Add intent fallback system

[MatthewScholefield]

This is the same pull request from 3 weeks ago, but with skill-wolfram-alpha blacklisted and a new fallback-wolfram-alpha added to the default skills. Note: since Wolfram's API key is currently not working, we this cannot be tested.

Original description:
This is a new system that allows other skills to register as fallbacks to handle general knowledge queries. For now, each skill self-assigns a priority where 0 is the highest and 100 is very low priority. When an intent failure occurs, each fallback gets run until one returns True meaning it found a response to speak to the user. Example usage:

skill-my-fallback/__init__.py:

from mycroft.skills.core import MycroftSkill
class MyFallbackSkilll(MycroftSkill):
    def __init__(self):
        MycroftSkill.__init__(self, name="MyFallbackSkill")

    def initialize(self):
        self.register_fallback(self.handle_fallback, 80)

    def handle_fallback(self, message):
        if 'what is' in message.data['utterance']:
            self.speak_dialog('the answer is always 42')
            return True
        return False

This PR also removes the multi utterance intent fail. It only makes sense to emit an intent_failure regardless of the amount of intents, especially since we only ever receive one utterance anymore.

When testing, check functionality of Wolfram Alpha for questions like what is a fox (and maybe wolfram skill reloading when /opt/mycroft/skill-wolfram-alpha/init.py is edited).

[coveralls]

Coverage decreased (-0.05%) to 41.339% when pulling 4db6539 on feature/fallbacks into e1a1751 on dev.

[forslund]

Looks extremely good. Unfortunately wolfram alpha seems to be unavailable to me right now so I haven't been able to test it.

Also PR 145 on mycroft-skills (MycroftAI/mycroft-skills#145) need to be merged before this one.

Would like to verify it with a working wolfram alpha before merging.

[coveralls]

Coverage decreased (-0.04%) to 41.221% when pulling fee8ffa on feature/fallbacks into 4c1fbcb on dev.