Grooper Help - Version 25.0
25.0.0023 2,165
  • Overview
  • Help Status

Data Field Container - Generate Descriptions

Data Field Container Data Field Container - Container Command Grooper.GPT

Adds or updates descriptions for this element and its descendant Data Elements using generative AI.

Remarks

The Generate Descriptions command enables automated documentation of data models by generating concise, human-readable descriptions for each Data Element. This is accomplished by leveraging a configured Data Generator (LLM) and user-supplied instructions to produce context-appropriate property descriptions.

How It Works

Below is an overview of the command's operation:

  • The command generates a JSON schema for the selected Data Field Container, including all properties.
  • It sends this schema, along with user instructions, to the specified Data Generator.
  • The LLM processes the schema and returns an updated version with descriptions added or modified.
  • The generated descriptions saved into the 'Description' property for each Data Element.

Benefits

The 'Description' property is widely used in Grooper for purposes such as:

  • Enhancing exported JSON schemas with meaningful documentation for each JSON property.
  • Improving clarity for users working in the Data Grid by deliering descriptions as tooltips.
  • Providing instructions for individual Data Elements to guide LLM-based data extraction.

Well-crafted descriptions enable both humans and LLMs to understand the purpose and expected content of each Data Element, improving the overall usability and extraction accuracy for the Data Model.

Configuration Guidance

  • Configure a Data Generator that is tuned for technical writing or schema documentation.
  • Provide clear, specific instructions in the 'Instructions' property to guide the style and content of generated descriptions.
  • Use the 'Overwrite' option to control whether existing descriptions are replaced or only missing ones are filled in.

How It Works

  1. The command generates a JSON schema for the selected Data Field Container, including all properties and structure.
  2. System messages and user instructions are sent to the Data Generator, requesting updated descriptions.
  3. The LLM returns a revised schema with descriptions added or updated as specified.
  4. The command imports these descriptions back into the data model, updating the configuration as needed.
  5. All changes are saved to the repository.

Properties

NameTypeDescription

See Also

Notification