* Trend sensor now uses linear regression to calculate trend
* Added numpy to trend sensor test requirements
* Added trendline tests
* Trend sensor now has max_samples attribute
* Trend sensor uses utcnow from HA utils
* Trend sensor now completes setup in async_added_to_hass
* Fixed linter issues
* Fixed broken import
* Trend tests make use of max_samples
* Added @asyncio.coroutine decorator to trend update callback
* Update trend.py
* Initial support for SNMP switch
Supports setting integer values
* Resolves styling issues
* Updates requirements
* Remove commented out import
* Changes default community to private. Fixes linting errors.
* More linter fixes and bugfix in #state()
* Refactors the code to make it simpler.
* Add Google Domains component
* Fixes for hound
* Add Google Domains tests
* Fixes for hound
* Clean up Google Domains
* Add timeout to Google Domains
* Remove whitespace from blank lines
* Update google_domains.py
* Update google_domains.py
* Fix: Last Played Media Title in plex would stay even when player was idle/off
Primary Fix is in the "if self._device" portion.
code in "if self._session" is a catch all but i'm not 100% if it is needed.
* Fixed lint issues with previous commit
* 1st Pass at refactoring plex refresh
Moved _media** into clearMedia() which is called in _init_ and
at start of refresh.
Removed redunant _media_* = None entries
Grouped TV Show and Music under single if rather than testing
seperately for now.
* Fixed invalid name for _clearMedia()
Removed another media_* = None entry
* Removed print() statements used for debug
* Removed unneeded "if" statement
* Changes
* Updated Requests to 2.18.4
* Updated plexapi to 3.0.3
* Removed function to convert_NA_to_None
* Removed function get_thumb_url
Type changes
* _session.player is now a list players
* na_type deprecated and to be removed
* plexapi has change na to None
Known Issues:
* Player controls currently broken
* Last location (library) stays while player idle
* Username is now Usernames and a list
* Fix for broken controls
* Removed errant print statement
* Removed depecrated na_type
* Updated Plex Sensor to use plexapi 3.0.3
Added support for Token to be used in sensor
Known Issues:
Username and Password broken for Plex Sensor use Token instead for now
* removed TODOs
* Fixes for private access violations
* Removed need for _local_client_fix
* Removed unused import and fixed parens
* Async rewrite
* Device and events now working with async core
* More async than before
* Methods have moved
* Remove check if parameter serial number is available since library handles reconnection anyway and we can expect user to set up configuration correctly
* Async rewrite
* Device and events now working with async core
* More async than before
* Methods have moved
* Remove check if parameter serial number is available since library handles reconnection anyway and we can expect user to set up configuration correctly
* Serial number is the only unique identifier, checks are still needed
* No async for setup_device
* Bump axis to 13 in preparation of friday the 13th
* Fix review comments
* Removed async after some discussions with @armills.
It wasn't possible to get past that aiohttp doesnt support digest auth, which makes it impossible to convert the full library at this point in time.
* Xiaomi Air Purifier 2 integration
* Flake8 errors fixed.
Changes based on review.
* Service domain ("fan") updated and services properly prefixed by xiaomi_miio.
* The underlying library is called python-miio now. Imports and requirements updated.
* Version bumped. The underlying library is called python-miio now.
* Fixes#10030. Extented Network Exceptions handling.
* Remove unused import. Replace ex.reason to ex.message to use custom exception instead of HTTPError
* Init commit of new whois sensor
* Updated requirements
* Resolved updated showing expired, added expired attr
* Added missing attribute in init, added whois to coveragerc
* Various PR comment changes
- Now more resiliant to invalid hostnames
- Removed various assumed STATE_UNKOWN setting
- Upfront check for valid hostname preventing the sensor starting with dud
- Resolved unit of measurement Day, Days, None issue
- Datetime formatting now done to iso 8601 standard
- Removed all expired usage, not really that useful
- Unused hass assignment
* More PR comment resolutions
- Resolved the dilemma with hosts / single host per sensor. Now running
single domain per sensor.
- Renamed host(s) to domain
* Moved coveragerc sensor location
* Re-phrased the expiration_date warning
* Resolved assumed updated_date existence
* Resolved missing indent
* Resolved discover_info typo
* Update whois.py
* add eph ember controls
* updates based on review
* remove unused import
* update to new version of pyephember
* added myself to codeowners as requested
* make codeowners alphabetical
* run fixed gen_requirements_all
* Update ephember.py
* Do not include program data in media_title if program data is undefined (None)
* Show thumbnail of currently playing program
* async setup
* Update requirements
* Initial revision of fail2ban sensor
* Verified working, added tests
* Re-factored code so that log reading isn't called for each sensor
* Lint fixes
* Removed errant reset of last ban, added test to verify bans persist through update
* Removed for loop in read_log and replaced with regex per review request
* Refactored update to use current ban array for last ban state
- also was missing return False in timer for default behavior
* Removed CONF_SCAN_INTERVAL from PLATFORM_SCHEMA.extend
- renamed DEFAULT_SCAN_INTERVAL to SCAN_INTERVAL
* SCAN_INTERVAL changed to timedelta
* Force travis rebuild (last build timed out)
* Using compiled regex now
* fixing a typo in the old library.
Should now work with both version 1 and version 2 hub
* version bump
* fix Scene shadowing
* fix requirements. (not sure whether I should commit the other generated files as well ?)
* Test including extra data on a no data trigger
* Match any dicts for default schema for event data
* Fix indentation
* Only check schema if one was configured
The default names for the feeds created by the EmonCMS component are
like 'emoncms1_feedid_10', but
- This is the display name. The ID should be lowercase and underscored,
but the display name should be readable. The ID gets derived from it
and comes out formatted correctly.
- EmonCMS lets you assign names to feeds, so it makes sense to use those
if they exist, rather than feed IDs. The ID is pretty meaningless and
basically means you have to override every name to make it readable.
- Including the ID identifying the EmonCMS instance (i.e. the '1') makes
the name clunkier and would only be useful for people with multiple
EmonCMS instances, which is likely to be an extremely small group since
one hub can run as many feeds as you need it to.
This changes the default behavior but still uses configured 'name' if
it's set, so it won't break the configuration of people who have
customized their feed names in HA config.
Sam Birch