Fill text inserts text pulled from value points or the data token. It is useful for dynamically creating AI service instructions.

How it Works

Fill text is accessed via preconfigured code points available to all organizations. The following are the steps to use fill text.

Import Point

Import fill_text:

{
    "imports": [
        "fill_text"
    ]
}

Use in Code Point

Fill text evaluates criteria within sets of double brackets, such as [[<criteria>]]. The following criteria are supported:

Data Token

[[path_in_data_token]]: Replaces with the value from the path within the data token.

Example

[[_this.type]] would be item, if the path _this.type has a value of item.

Data Token or Other Value

[[path_in_data_token?Or this instead]]: Replaces with the value from the path within the data token if it exists; if not, it replaces with the alternative text.

Example

[[_this.type?hello]] would be item, if the path _this.type has a value of item. Alternatively, if _this.type is undefined, it would replace with hello.

Value Point

[[@name_of_value_point]]: Replaces with the value from a value point.

Example

[[@organization_name]] would be melodyarc, if a value point named organization_name had a value of melodyarc.

List from Value Points in Partition

[[*partition_name.key_in_value_key]]: Retrieves a defined key from within the value key from all value points within a partition and formats it into a markdown-compliant bulleted list. Can optionally sort by adding.

📘

Sorting

If value points have a key of value.sort, that value will be used to sort the bullets in an ascending order.

Example

[[*label-message.option]] would be - item 1 - item 2 - item 3, if there were three value points with a partition of label-message and each had a value of value.option equal to either item 1, item 2, or item 3.