* Voice command API.AI. First import
* Fixes suggested by hound
* Fixing comments
* Fix pylint and pydocstyle errors
* Change how speech is defined
Also clean some unused constants, remove card type (not used), define
a message when action is not defined and improve the message when
action is unknown.
* Change how speech is defined
Clean some constants.
Improve error messages.
Delete card type, not used.
* Tests for new Api.ai component
* Use async_add_job to python compatibility. New test to measure response time
* Add async_action option to choose between waiting or not for the action to execute
* Travis-ci needs more time
* Removed timeout tests
* Removed timeout tests
* Added apiai to .coveragerc as specified by PR doc
Adrián López