* Unused pylint suppression cleanups
* Remove outdated pylint bug references
* Add flake8-noqa config and note to run it every now and then
* Add codes to noqa's
* Unused noqa cleanups
* Don't expose flows that aren't initialised.
If a flow init does not return immediately then there is a window where our
behaviour is screwy:
* Can try to configure a flow that isn't ready
* Can show notifications for discoveries that might yet return an abort
This moves the flow discovery events and notifications to after the flow is
initialised and hides flows that don't have a cur_step from async_progress
* Fix tradfri test
* Black.
* Lint fixes
* Move tradfri component to use new zeroconf discovery
* Will this work?
* Remove prints
* Correct order in generated zeroconf
* Update test_init.py
* Update test_init.py
* Update test_init.py
* Update test_init.py
* Fix light tests [skip ci]
* Fix tests/common
* Fix some mqtt tests [skip ci]
* Fix tests and component manifests which have only one platform
* Fix more tests and manifests
* Fix demo/notify tests
* Rollback test for demo.geo_location
* Clean up leftover config schema option
* Allow import groups via new config yaml setup
* Fix and add test
* Add a test without groups for legacy import
* Change default import groups to False
* Fix I/O in test
