Using Lists or Slots in Trigger Sentences?

[queen anchor]

Alright, so my wife and I have a few apps that are really terrible about letting us know when we have a new message(s). Or we turn off notifications from that app because their notifications are all-or-nothing and we hate the dozens of pushy notifications they send each day. So I set up buttons I called "Ping" buttons found in a Ping subview in a dashboard (They all call a corresponding script). But, I thought it might be easier if we could ask Home Assistant to ping other person.

But, I really don't want to set up 5 different triggers for each app.

I'd rather have one trigger with the sentence similar to
tell my wife to check her {app_name}

and for whatever is said in {app_name} to determine which choice of actions is run. So if I say "tell my wife to check her snapchat", it'll run a script that sends her a notification to check her snapchat.

I've been trying to read up on intent, custom sentences, and custom responses and such, but so much of it gives complex examples I don't quite understand or talk about files that don't pre-exist in my files. And when I mention lists in searches, I get results about to-do lists which have nothing to do with what I'm trying to do.

(slightly related, I also re-did setting the lights to certain colors just so the assistant would respond with "as you wish").

Can I get a super beginner's guide on setting such wildcard words? Or whatever it takes to set this up correctly?

[gray tinsel]

Here's example:

language: "en"
intents:
  AnyListAddItem:
    data:
      - sentences:
        - "add {item} to <articles> {anylist} list"

  AnyListRemoveItem:
    data:
      - sentences:
        - "remove {item} from <articles> {anylist} list"

  AnyListGetItems:
    data:
      - sentences:
        - "(whats | what's | what is | what do i have) (in | on) <articles> {anylist} list"

lists:
  item:
    wildcard: true
  anylist:
    values:
      - in: "(costco | costa | cosco)"
        out:
          name: "Costco"
          entity_id: "todo.anylist_costco"
      - in: "errands"
        out: 
          name: "Errands"
          entity_id: "todo.anylist_errands"

expansion_rules:
  articles: "[the | my]"

[queen anchor]

1: I understand nothing of what you just put

2: my post has nothing to do with to-do lists.

[gray tinsel]

2. I don't want to write it for you. 🙂 It's example of how you can use slots and lists in custom intents. To-do lists here are just things that I use this intent for, so happened.

3. If you don't understand this, then what you tried so far? Post your current state.

[stiff basalt]

Maybe take a look at https://www.home-assistant.io/voice_control/custom_sentences/ to get started? It describes how to add a custom sentence trigger to an automation with a wild card slot like you are looking for. From there you can use the value in that slot and do what you need with it. 🙂

[queen anchor]

I set up a test

`alias: wildcard test
description: ""
triggers:

• trigger: conversation
  command:
  • I pick [the color] {color}
    conditions: []
    actions:
• set_conversation_response: Me, too. {{trigger.slots.color}} is my favorite.
  mode: single`

This works as you'd expect. But, when I try to add a "choose" with an option based on the color I said, all I get back is "Done". I thought a template condition would work fine.

`alias: wildcard test
description: ""
triggers:

• trigger: conversation
  command:
  • I pick [the color] {color}
    conditions: []
    actions:
• choose:
  • conditions:
    • condition: template
      value_template: "{{trigger.slots.color}} == "red"". ## (FOUND THE PROBLEM. IT'S HOW THIS IS FORMATTED)
      sequence:
    • set_conversation_response: Me, too. {{trigger.slots.color}} is my favorite.
      mode: single`

[zenith vault]

try {{ 'red' in trigger.slots.color | lower }}

Regarding your original request, please show us the actions of your “ping” script. The use of wildcards is excessive; it is better to use a custom sentence in YAML with a predefined list of several options, as shown above.

[queen anchor]

I figured out what I was trying to do.

For anyone else stumbling on this who may have also been confused on how to do something like this, here's a small example using lights because those are easy to understand.

This example will change the color of a single light. Like any automation, we start with the trigger. The trigger will be a sentence trigger as we're setting up a custom sentence we can speak into our HA voice assistant.

triggers:
- trigger: conversation
command: set the lightbulb to {color}

Yes, just write it like that. Normally, if I'd want it red, here's how a light.turn_on action would look in YAML

actions:
- action: light.turn_on
metadata: {}
data:
rgb_color:
- 255 # red
- 0 # green
- 0 # blue
target:
entity_id: light.the_lightbulb

If you add the choose building block, you can add conditions to each choice so that only the correct one gets applied. It looks like this in YAML:

actions:
- choose:
- conditions: [] # I left the condition blank for now
sequence:
- action: light.turn_on # here's the action we saw earlier
metadata: {}
data:
rgb_color:
- 255
- 0
- 0
target:
entity_id: light.the_lightbulb

Now, if I add a condition based on a template, I could write {{ trigger.slots.color == 'red' }} in the conditions. In YAML, that makes it look like this:

actions:
- choose:
- conditions:
- condition: template
value_template: "{{ trigger.slots.color == 'red' }}" # this is the magic
sequence:
- action: light.turn_on
metadata: {}
data:
rgb_color:
- 255
- 0
- 0
target:
entity_id: light.the_lightbulb

What's actually happening? The documentation has this to say:

Adding one or more {lists} to your trigger sentences will capture any text at that point in the sentence. A slots object will be available in the trigger data. This allows you to match sentences with variable parts...

The way this makes sense to me is I think of it as putting a "slot" in the sentence using the {curly brackets}. You can put anything in the curly brackets to name the slot and use trigger.slots.[your_chosen_slot_name] to refer to it later in the automation.