Unified mocks

[nmlorg]

As part of the general-purpose .transcript-based test runner, I want to unify all or most of the various Bot and other mocks.

I'm toying with the idea of creating a [partial] mock of telegram-bot-api, which can intercept requests calls to https://api.telegram.org/... and respond to things like getme, sendmessage, etc., including keeping track of if bots have existing chats open with users (returning 403 if not), etc.

def test_something(telegram):
    bot = ntelebot.bot.Bot('1234:token')
    with pytest.raises(ntelebot.errors.Forbidden):
        bot.send_message(chat_id=1000, text='Should fail')
    telegram.send_update(1234, {'message': {'from': {'id': 1000}, 'chat': {'id': 1234}, 'text': 'Initial message'}})
    bot.send_message(chat_id=1000, text='Should succeed')

This could alternatively be part of [Mock]Bot:

def test_something(telegram):
    bot = ntelebot.bot.Bot('1234:token')
    with pytest.raises(ntelebot.errors.Forbidden):
        bot.send_message(chat_id=1000, text='Should fail')
    bot.handle_update({'message': {'from': {'id': 1000}, 'chat': {'id': 1234}, 'text': 'Initial message'}})
    bot.send_message(chat_id=1000, text='Should succeed')

(or maybe both, with the latter just being a convenient way to access the former).

 
The point would be to be able to clean up things like:

ntelebot/ntelebot/test_preprocess.py

Lines 40 to 45 in 9e42260

	if chat_id in self.unauthorized:
	raise ntelebot.errors.Forbidden()
	self.messages[chat_id, None] = text
	self.parse_modes[chat_id, None] = parse_mode
	return reply_to_message_id and 'REPLY' or 'SEND'

used in places like:

ntelebot/ntelebot/test_preprocess.py

Lines 160 to 162 in 9e42260

	assert ctx.reply_text(response_text) == 'REPLY'
	assert bot.messages[chat['id'], None] is response_text
	bot.messages.clear()