Definition

A label is a kind of reusable template.

Here is an example :

? HELLO
    - {MY_LABEL}

? HELLO JOHN
    - {MY_LABEL}

[MY_LABEL]
    - Hello there

In this script, MY_LABEL is label, the syntax follows this rules :

Starts a line
Label identifier are between square brackets
Label identifier is composed of letters only and underscores.
Label is followed by templates

In that case, each templates call the label ‘MY_LABEL’, so the rendered template is Hello there

Parameters

A label can define parameters :

? HELLO
    - {MY_LABEL}

? HELLO JOHN
    - {MY_LABEL("John")}

[MY_LABEL(name)]
    - Hello %name

MY_LABEL now defines a ‘name’ parameter, and can be used in template.

Hello -> Hello
Hello john -> Hello John

A label can have multiple Parameters, the syntax follows this rules :

all parameters are in curly brackets, between label identifier and closing square brackets
parameters are separated by coma

To call a label with parameters, it follows the same rules, but you need to put values in double quotes.

Default

Athena defines a special label called : DEFAULT

This label is called each time the engine cannot find a matching scenario :

? HELLO
    - {MY_LABEL}

? HELLO JOHN
    - {MY_LABEL}

[MY_LABEL]
    - Hello there

[DEFAULT]
    - Sorry i did not get that

In this script, if the user says ‘Bye’, the agent will respond with ‘Sorry i did not get that’

Sub Scenarios

A label can have sub scenarios :

? HELLO
    - {MY_LABEL}

[MY_LABEL]
    - How are you today ?
        ? FINE
            - Glad to hear that !
        ? BAD
            - What can I do for you ?

This is equivalent to :

? HELLO
    - How are you today ?
        ? FINE
            - Glad to hear that !
        ? BAD
            - What can I do for you ?

Be wary that any sub scenario of the scenario will be overriden by the label’s sub scenario, even if none. To avoid this, cf. https://my.satisfaction.ai/docs/athena/en/labels.html#sub-scenarios-combined :

? HELLO
    - {MY_LABEL}
        ? GREAT
            - That's good :)
        ? TERRIBLE
            - Ohhhhhh

[MY_LABEL]
    - How are you today ?
        ? FINE
            - Glad to hear that !
        ? BAD
            - What can I do for you ?

This is equivalent to : (The sub scenario is removed.)

? HELLO
    - How are you today ?
        ? FINE
            - Glad to hear that !
        ? BAD
            - What can I do for you ?

Successive call label

A template can have successive calls to labels :

? HELLO
    - {YOUR_NAME} {MY_LABEL}
        ? GREAT
            - That's good :)
        ? TERRIBLE
            - Ohhhhhh

[YOUR_NAME]
    - Hello John

[MY_LABEL]
    - How are you today ?
        ? FINE
            - Glad to hear that !
        ? BAD
            - What can I do for you ?

All call labels are generated, from the left to the right. Here we have two generations :

First :

? HELLO
    - Hello John {MY_LABEL}

[MY_LABEL]
    - How are you today ?
        ? FINE
            - Glad to hear that !
        ? BAD
            - What can I do for you ?

You can notice in this example that the sub scenario is removed, even if the label YOUR_NAME doesn’t have a sub scenario.

Second :

? HELLO
    - Hello John How are you today ?
        ? FINE
            - Glad to hear that !
        ? BAD
            - What can I do for you ?

Call label as parameter

A label with parameters can be called with a call to a label :

? HELLO
    - {YOUR_NAME("{BOT_NAME}")} {MY_LABEL}
        ? GREAT
            - That's good :)
        ? TERRIBLE
            - Ohhhhhh

[BOT_NAME]
    - Alf

[YOUR_NAME(bot_name)]
    - Hello John, my name is %bot_name

[MY_LABEL]
    - How are you today ?
        ? FINE
            - Glad to hear that !
        ? BAD
            - What can I do for you ?

Never forget to surround the call label between double quotes !

Successive outputs are :

{YOUR_NAME(“Alf”)} {MY_LABEL}
Hello John, my name is Alf {MY_LABEL}
Hello John, my name is Alf How are you today ?

Sub Scenarios combined

Because a call to a label override the current sub scenario, a syntax exists to modify the behavior :

? HELLO
    - {+MY_LABEL}
        ? GREAT
            - That's good :)
        ? TERRIBLE
            - Ohhhhhh

[MY_LABEL]
    - How are you today ?
        ? FINE
            - Glad to hear that !
        ? BAD
            - What can I do for you ?

The + before the label identifier means that the sub scenario will be aggregated to the current sub scenario :

Which is equivalent to

? HELLO
    - How are you today ?
        ? GREAT
            - That's good :)
        ? TERRIBLE
            - Ohhhhhh
        ? FINE
            - Glad to hear that !
        ? BAD
            - What can I do for you ?