* Significantly speed up single entity/response service calls
Since the majority of service calls are single entity, we can
avoid creating tasks in this case. Since the multi-entity
service calls always check the result and raise, we can switch
the asyncio.wait to asyncio.gather
* Significantly speed up single entity/response service calls
Since the majority of service calls are single entity, we can
avoid creating tasks in this case. Since the multi-entity
service calls always check the result and raise, we can switch
the asyncio.wait to asyncio.gather
* revert
* cannot be inside pytest.raises
* one more
* Update homeassistant/helpers/service.py
* Ensure async_get_system_info does not fail if supervisor is unavailable
fixes#96470
* fix i/o in the event loop
* fix tests
* handle some more failure cases
* more I/O here
* coverage
* coverage
* Update homeassistant/helpers/system_info.py
Co-authored-by: Paulus Schoutsen <balloob@gmail.com>
* remove supervisor detection fallback
* Update tests/helpers/test_system_info.py
---------
Co-authored-by: Paulus Schoutsen <balloob@gmail.com>
* Differentiate between device info types
* Update allowed fields
* Update homeassistant/helpers/entity_platform.py
Co-authored-by: Martin Hjelmare <marhje52@gmail.com>
* Split up message in 2 lines
* Use dict for device info types
* Extract device info function and test error checking
* Simplify parsing device info
* move checks around
* Simplify more
* Move error checking around
* Fix order
* fallback config entry title to domain
* Remove fallback for name to config entry domain
* Ensure mocked configuration URLs are strings
* one more test case
* Apply suggestions from code review
Co-authored-by: Erik Montnemery <erik@montnemery.com>
---------
Co-authored-by: Martin Hjelmare <marhje52@gmail.com>
Co-authored-by: Erik Montnemery <erik@montnemery.com>
* Mark scripts as response optional, make it always return a response if return_response is set
* Update test_init.py
* Revert "Update test_init.py"
This reverts commit 8e113e54db.
* Split + add test
* Tune keep alives for polling integrations
aiohttp closes the connection after 15s by default,
and httpx closes the connection after 5s by default.
We have a lot of integrations that poll every 10-60s
which create and tear down connections over and over.
Set the keep alive time to 65s to maximize connection
reuse and avoid tls negotiation overhead
* Apply suggestions from code review
* adjust
* Update log message when referenced entity not found
* Update homeassistant/helpers/service.py
Co-authored-by: Martin Hjelmare <marhje52@gmail.com>
* Update test
---------
Co-authored-by: Martin Hjelmare <marhje52@gmail.com>
* Allow returning a script variable from a script
* Don't allow returning a template result
* Raise if response variable is undefined
* Add test
* Update homeassistant/helpers/script.py
Co-authored-by: Paulus Schoutsen <balloob@gmail.com>
* Format code
---------
Co-authored-by: Paulus Schoutsen <balloob@gmail.com>
* Add image platform to the template integration
* Set a default name
* Fix typo
* Add tests
* Improve test coverage
* Derive content-type from fetched image
* Add service response data for listing calendar events
Add the capability of response data for for the entity component.
* Rename input arguments and add service description
* Improve list events to be more user friendly
Allow the end date to be determined based on a relative time duration. Make the start time optional and set to "now". Add additional test coverage. Update demo calendar to actually perform date range checks.
* Wrap docstrings properly.
* Increase test coverage
* Update to use new API calls
* Readability improvements
* Wrap docstrings
* Require at least one of end or duration
* Check for multiple entity matches earlier in the request
* Update documentation strings
* Improve service response data APIs
Make the API naming more consistent, and require registration that a
service supports response data so that we can better integrate with
the UI and avoid user confusion with better error messages.
* Improve test coverage
* Add an enum for registering response values
* Assign enum values
* Convert SupportsResponse to StrEnum
* Update service call test docstrings
* Add tiny missing full stop in comment
---------
Co-authored-by: Franck Nijhof <frenck@frenck.nl>
* Allow scripts service actions to save return values
* Simplify script service response data
* Rename result_variable to response_variable based on feedback
* Speed up entity service calls
- Avoid permissions check if the caller is an admin
- Use set intersection instead of linear search of entity platforms to find entities
* tweak
* fix light test to not use an admin user
* Update return signature of service calls
* Add timeout error handling in websocket api for service calls
* Update recorder tests to remove assertion on service call
* Remove timeout behavior and update callers that depend on it today
* Fix tests
* Add missing else
* await coro directly
* Fix more tests
* Update the intent task to use wait instead of timeout
* Remove script service call limits and limit constants
* Update tests that depend on service call limits
* Use wait instead of wait_for and add test
* Update homeassistant/helpers/intent.py
Co-authored-by: Martin Hjelmare <marhje52@gmail.com>
---------
Co-authored-by: Paulus Schoutsen <balloob@gmail.com>
Co-authored-by: Martin Hjelmare <marhje52@gmail.com>
* Name unnamed binary sensors by their device class
* Update type annotations
* Fix loading of entity component translations
* Add test
* Update integrations
* Set abode and rfxtrx binary_sensor name to None
* Revert changes in homekit_controller
* Remove unneeded checks for Entity.platform
* Update tests
* Prevent breaking integrations without an EntityComponent
* Warn when entity has no platform
* support live reload of intent_script
* add services.yaml
* update tesls for full code coverage
* Update based on feedback
* fix intent_script reload when no intent_script config
* Update homeassistant/helpers/intent.py
* update tests to handle no_existing better
---------
Co-authored-by: Paulus Schoutsen <paulus@home-assistant.io>
* Migrate restore_state helper to use registry loading pattern
As more entities have started using restore_state over time, it
has become a startup bottleneck as each entity being added is
creating a task to load restore state data that is already loaded
since it is a singleton
We now use the same pattern as the registry helpers
* fix refactoring error -- guess I am tired
* fixes
* fix tests
* fix more
* fix more
* fix zha tests
* fix zha tests
* comments
* fix error
* add missing coverage
* s/DATA_RESTORE_STATE_TASK/DATA_RESTORE_STATE/g
* Fix refactoring error with updating in sequence
see #93649
* coverage
* make sure entities are being updated in parallel
* make sure entities are being updated in sequence
* Defer creating tasks we know are going to wait on the entity platform semaphore
When looking at whats going on with aiomonitor-ng, I noticed
we end up creating a lot of tasks that block waiting for the
executor because of the entity platform parallel_updates semaphore.
When we know the tasks are going to block we now await them
in sequence to avoid feeding the loop a herd of tasks that
will block on the semaphore
* change during iteration fix
* change during iteration fix
* cleanup
* cleanup
* fix vizio test
* Improve performance when serializing small bits of JSON
Making json_bytes a partial reduced the run time to
build the small JSON messages by ~18.75%
We serialize a lot of small messages over the websocket
* typing
* Switch some frequently called call_laters to call_at
call_at is a bit faster than call_later since call_later
is a wrapper around call_at.
We call at lot of these at startup so it helps a bit when
we are resource constrained
* update test
* update test
* Disable cleanup_closed for aiohttp.TCPConnector with cpython 3.11.2+
There is currently a relatively fast memory leak when using
cpython 3.11.2+ and cleanup_closed with aiohttp
For my production instance it was leaking ~450MiB per day
of `MemoryBIO`, `SSLProtocol`, `SSLObject`, `_SSLProtocolTransport`
`memoryview`, and `managedbuffer` objects
see https://github.com/aio-libs/aiohttp/issues/7252
see https://github.com/python/cpython/pull/98540
* Update homeassistant/helpers/aiohttp_client.py
* Support marking an assist pipeline as preferred
* Adjust
* Revert unneeded change
* Send preferred pipeline id in pipeline list
* Don't use property functions for the preferred pipeline
* Switch back to using call_later for the slow entity update warning
I had originally changed this to create a task and wait
in #41184 but that does not make sense anymore with newer
cpython as the profile now shows the original method is cheaper
(or I did it wrong the first time)
* fix missing block till done since there is no longer a task being created which would run the event loop once
* Fix memory churn in state templates
The LRU for state templates was limited to 512 states. As soon
as it was exaused, system performance would tank as each template
that iterated all states would have to create and GC any state
> 512
* does it scale?
* avoid copy on all
* comment
* preen
* cover
* cover
* comments
* comments
* comments
* preen
* preen
* Speed up entity filter when there are many glob matchers
Since we do no care about which glob matches we can
combine all the translated globs into a single regex
which reduces the overhead
* delete unused code
* preen
* Avoid more task creation in the discovery helper
There is no longer a reason to awaiti the jobs being dispatched
since nothing was using the result and there is no risk of
job being garbage collected prematurely anymore since
the task revamp
* Update homeassistant/helpers/discovery.py
* Allow passing an optional name to async_track_time_interval
This is the same idea as passing a name to asyncio.create_task which
makes it easier to track down bugs
* more
* short
* still cannot find it
* add a few more
* test