* TTS Component / Google speech platform
* Change file backend handling / cache
* Use mimetype / rename Provider function / allow cache on service call
* Add a memcache for faster response
* Add demo platform
* First version of unittest
* Address comments
* improve error handling / address comments
* Add google unittest & check http response code
* Change url param handling
* add test for other language
* Change hash to sha256 for same hash on every os/hardware
* add unittest for receive demo data
* add test for error cases
* Test case load from file to mem over aiohttp server
* Use cache SpeechManager level, address other comments
* Add service for clear cache
* Update service.yaml
* add support for spliting google message
* Add MAC vendor lookup for device_tracker.
* Test vendor mac lookup and fix device attribute.
* Generate requirements.
* Style.
* Use hyphen instead of underscore to satisfy 'idna'.
https://github.com/kjd/idna/issues/17
* Resort imports.
* Refactor macvendor to use macvendors.com API instead of netaddr library.
* Test vendor lookup using macvendors.com api.
* Remove debugging.
* Correct description.
* No longer needed.
* Device tracker is now an async component. Fix ddwrt tests.
* Fix linting.
* Add test case for error conditions.
* There is no reason to retry failes vendor loopups as they won't be saved to the file anyways at that point.
* Sorry, bad assumption, this only made things worse.
* Wait for async parts during setup component to complete before asserting results.
* Fix linting.
* Is generated when running 'coverage html'.
* Undo isort.
* Make aioclient_mock exception more generic.
* Only lookup mac vendor string with adding new device to known_devices.yaml.
* Undo isort.
* Revert unneeded change.
* Adjust to use new websession pattern.
* Always make sure to cleanup response.
* Use correct function to release response.
* Fix tests.