* Clarify InlineKeyboardButton callback-data docstring
Fixes#1267
* Improve documentation slightly
- Prettier changelog
- Link to examples on frontpage
- Link to wiki on frontpage
- Put telegram.ext in own sidebar thingy
* Improve ext.Filters documentation
Superseeds #1089
- Remove `:obj:Filter` as it's pretty obvious that it's a filter since it's in this module. It also made the html output about 1/3 times longer to scroll through.
- Add an __all__ to reorder so Filter is at the top instead of BaseFilter
- Add a proper docstring for document that documents the attributes (this is the #1089 part)
- Also fix a couple of grammar errors.
* Fix indentation in telegram.rst
* Add `git describe --long --tags` to the `python -m telegram` output
* Use admonition for examples to make them *pop* better
* Remove automodule telegram
Sphinx was whining a lot and it doesn't actually seem to be needed in any way?
* Fix Filters.document docstring per CR
Co-Authored-By: jsmnbom <jsmnbom@gmail.com>
* Fix spelling error in docs index
Co-Authored-By: jsmnbom <jsmnbom@gmail.com>
* Setting needs_sphinx to 1.7.9
so we don't have more issues with incongruent versions
* Fixed last warnings
* Make pre-commit happy
* Fix unresolvable promises
* added async test and description
* added test_none_on_first_message for conv_handler
* Small change in ConversationHandler docstring
* Fix test to work with new commandhandler
* Add WAITING state and behavior
* Remove `run_async_timeout` and `timed_out_behavior` arguments
* replace with `WAITING` constant and behavior from states
* never wait for promise to resolve (will hang up entire update queue
* see #1250 for discussion
* Fixing pytest version to 4.2.0
Pytest 4.2.1 has a weird bug on top level collect in 4.2.1 Fixing version to 4.2.0
* Make dispatcher use one context per update
It gives user the option to `overload` context with their own properties in a lower group handler if they like
* Improve callbackcontext & run_async docstring
- Add note about how you can add custom attributes to context.
- Add warnings about how run_async and custom attributes should not be used together.
* Small documentation improvements. [ci skip]
* update_filter attribute on filters
Makes it possible to have filters work on an update instead of message, while keeping behavior for current filters
* add update_type filter
* Messagehandler rework
- remove allow_edited (deprecated for a while)
- set deprecated defaults to None
- Raise deprecation warning when they're used
- add sensible defaults for filters.
- rework tests
* Commandhandler rework
* Remove deprecation test from new handler
* Some tweaks per CR
- rename update_types -> updates
- added some clarification to docstrings
* run webhook set test only on 3.6 on appveyor
* update_filter attribute on filters
Makes it possible to have filters work on an update instead of message, while keeping behavior for current filters
* add update_type filter
* Messagehandler rework
- remove allow_edited (deprecated for a while)
- set deprecated defaults to None
- Raise deprecation warning when they're used
- add sensible defaults for filters.
- rework tests
* Commandhandler rework
* Remove deprecation test from new handler
* Some tweaks per CR
- rename update_types -> updates
- added some clarification to docstrings
* run webhook set test only on 3.6 on appveyor
* Changes per CR
* Update travis to build v12
* small doc update
* try to make ci build version branches
* doc for BaseFilter
* Modify regexfilter and mergedfilter
Now returns a list of match objects for every regexfilter
* Change callbackcontext (+ docs)
* integrate in CommandHandler and PrefixHandler
* integrate in MessageHandler
* cbqhandler, iqhandler and srhandler
* make regexhandler a shell over MessageHandler
And raise deprecationWarning on creation
* clean up code and add some comments
* Rework based on internal group feedback
- use data_filter instead of regex_filter on BaseFilter
- have these filters return a dict that is then updated onto CallbackContext instead of using a list is before
- Add a .match property on CallbackContext that returns .matches[0] or None
* Fix and add test for callbackcontext.match
* Lots of documentation fixes and improvements [ci skip]
* Remove message decorator to fix default timeouts.
* Make message wrapper method private.
* Make tests pass
* Fix callbackquery shortcuts
Closes#1180
* Fix callbackquery shortcut tests
* Fix merge
* Address CR
* Add missing default timeout=20 for some bot file uploads
* Fix wrong return value in convhandler
Probably stems from a combination of bad merge plus quickly merged hacktoberfest PR.
* Improve regex docstring and add test case
* Add Ambro17 as contributor
* rename regex filter
* Fix sphinx documentation for Filters.regex
* Add spacing to render Note with blue background
* Add convenience classmethods for InlineKeyboardMarkup (#1186)
* Switch to row and column methods
* Also add convenience classmethods for ReplyKeyboardMarkup
* Add some simple tests
* Add conflicting bot id to conflict error message.
* Add test and comment to conflict error
* Don't extract bot id in Conflict exception per PR comments
* update MAX_CAPTION_LENGTH
Telegram silently changed the max length for captions to 1024 chars.
* Update test_constants.py
* change docstrings to reflect new length
* remove message
* clear message and proper match
* handler for ConversationHandler.END (timeout one) #1136
* review fixes
* review fixes
* review fixes
* review fixes
* docs and tests
* fixing stuff
* Fix problem
* fix conftest
* now it should work
* Add ConversationTimeoutContext
As discussed in the developers group. Use a class as the jobs context over using a dict.
* less verbosity
* Update data.py to be compatible with example
for now, if you process personal_info with example code, then you got an error if there is no set option to get native fist and last name.
setting default value will allow to process personal_info without native name/surname transation
* fixing line length
i hope i understood right this. sorry for delay, just starting using github!
* BasePersistence
* basic construct
* Keep working
* Continue work
Add tests for Basepersistence
* Finish up BasePersistence and implementation
* PickelPersistence and start tests
* Finishing up
* Oops, left in some typings
* Compatibilty issues regarding py2 solved
For Py2 compatibility
* increasing coverage
* Small changes due to CR
* All persistence tests in one file
* add DictPersistence
* Last changes per CR
* forgot change
* changes per CR
* call update_* only with relevant data
As discussed with @jsmnbom
* Add conversationbot Example
* should not have committed API-key
* Patch urllib3RequestField to make it *not* support RFC2231
* Add new required test data file
* Fix on py2
* Remove weird legacy code from inputfile
Not needed anymore, and also makes it not work... so that was not great
* Fix passport decryption failing at random times
Sometimes a decrypted secret was being treated as b64 and therefore got decoded even further. Fix by decoding b64 right before call to decrypt so we have better control of when not to do it
* Bot api 4.1
Telegram passport 1.1
Added support for middle names.
Added support for translations for documents
Add errors for translations for documents
Added support for requesting names in the language of the user's country of residence
Replaced the payload parameter with the new parameter nonce
NOTE: Scope stuff is NOT implemented, as we wanna STRONGLY encourage users to use the telegram provided SDKs anyway (and not generate telegram auth links in their bot, but rather on a server)
* Minor fixes
* Add hash to EncryptedPassportElement
For use with PassportElementErrorUnspecified apparently
Telegram Passport (#1174):
- Add full support for telegram passport.
- New types: PassportData, PassportFile, EncryptedPassportElement, EncryptedCredentials, PassportElementError, PassportElementErrorDataField, PassportElementErrorFrontSide, PassportElementErrorReverseSide, PassportElementErrorSelfie, PassportElementErrorFile and PassportElementErrorFiles.
- New bot method: set_passport_data_errors
- New filter: Filters.passport_data
- Field passport_data field on Message
- PassportData is automagically decrypted when you specify your private key when creating Updater or Bot.
- PassportFiles is also automagically decrypted as you download/retrieve them.
- See new passportbot.py example for details on how to use, or go to our telegram passport wiki page for more info
- NOTE: Passport decryption requires new dependency `cryptography`.
Inputfile rework (#1184):
- Change how Inputfile is handled internally
- This allows support for specifying the thumbnails of photos and videos using the thumb= argument in the different send_ methods.
- Also allows Bot.send_media_group to actually finally send more than one media.
- Add thumb to Audio, Video and Videonote
- Add Bot.edit_message_media together with InputMediaAnimation, InputMediaAudio, and inputMediaDocument.
Other Bot API 4.0 changes:
- Add forusquare_type to Venue, InlineQueryResultVenue, InputVenueMessageContent, and Bot.send_venue. (#1170)
- Add vCard support by adding vcard field to Contact, InlineQueryResultContact, InputContactMessageContent, and Bot.send_contact. (#1166)
- Support new message entities: CASHTAG and PHONE_NUMBER. (#1179)
- Cashtag seems to be things like $USD and $GBP, but it seems telegram doesn't currently send them to bots.
- Phone number also seems to have limited support for now
- Add Bot.send_animation, add width, height, and duration to Animation, and add Filters.animation. (#1172)
Co-authored-by: Jasmin Bom <jsmnbom@gmail.com>
Co-authored-by: code1mountain <32801117+code1mountain@users.noreply.github.com>
Co-authored-by: Eldinnie <pieter.schutz+github@gmail.com>
Co-authored-by: mathefreak1 <mathefreak@hi2.in>
* Commandhandler reworked
* Make CommandHandler strict
Only register valid botcommands, else raise ValueError
* Add PrefixHandler
* declare encoding on test_commandhandler
* Fix some tests dependend on CommandHandler
* CR changes
* small docfix.
* Test all possibilities for PrefixHandler
See https://github.com/python-telegram-bot/python-telegram-bot/wiki/Transition-guide-to-Version-11.0 under Context based callbacks and Filters in handlers for a good guide on the changes in this commit.
* Change handlers so context is supported
* Attempt to make parameter "guessing" work on py < 3.5
* Document use_context in all handlers
* Add Context to docs
* Minor fixes to context handling
* Add tests for context stuff
* Allow the signature check to work on py<3.5 with methods
* Fix order of operations
* Address most issues raised in CR
* Make CommandHandler no longer support filter lists
* Fix indent
(pycharm can be an arse sometimes)
* Improve readability in conversationhandler
* Make context have Match instead of groups & groupdict
* Remove filter list support from messagehandler too
* Small fix to StringCommandHandler
* More small fixes to handlers
* Amend CHANGES
* Fix tests and fix bugs raised by tests
* Don't allow users to ignore errors without messing with the warning filters themselves
* Ignore our own deprecation warnings when testing
* Skipping deprecationwarning test on py2
* Forgot some changes
* Handler: Improved documentation and text of deprecation warnings
* HandlerContext: Keep only dispatcher and use properties; improved doc
* Complete fixing the documentation.
- Fixes per Eldinnie's comments.
- Fixes per warnings when running sphinx.
* Some small doc fixes (interlinks and optionals)
* Change add_error_handler to use HandlerContext too
* More context based changes
Context Based Handlers -> Context Based Callbacks
No longer use_context args on every single Handler
Instead set dispatcher/updater .use_context=True to use
Works with
- Handler callbacks
- Error handler callbacks
- Job callbacks
Change examples to context based callbacks so new users are not confused
Rename and move the context object from Handlers.HandlerContext to CallbackContext, since it doesn't only apply to handlers anymore.
Fix tests by adding a new fixture `cpd` which is a dispatcher with use_context=True
* Forgot about conversationhandler
* Forgot jobqueue
* Add tests for callbackcontext & for context based callback job
* Fix as per review :)
* Use python 3 on read the docs
This has multiple improvements on the inspection of classes. Mainly that method signatures propely display the arguments for the method.
* Add note to telegram.Bot about camelCase
Switching to py3 means that the camelCase method descriptors will no longer be displayed in the docs. Adding a note to Bot to clarify they exist.
* Add Alias links for all camelCase
* Added extra Filters for File Type and Category
Added extra Filters for File Type and Category, based on the provided Mime-Type
* Added tests for the new Filters
Added Tests for the Category and File Types Filters.
* Fixed Tests
Fixed the Tests from the last commit
* Little Fix
* Revert of unwanted changes
* Update filters.py
* Changed the strings
* Changed file_type to mime_type
* Fixed Tests
* Fix in telegram.Message
The `_parse_(html|markdown)` methods now properly return `None` on an empty `Message.text` or an empty `Message.caption`
* As per CR
* Fix conversationhandler
As found by @nmlorg and described in #1031closes#1031
* adding another test and definitely finish conversationhandler
It seems another problem was when the job executed the timeout, it wasn;t removed from `self.conversation_timeouts` which made it still fail because job would be present in the handler dict, although it was already disabled.
This should fix it properly.
* add documentation for telegram.utils.promise and .request
* improve documentation for telegram.utils.promise and .request
* add missing 's' to new_chat_member(s) in all docstrings
* fix docs for `set_chat_photo`
[CI skip]
- File.download_as_bytearray - new method to get a d/led file as bytearray
This is much more convenient and straight forward than using a file
object.
- File.download(): Now returns a meaningful return value
- File.download*(): New and/or better unitests
We have some objects that have exactly the same to_dict() method, only specifying that `from_user` should be `from` in the `data`-dict. I refractored this logic to `TelegramObject` and removed those to_dicts() from the code.
TimedOut exception is an expected an normal event. To reduce noise and
make things more "fluent" we now:
- Make sure that we don't sleep after the timeout but rather retry
immediately.
- Log debug instead of error level.
Fixes#802
* Bot.get_file now allows passing a file in addition to file_id
* Add .get_file() to Audio, Document, PhotoSize, Sticker, Video, VideoNote and Voice
* Add .send_*() methods to User and Chat
* Added support for new field `telegram.Message.connected_message`
* Added support for new field `telegram.Message.connected_message`
* Added support for parse_mode in captions
* Added parse_mode parameter for captions in InlineQueryResult*
* Added supports_streaming parameter in telegram.Bot.send_video and telegram.InputMediaVideo
Fixed Docstrings for parse_mode in captions
* pypy3.5 unitests are now running with a new version due internal errors on travis.
closes#1005
BaseServer.handle_error() default behaviour is to print to stdout or
stderr (depends on the python version). In case that the file descriptor
is closed an additional exception will be raised, causing the webhook
thread to quit.
Fixes#970
- Fix JobQueue.jobs to obtain a lock on the internal queue object prior
to iterating over it.
- Rename JobQueue.queue to JobQueue._queue. This shouldn't be
accessible by the user directly, but rather only with sanitized
thread safe methods.
- JobQueue.interval_seconds - access self.interval only once to avoid
race conditions.
Fixes#968
- Use string `format` instead of dict comprehension.
- Better documentation to signify the semantics difference between
`name` and `full_name`.
- Use string `format` instead of dict comprehension.
- Better documentation to signify the semantics difference between
`name` and `full_name`.
* Removed the NOTE and mentinoed the "@" prefix.
* add support for 3.5 api
* removed "unused" import by accident
* Hardcoded values
Appearantly TG decided to change the size of a send image (again)
* test_official
* Improve coverage
* Finishing up
* spelling error
* pytest fixed tot < than 3.3 for python 3.3 support
* flake8
* rollback requirements
* as per CR
* object for provider_data
Make it possible to send an object that will be json-serialized for send_invoice + tests
* shorten error message
* using string_types
* Bot API 3.3
* simpler helper methods (accept only custom names now)
attempt to fix circular import (not sure if its the correct way tho)
added helper methods into User object
* fix User objects in tests to contain is_bot required value
* fix User objects in tests to contain is_bot required value
* delete extra line that caused flake8 error
* fix swapped arguments
The idea was nice, but it really complicated things for us and for the
user.
If a user wants to run more than one handler on an update, he can put
the handlers in different groups or he can have a single handler.
If a user wants to have multiple handlers in the same group which only
one of them should run on the update, he should use check_update().
Since we haven't released this code yet, there's no problem with
backward compatability.
Required fixes:
- CallbackQuery is now comparable.
- Message.effective_attachment, Message.photo,
Message.new_chat_members, Message.new_chat_photo &
Game.text_entitties semantic fixes - when they are not defined,
return an empty list.
- Docstring fix to Update class.
offset for get_user_profile_photo can be zero.
edit_message_caption should raise ValueError in compliance with other methods. Also changed to snake_case
edit_message_reply_markup was even weirder
`text_html` & `text_markdown` reverted to the old semantics - URLs are not converted to hyperlinks.
To get the new behaviour there are matching `text_html_urled` & `text_markdown_urled` properties.
fixes#773
- Added all the new and shiny features from API 3.1.
- Not API 3.1 changes:
- Use future.utils.string_types for string isinstance checks.
- Stall between retries of test_set_webhook_get_webhook_info() &
test_delete_webhook().
So just remove it for now. Better than doing annoying stuff with parsing a **kwargs dict in my opinion. It didn't even *really* need to be kwarg only anyways I guess...
* Clean up Bot code a bit
- Move decorators to module. It really wasn't clear how decorators inside classes work, and why they didn't have a self parameter, but still wasn't static. This also makes them effectively private without having to underscore them, which I think we should have done long time ago atm. Note that this might break backwards compatibility slightly (only if people are daft enough to have used the decorators themselves)
- Don't call _message_wrapper directly. Ever. Instead always use the message decorator, since it's what it's there for. Closes#627
- Don't use the message decorator if the method isn't supposed to return a message. The decorator could handle values like True (which is often the return value), but to someone reading the code, it seems like it's a message returning method even when it wasn't.
- Always document timeout and **kwargs
- Log all methods
* Add test to make sure timeout propagates properly despite decorators
* Rename shortcut functions to snake_case
* More function renaming
* Example function rewrite
* Add myself to authors.rst
* More function renaming
* Rename mockbot test functions
* Break comment line for flake max line length
* Change hard-coded 20. to timeout paramter
Passing timeout parameter to _message_wrapper in send_audio instead of hard-coded 20 seconds
* add myself to contributors
* add support for User.language_code
* Add language filter
Useful is you wanna do something like restrict your shop to a single or a few locales or something like that.
* Allow edited as seperate input
In short made it possible to tune messagehandler more to your wishes. and choose exactly what updates to receive. messages, edited_message or channel_post or a combination.
- Added the edited_updates argument to MessageHandler
- Added DepricationWarning when using allow_edited
- replaced _is_allowed_message and _is_allowed_channel_post with _is_allowed_update
- Modified tests to reflect new way
* oops
Spelled deprecation wrong
made an error in the _is_allowed_update.
* Python 2 does not have assertWarns.
* remove unneeded statements
* Improved CommandHandler
Now you can pass list of commands instead of one command
* Added tests for list of commands
* Return backward compatibility
Renamed `commands` to `command` in CommandHandler
* Added test for a command not in the list
* Fixed py2 unicode command support in `CommandHandler`
* Implemented `Bot.delete_message()` and `Message.delete()` methods
* Added tests for `Bot.delete_message()` and `Messsage.delete()`
* Added myself to `AUTHORS.rst`
* Using `assertRaisesRegexp` instead of `assertRaisesRegex` in tests
Now tests can run on Python 2.7 without raising 'AttributeError'
* Fix docstring
Added a warning to the docstring describing that this method is undocumented and not guaranteed to work
* Fix Bug #571
ConversationHandler will not process CallbackQuery if per_chat=True and
the CallbackQuery has no message attached to it (as is the case with
buttons on inline results)
* Adds test case for CallbackQuery without Chat
Attempting to import urllib3.contrib.socks may fail if PySocks is not
installed. Most users won't care for that.
Only import that module if the user requested to use a socks proxy.
* Introduce MessageQueue
* minor documentation and terminology fixes according to the review
* minor documentation and terminology fixes according to the review
* minor documentation and terminology fixes according to the review
* pep8 fix
* conversationhandler.py: add per_chat, per_user and per_message
* test_conversationhandler.py: test case per_user=False
* test_conversationhandler.py: add test for callbackqueryhandlers
* ✏️ Fix accidental typo in logging format
* stripping token of whitespaces before starting bot
* Line feed
* CommandHandler checks if message is for this bot
* CommandHandler checks if message is for this bot
- Added tests
* Fixed tests in test_conversationHandler to work with new commandhandler structure
* type in conversationhandler test
* Added methods to generate the original markdown/html formatted string from the entities contained in an update
* Added
* Moved the html/markdown parsing methods to `Message`
* Moved extract_* methods from helpers to the appropriate location
* Refactored text_markdown and text_html
* Refactored text_markdown and text_html for efficiency
* Fixed method call in conversationhandler
* Fixed method call in handler
* Fixed `make test` command on windows systems
* Improved method documentation
- Job.job_queue is now weakref.proxy reducing the risk of cyclic
pointers preventing Job object from being deleted.
- JobQueue._put(): raise if both next_t and job.interval are None
- Don't put repeating job back to queue if user had disabled it was
disabled during the time of execution.
- New method: Job.is_removed() - promising a consistent API (instead of
access to private member Job._remove)
- Documentation fixes.
* New fields channel_post and edited_channel_post for Update
refs #468
* setGameScore() changes
- Changed behaviour: messages with high scores will be update with new
high scores by default. (documentation fix)
- Use (new) disable_edit_message in setGameScore to disable the above new
behaviour.
- The edit_message parameter from setGameScore is no longer in use. For
backward compatibility, it will be taken into account for a while,
unless disable_edit_message is passed explicitly.
refs #468
* New field forward_from_message_id for Message.
refs #468
* New parameter cache_time for answerCallbackQuery
refs #468
* replykeyboardhide renamed to replykeyboardremove
refs #468
* Unitests for updated setGameScore semantics
refs #468
* Backward compatibility for ReplyKeyboardHide
refs #468
* Fix docstrings of wrapper methods in Message
* Unitest new field forward_from_message_id of Message
refs #468
* Fix testMaxCaptionLength
Telegram servers changed their behaviour - now they truncate a long
caption instead of returning an error.
* MessageHandler: Added support for channel posts
* Fix flake8 complaints in a manner which yapf will like it too.
* fix rst markup
* Fix telegram API change, returning '404 Not found' with raising own TelegramError rather native exception
* Change exception to InvalidToken in test and request util
* Added myself to AUTHORS. Thx for appreciation :)
* Adding timeunit and day support to the jobqueue
* Adding tests
* Changed the file permission back to 644.
* Changed AssertEqual argument order to (actual, expectd).
* Removed the TimeUnit enum and unit param, instead use datetime.time for interval.
* Removing the TimeUnits enum and unit param in favour of optionally using a datetime.time as the interval.
* Removing the TimeUnits enumeration, forgot the remove it in the last one.
* Removed some old docstrings refering to the TimeUnits enum.
* Removed the old TimeUnits import.
* Adding some error handling for the 'days' argument (only a 'tuple' with 'Days')
* Writing the error message directly in the exception.
* Moving a debug statement wrongfully saying a job would be running on days it wouldn't.
* Writing error messages directly in the exceptions instead of making an extra variable.
* Replacing datetime.time in favour of datetime.timedelta because of the get_seconds() method.
* Adding error handling for the method .
* Splitting the tests up in multiple ones, no float test because I haven't found a reliable way to test it.
* Excluding .exrc file.
* Removing \ at EOF of ValueError.
* Replacing Enums with plain new-style classes.
* Using numbers.number to check for ints/floats instead of seperate int/float checks.
* Fixing typo, number -> Number.
* Changed lower_case Days attributes to UPPER_CASE.
* Different formatting for Days class, removed the get_days function in favour of a tuple.
* Removed redundant function get_days.
* Edited the docstring for next_t to also take datetime.timedelta.
* Removed for-loop in favour of any().
* Changed docstring for interval.
* Removed debug print.
* Changing some docstrings.
* Changing some docstrings (again).
* update sphinx source files to properly build latexpdf and improve html build
* fix docstrings and sphinx sources to get rid of warnings
* add telegram.contrib.rst
* initial commit for user_data
* add chat_data and use defaultdict
* fix chat_data copy-paste error
* add test for user_data and chat_data
* fix case where chat is None
* remove braces from import line
While not stickily necessary for most classes (since user isn't directly creating them) it still unifies our approach.
However for some like ReplyKeyboardHide where users are making the classes themselves it should improve IDE autocomplete support.
* fixes broken test with PhotoSize, Game and Animation
However:
testSendGame and test_set_game_score both produces *BadRequest: u'Wrong file identifier/HTTP URL specified'*.
* fixes test_set_game_score
* adds to_dict method to Game to prevent extra collection type checks in base.TelegramObject
* Properly split and handle arguments in CommandHandler
* Update the docstring for pass_args in CommandHandler
* Properly split and handle arguments in StringCommandHandler
* bot.py: add create_references method
* create bot reference in webhook handler, use create_references on new updates
* message.py: implement reply_text
* echobot2.py: use Message.reply_text
* fix create_references in webhook handler
* add some more instance methods
* Chat.kick_member and unban_member
* bot.py: Create bot references in outgoing messages
* add tests for everything testable
* test_updater.py: add create_references method to MockBot
* remove Bot.create_references and refactor TelegramObject.de_json to take the additional parameter bot
* List bot as named kwarg where used
* file.py: Use Bot.request property instead of Bot._request attr
Should close#400.
* Add parse_entity
* Add parse_entities
* Add MessageEntity types as constants to MessageEntity.
* Add MAX_MESSAGE_ENTITIES to constants.py
Note: the value has been found by experimentation as opposed to extracted from the api docs.
* Add tests for parse_entity and parse_entities
* Create a Request class which maintains its own connection pool
* When creating a Bot instance a new Request instance will be created if one wasn't supplied.
* Updater is responsible for creating a Request instance if a Bot instance wasn't provided.
* Dispatcher: add method to run async functions without decorator
* Dispatcher can now run as a singleton (allowing run_async decorator to work) as it always did and as multiple instances (where run_async decorator will raise RuntimeError)
