* increasing python-websockets' version number so now it works with python 3.7
* required version for websockets increased to work with Python 3.7
* script/gen_requirements_all.py is done
* Don't treat typing as an "in-between" module for import order
That was a < 3.5 era thing.
* Tighten scope of some pylint unused-import disables
To avoid isort moving a top level one around, undesirably broadening its
scope.
* Allow device registry to optionally store config entries
* Connections and identifiers are now sets with tupels
* Make config entries mandatory
* Fix duplicate keys in test
* Rename device to device_info
* Entity platform should only create device entries if config_entry_id exists
* Fix Soundtouch tests
* Revert soundtouch to use self.device
* Fix baloobs comments
* Correct type in test
* First draft
* Generate device id
* No obscure registry
* Dont store config_entry_id in device
* Storage
* Small mistake on rebase
* Do storage more like entity registry
* Improve device identification
* Add tests
* Remove deconz device support from PR
* Fix hound comments, voff!
* Fix comments and clean up
* Fix proper indentation
* Fix pydoc issues
* Fix mochad component to not use self.device
* Fix mochad light platform to not use self.device
* Fix TankUtilitySensor to not use self.device
* Fix Soundtouch to not use self.device
* Fix Plex to not use self.device
* Fix Emby to not use self.device
* Fix Heatmiser to not use self.device
* Fix Wemo lights to not use self.device
* Fix Lifx to not use self.device
* Fix Radiotherm to not use self.device
* Fix Juicenet to not use self.device
* Fix Qwikswitch to not use self.device
* Fix Xiaomi miio to not use self.device
* Fix Nest to not use self.device
* Fix Tellduslive to not use self.device
* Fix Knx to not use self.device
* Clean up a small mistake in soundtouch
* Fix comment from Ballob
* Fix bad indentation
* Fix indentatin
* Lint
* Remove unused variable
* Lint
* Handle missing mpd capabilities
It is possible to configure mpd without volume or playlist support.
Gracefully degrade when either of these features appears to be missing.
Resolves: #14459, #15927
* Use longer name for exception
* Only return support flags post-connection
* Small consistency fixes to mpd.py for review.
* add pjlink media player component
* retrieve pjlink device name from projector if name isn't specified in configuration
* update .coveragerc
* fix style
* add missing docstrings
* address PR comments from @MartinHjelmare
* fix code style
* use snake case string for source names
* add missing period at the end of comment string
* rewrite method as function
* revert to use source name provided by projector
* Add media_player.dlna_dmr component
* PEP 492
* Move DIDL-template up
* Remove max_volume-override option
* Remove picky_device support
* Use DEFAULT_NAME
* Make supported_features static
* Remove unneeded argument
* Proper module-docstring
* Add http dependency
* Remove additional_configuration options, no longer used
* Change default name to 'DLNA Digital Media Renderer'
* Use python-didl-lite for DIDL-Lite-xml construction/parsing
* Handle NOT_IMPLEMENTED for UPnP state variables RelativeTimePosition and CurrentMediaDuration
* Use UPnP-UDN for unique_id
* Proper handling of upnp events
* Keeping flake8 happy
* Update requirements_all.txt
* Make UDN optional
* Ensure NotifyView is started, before using it
* Only subscribe to services we're interested in
* Don't update state_variables if value has not been changed + minor refactoring
* Improve play_media, follow flow of DLNA more closely
* Hopefully fix ClientOSError problems
* Flake8 fixes
* Keep pylint happy
* Catch errors and report gracefully
* Update async_upnp_client to 0.11.0
* Don't be so noisy
* Define/use constants for HTTP status codes
* Add discovery entry for dlna_dmr
* More robustness with regard to state variable not being set (yet)
* Keep privates hidden
* Handle NOT_IMPLEMENTED for CurrentTrackMetaData state variable
* Fixes in async_upnp_client + renew UPnP subscriptions regularly
* Not too eager
* Refactor duplicate code to _current_transport_actions and improve parsing of actions
* Support RC:1 to RC:3 and AVT:1 to AVT:3
* Moved DLNA-specifics to async_upnp_client.dlna.DmrDevice
* Use our own HTTP server to listen for events.
* More clear and explicit log message for easier troubleshooting
* Follow changes by hass, fixes traceback
* Fix not being able to do next
* Changes after review by @MartinHjelmare
* Linting
* Use homeassistant.util.get_local_ip
* Moved upnp event handling to async_upnp_client
* Keeping pylint happy
* Changes after review by @MartinHjelmare
* kodi add unique id based on discovery
* initialize unique_id to None
* use netdisco-extracted mac_address
* use an uuid instead of mac for real uniqueness
* add missing docstring
* verify that there is no entity already for the given unique id
* whitespace fix
* Add a component for Sisyphus Kinetic Art Tables
The [Sisyphus Kinetic Art Table](https://sisyphus-industries.com/) uses a
steel ball to draw intricate patterns in sand, thrown into sharp relief by a
ring of LED lights around the outside.
This component enables basic control of these tables through Home Assistant.
* Fix lints
* Docstrings, other lints
* More lints
* Yet more.
* Feedback
* Lint
* Missed one piece of feedback
* - Use async_added_to_hass in media player
- async_schedule_update_ha_state in listeners
- constants for supported features
- subscripting for required keys
- asyncio.wait
- update to sisyphus-control with passed-in session
* Update requirements
* lint
* fix aiohttp InvalidURL exception when fetching media player image
The first call for the HA proxy (`/api/media_player_proxy/media_player.kodi?token=...&cache=...`)
is receiving relative urls that are failing, this is a simple fix to precede the base_url when hostname is None.
* fix import location and sort stdlib imports
## Description:
More fixes flagged by pylint 2 that don't hurt to have before the actual pylint 2 upgrade (which I'll submit soon).
## Checklist:
- [ ] The code change is tested and works locally.
- [x] Local tests pass with `tox`. **Your PR cannot be merged unless tests pass**
* Add sound mode support
* continuation line indent
* indentation
* indentation
* Remove option to configure sound_mode_dict
* Sound mode support
- removed the sound_mode_raw propertie because it was not used, (still available through self._sound_mode_raw (as device attribute for automations and diagnostics)
* Detect sound mode support from device
Removed the config option to indicate if sound mode is supported.
Added detection if sound mode is supported from the receiver itself.
Pushed denonavr library to V.0.7.4
* Pushed denonavr to v.0.7.4
* Add python 3.7 to travis and tox
* Use pyyaml from github
* Don't version constraints
* Fix version tag
* Change to new pyyaml release
* Python 3.7 requires xenial
* Fix namespace detection
* Use correct RegEx type
* Update pexpect to 4.6
* Use correct validation for dictionaries
* Disable Py37 incompatible packages
* Upgrade all pexpect to 4.6
* Add explicit None as default param
## Description:
Switch to own packaged version of pylgnetcast
Request to make a pypi package didn't get any response: https://github.com/wokar/pylgnetcast/issues/1
**Related issue (if applicable):** #7069
* Show running apps as sources for Fire TV
* Fix unnecessary 'else' after 'return' (no-else-return)
* Remove 'pylint: disable=unused-argument'
* cleanup
* Improve volume support for Vizio Smartcast
* Vizio: avoid an error when 'self._device.get_current_volume()' returns 'None'
* Improve volume support for Vizio Smartcast
* Vizio: avoid an error when 'self._device.get_current_volume()' returns 'None'
* First line should end with a period
* Epson projector support. Version based on external library
* Epson projector support. Version based on external library
* modified epson according to MartinHjelmare review.
Added description of cmode to services.yaml
* renamed EPSON_SCHEMA to epson_schema
* removed method of getting cmode property
* removed unnecessary checks
change name of cmode service
* renamed SERVICE_ATTR_CMODE to SERVICE_SELECT_CMODE
Current uuid is ok when using only 1 snapserver
New uuid is needed when using multiple snapserver
Because the client can connect to more snapservers and
then uuid based on client MAC is not enough
* added new platform for the Unitymedia Horizon HD Recorder
* improve connection handling of the horizon platform
* remove unneeded parameters and fix spelling in the horizon platform
* abort or raise exception if connection to the device could not be established
* remove channel/source list and SELECT_SOURCE feature
* remove useless type check after cast and use a try block instead
* abort or raise exception if reconnect to device fails
* remove protocol specific code and restructure sending logic accordingly
* fix indentation to be pep8 complaint
* remove unused methods/properties
* fix unnecessary pylint commands and use a return to abert outside of setup_platform
* directly access config values
Unique id for client is generated from prefix 'snapcast_client_'
and MAC address
Unique id for group is generated from prefix 'snapcast_group_'
and UUID provided by snapcast library
When Samsung TV is idle for a period of time after issued a command,
subsequent 'turn_off' command won't turn off the TV. The issue is seen
in Samsung models with websocket as discussed in #12302.
== Reproducible Steps
1. Turn on TV (either via HA or directly).
2. Issue some commands e.g. volume ups / downs.
3. Wait for ~1 minute.
4. Issue turn_off command via HA. TV won't turn off.
5. Issue subsequent turn off commands won't turn off TV still.
6. However, issue some other commands e.g. volume ups / downs multiple
times in a row and then turn_off will turn off the TV.
== Root Cause
The underlying websocket connection opened by samsungctl get closed
after some idle time. There was no retry mechanism so issued commands
would intermittently fail but the subsequent one would succeed when
`_remote` get recreated. With `turn_off()`, however, there is an
additional call to `self.get_remote().close()` which indirectly caused
new connection to be created and then closed immediately. This causes the
component to stuck in failure mode when turn_off command is repeatly
issued.
== The Fix
Recreate the connection and retry the command if connection is closed
to avoid silent failures due to connection closed. Also set `_remote`
to None after calling close() to put it in correct state.
This fix eliminates intermittent command failure and failure mode in
turn_off().
* Media player: general sound mode support
* General sound mode support
* White spaces
* Add sound mode support to demo media player
* white space
* remove unnessesary code
* Version Requirement bump
Bump required version to 0.5 to allow component to work with Genie Mini clients using the clientAddr variable.
* Ran script/gen_requirements_all.py as requested.
The Universal media player inherits the states of the first child player that is not in some sort of "Off" state (including idle.) It was not considering the "unavailable" state to be off. Now it does.
This commit bumps the pycmus version used by the cmus component. There
was a bug in the previous version used, 1.0.0, when running in local
mode. This was caused by a mtreinish/pycmus#1 and also was reported in
the home-assistant forums (but not as an issue):
https://community.home-assistant.io/t/cant-install-cmus-component/7961
Version 0.1.1 of pycmus fixes this issue so it should work properly for
users running cmus and home-assistant on the same machine.
* onkyo: add support for max volume range
* onkyo: make flake8 happy
* onkyo: fix PEP8 D205 on line 181
* onkyo: use range for max_volume configuration
* onkyo: fix line too long