API 4.5

[Bibo-Joshi]

PR for the API 4.5 changes. Closes #1689

• Two new messageEntities, strikethrough and underline
• nested messageEntities
• New parsemode, MardownV2
• New file_unique_id for Animation, Audio, Document, PassportFile, PhotoSize, Sticker, Video, VideoNote, Voice, File
• small_file_unique_id and big_file_unique_id to the object ChatPhoto.
• custom_title for ChatMember
• new method setChatAdministratorCustomTitle
• new slow_mode_delay for Chat object
• Before merge, merge master and revert temporarily skip tests failing b/c missing api 4.5 #1738
• When merged, update this wiki page.

---

Original description:

Cherry-picked from #1464 since we're not sure yet how Telegram will actually handle nested entities in API 4.5.

This only allows something like _bold *within* italic_ and not *bold _overlapping* italic_. The announcement is not every clear, if that is also allowed …

Leaving this here as a reminder, that we already have some code for that case.

[Bibo-Joshi]

The PR is ready for review now. The test failures seem unrelated (TestBot.test_pin_and_unpin_message and TestConversationHandler.test_per_message_false_warning_is_only_shown_once are nothing we fiddled with)

[dmytrohoi]

Reviewers, what do you think:

1. Should we add a shortcut for bot.set_chat_administrator_custom_title() to ChatMember class from chatmember.py?
2. I think because of this description, This is an obsolete mode saved for backward compatibility. To use this mode, pass Markdown to the parse_mode field from chapter "Markdown" in Telegram Bot API documentation, we need to use MarkdownV2 as * default * option in all related to markdown functions.
3. Instead of the markdown_smt_v2 statements, perhaps we need to use the markdown_v2_smt? (one of them is commented in this review)

[Bibo-Joshi]

1. Should we add a shortcut for bot.set_chat_administrator_custom_title() to ChatMember class from chatmember.py?
2. I think because of this description, This is an obsolete mode saved for backward compatibility. To use this mode, pass Markdown to the parse_mode field from chapter "Markdown" in Telegram Bot API documentation, we need to use MarkdownV2 as * default * option in all related to markdown functions.
3. Instead of the markdown_smt_v2 statements, perhaps we need to use the markdown_v2_smt? (one of them is commented in this review)

1. and 3. sound good to me. Could do it in the next days, but if you want to, please go ahead.

2. There is no default value for parse_mode anywhere, there is only Message.reply_markdown (correct me, if I'm missing something). For the latter, there is Message.reply_markdown_v2 now. For the former, Add default values #1490 is the PR, but it's left to the user to choose a default value, which makes sense because just putting MarkdownV2 as default would break backwards compitibility since it requires more characters to be escaped than v1

[Eldinnie]

Overall looking good. Just some questions and remarks added

[Eldinnie]

Codecov, mainly is unhappy because of the added checks for py2 (in parse_entities) which are never hit, because we don't test 2.7 anymore. You can ignore it

[Bibo-Joshi]

Thanks for the review @Eldinnie! Will try to incorporate the changes on the WE at the latest

[Poolitzer]

@Bibo-Joshi Since we want to get this done asap, I guess I can chime in this afternoon for me, if thats alright with you

[Bibo-Joshi]

Latest commit closes #1654 . Put this here since we're fiddling with _parse_markdown anyway and we would have merge conflicts else …

[gpchelkin]

Just a friendly reminder, I think it better belongs here. Please bump supported Bot API version number in https://github.com/python-telegram-bot/python-telegram-bot/blob/master/README.rst

It says 4.1 now, but one could tell from the changelogs, that actually it is 4.4 now. I think it might be repelling or misleading for new users of this great wrapper :-)

[Bibo-Joshi]

pytest failures seem unrelated (test_conversation_timeout_cancel_conflict)
codecov complains that in Message._parse_*, the parts for if sys.maxunicode == 0xffff are not hit, but I woudn't know how to test that …