Add SMART.KASAHUB to the map of supported devices.
This also adds fixture files for KH100, KE100, and T310, and adapts affected modules and their tests accordingly.
---------
Co-authored-by: Steven B <51370195+sdb9696@users.noreply.github.com>
* Added https://github.com/mihai-dinculescu/tapo/ to the list of related
projects.
* Changed the `kasa.tapo` to `kasa.smart`
---------
Co-authored-by: Steven B <51370195+sdb9696@users.noreply.github.com>
When smart devices encounter an error during a multipleRequest they
return the previous successes and the current error and stop processing
subsequent requests. This checks the responses returned and
re-queries individually for any missing responses so that individual
errors do not break other components.
Also updates CI pypy versions to be 3.9 and 3.10 which are the currently
[supported
versions](https://www.pypy.org/posts/2024/01/pypy-v7315-release.html).
Otherwise latest cryptography doesn't ship with pypy3.8 wheels and is
unable to build on windows.
Also updates the `codecov-action` to v4 which fixed some intermittent
uploading errors.
This changes `state_information` to return the names and values of
all defined features.
It was originally a "temporary" hack to show some extra, device-specific
information in the cli tool, but now that we have device-defined
features we can leverage them.
Allow controlling the color temperature via features interface:
```
$ kasa --host 192.168.xx.xx feature color_temperature
Color temperature (color_temperature): 0
$ kasa --host 192.168.xx.xx feature color_temperature 2000
Setting color_temperature to 2000
Raised error: Temperature should be between 2500 and 6500, was 2000
Run with --debug enabled to see stacktrace
$ kasa --host 192.168.xx.xx feature color_temperature 3000
Setting color_temperature to 3000
$ kasa --host 192.168.xx.xx feature color_temperature
Color temperature (color_temperature): 3000
```
This improves the initial update cycle to fetch the information as early
as possible and avoid requesting unnecessary information (like the child
component listing) in every subsequent call of `update()`.
The initial update performs the following steps:
1. `component_nego` (for components) and `get_device_info` (for common
device info) are requested as first, and their results are stored in the
internal state to allow individual modules (like colortemp) to access
the data during the initialization later on.
2. If `child_device` component is available, the child device list and
their components is requested separately to initialize the children.
3. The modules are initialized based on component lists, making the
queries available for the regular `update()`.
4. Finally, a query requesting all module-defined queries is executed,
including also those that we already did above, like the device info.
All subsequent updates will only involve queries that are defined by the
supported modules. This also means that we do not currently support
adding & removing child devices on the fly.
The internal state contains now only the responses for the most recent
update (i.e., no component information is directly available anymore,
but needs to be accessed separately if needed). If component information
is wanted from homeassistant users via diagnostics reports, the
diagnostic platform needs to be adapted to acquire this separately.
Caching pre-commit halves the linting time and the `action/setup-python`
cache does not handle `--extras` [properly
](https://github.com/actions/setup-python/issues/505) so switching to
action/cache for the poetry cache
Currently the CI is very slow for pypy vs cpython, one job is 24m vs 3m
on cpython. This PR enables poetry environment caching and bypasses
coverage checking for pypy.
N.B. The poetry cache is keyed on a hash of the `poetry.lock` file.
This should prevent child fixtures from hubs breaking tests due to
missing discovery info. To get these devices in `filter_fixtures`
include protocol string of `SMART.CHILD`.
Will create separate fixture files if the model of the child devices
differs from the parent (i.e. hubs). Otherwise the child device info
will be under `child_devices`
This is in preparation for tests based on supporting features amongst
other tweaks:
- Consolidates the filtering logic that was split across `filter_model`
and `filter_fixture`
- Allows filtering `dev` fixture by `component`
- Consolidates fixtures missing method warnings into one warning
- Does not raise exceptions from `FakeSmartTransport` for missing
methods (required for KS240)
Previously, only v1 of on_off_gradually is supported, and the newer versions are not backwards compatible.
This PR adds support for the newer versions of the component, and implements `number` type for `Feature` to expose the transition time selection.
This also adds a new `supported_version` property to the main module API.
Adds initial support for H100 and its alarmmodule.
Also implements the following modules for T315:
* reportmodule (reporting interval)
* battery
* humidity
* temperature
* Initialize children's modules (and features) using the child component negotiation results
* Set device_type based on the device response
* Print out child features in cli 'state'
* Add --child option to cli 'command' to allow targeting child devices
* Guard "generic" features like rssi, ssid, etc. only to devices which have this information
Note, we do not currently perform queries on child modules so some data may not be available. At the moment, a stop-gap solution to use parent's data is used but this is not always correct; even if the device shares the same clock and cloud connectivity, it may have its own firmware updates.
If a handshake session becomes invalid the device returns 403 on send and an `AuthenticationError` is raised which prevents a retry, however a retry would be successful. In HA this causes devices to go into reauth flow which is not necessary.
