Action types
This page assumes you already know how the mapping attribute works.
An action type is a reduced version of the Home Assistant script syntax. It defines one action or a sequence of actions to execute when an event is fired. The available action types are predefined actions, service calls, scene activation and delays.
...
mapping: # or merge_mapping
## Predefined actions
# `toggle` is a light predefined action
<event>:
action: toggle
# The previous action type can be reduced as:
<event>: toggle
## Call services
# Call any service the same way as it is done through `Developer Tools > Services` in HA
<event>:
service: script.my_script
data:
attr1: 42
attr2: foo
# `entity_id` can be passed directly like this or through `data`.
# Additionally, if the service is within the same domain
# (light, media_player, etc) as the main entity from the configuration,
# and entity_id is not passed, then it will use the one from the configuration.
# This is handy, so there is no need to repeat the same entity over and over again.
# Priority order for entity_id:
# - Inside data
# - In the same level as "service"
# - From the main config if the domain matches
<event>:
service: light.turn_on
entity_id: light.my_light
## Scene activation
# Activate any HA Scene
<event>:
scene: scene.my_scene
## Delay
# `delay` is useful when defining a list of actions, and you want
# an action to be triggered after some defined time.
# The value of the attribute only accepts seconds.
<event>:
- on_min_brightness # predefined action
- delay: 5 # wait 5 seconds
- on_full_brightness # predefined action
The <event> key is the event from your controller and integration.
If an action is still executing (most likely because of a delay) and the same event gets fired again, the behaviour depends on the mode parameter, which works like the Home Assistant automation modes. By default (mode: single), the new execution is ignored and a warning is logged. Use mode: restart to cancel the running action and start a new one.
Actions are executed sequentially, so it is not recommended to use a list of hold predefined actions, since it will not result in the expected behaviour. This is because hold actions are blocking operations, and they do not finish until a release action is fired.