists and then return it like chat.link does. If the chat doesnt have a username and is a group/channel, check if the bot is an admin with add member rights (I think this is required), then check if get_chat returns a link. If not, export a new one. Then return the invite link. And definitely document that calling this might result in three API calls

Now that v20.0 is out, we can update the requirements files to be a bit more permissive by specifying e.g. ~=20.0. Contrubtions that were already updated to one of the pre-release versions should be rather easy to update to v20.0 as few breaking changes have happend since then. A PR in this direction would be welcome and would probably only have to update the requirements.txt files as well as the README.md files.

PTB_sqlalchemy_persistence

Because other data in the project is already stored in mysql i've tried to adapt the code for using mysql instead of postgres. (And never run postgres myself, maybe this is already the source of the bug.)

Steps to reproduce

1. Altering the test for postgres-uri in line 71.
2. Use sqlalchemy-scoped-session bound to a mysql-database.
3. start the updater

Actual behaviour

File "persistence.py", line 108, in __load_database: self._chat_data = defaultdict(dict, self._key_mapper(data.get("chat_data", {}), int)) AttributeError: 'str' object has no attribute 'get'

For some reasons 'data' is of type 'str'. ("{}" after calling __init_database())

Solution

adding two lines of code at 105:

if isinstance(data, str):
    data = json.loads(data)

Version of Python, python-telegram-bot & dependencies:

$ python -m telegram python-telegram-bot 13.11 Bot API 5.7 certifi 2021.10.08 Python 3.9.7 (default, Sep 16 2021, 16:59:28) [MSC v.1916 64 bit (AMD64)]

mysql 8.0 sqlalchemy 1.4.27 conda 4.11.0 Windows 11

I thought somebody could find this interessting. Apart from this minor fix could this module named more general.

Module in question: ptb_sqlalchemy_jobstore

Steps to reproduce

1. Create a ConversationHandler with conversation_timeout set to any supported value other than None.

2. Add PTBSQLAlchemyJobStore to dispatcher.job_queue.scheduler .

3. Start the bot.

4. Enter ConversationHandler created in step 1.

Expected behaviour

Conversation Timeout job should be added to job store.

Actual behaviour

Job cannot be stored due to serialisation error.

Configuration

MacOs 12.0.1, python 3.9.7

Version of Python, python-telegram-bot & dependencies:

python-telegram-bot 13.8.1
Bot API 5.4
certifi 2021.10.08
Python 3.9.7 (default, Oct 22 2021, 13:39:39)  [Clang 13.0.0 (clang-1300.0.29.3)]
ptbcontrib 0.0.0

Logs

Traceback (most recent call last):
  File "venv/lib/python3.9/site-packages/telegram/ext/conversationhandler.py", line 497, in _schedule_job
    self.timeout_jobs[conversation_key] = j_queue.run_once(  # type: ignore[union-attr]
  File "venv/lib/python3.9/site-packages/telegram/ext/jobqueue.py", line 196, in run_once
    j = self.scheduler.add_job(
  File "venv/lib/python3.9/site-packages/apscheduler/schedulers/base.py", line 444, in add_job
    self._real_add_job(job, jobstore, replace_existing)
  File "venv/lib/python3.9/site-packages/apscheduler/schedulers/base.py", line 868, in _real_add_job
    store.add_job(job)
  File "venv/lib/python3.9/site-packages/ptbcontrib-0.0.0-py3.9.egg/ptbcontrib/ptb_sqlalchemy_jobstore/jobstore.py", line 65, in add_job
    super().add_job(job)
  File "venv/lib/python3.9/site-packages/apscheduler/jobstores/sqlalchemy.py", line 95, in add_job
    'job_state': pickle.dumps(job.__getstate__(), self.pickle_protocol)
TypeError: cannot pickle '_thread.lock' object
2021-11-11 23:15:26,089 - cannot pickle '_thread.lock' object
Traceback (most recent call last):
  File "venv/lib/python3.9/site-packages/telegram/ext/conversationhandler.py", line 497, in _schedule_job
    self.timeout_jobs[conversation_key] = j_queue.run_once(  # type: ignore[union-attr]
  File "venv/lib/python3.9/site-packages/telegram/ext/jobqueue.py", line 196, in run_once
    j = self.scheduler.add_job(
  File "venv/lib/python3.9/site-packages/apscheduler/schedulers/base.py", line 444, in add_job
    self._real_add_job(job, jobstore, replace_existing)
  File "venv/lib/python3.9/site-packages/apscheduler/schedulers/base.py", line 868, in _real_add_job
    store.add_job(job)
  File "venv/lib/python3.9/site-packages/ptbcontrib-0.0.0-py3.9.egg/ptbcontrib/ptb_sqlalchemy_jobstore/jobstore.py", line 65, in add_job
    super().add_job(job)
  File "venv/lib/python3.9/site-packages/apscheduler/jobstores/sqlalchemy.py", line 95, in add_job
    'job_state': pickle.dumps(job.__getstate__(), self.pickle_protocol)
TypeError: cannot pickle '_thread.lock' object

This idea came up in a discussion group, https://t.me/c/1101839433/8235. One can increase the major version every time a new module is added, and a minor increase if a module gets updated.

The benefit would be to enable installing ptbcontrib with pip install git+https://github.com/python-telegram-bot/[email protected].

This should be possible, according to this document: https://github.com/release-it/release-it/blob/master/docs/ci.md#github-actions

Great issue for someone foreign to the project trying to make a small change :)