Label Info

The 'Text' property defines the primary string that will be matched against document content to identify this label.

• This value should match the actual label as it appears on the document, including case and punctuation if relevant.
• If the label may appear in multiple forms (e.g., "Invoice Number", "Inv #", "Bill No."), use the 'Alternate Versions' property to specify additional accepted forms.
• The text is used by label-aware extractors and classification logic to locate the corresponding Data Element.

For best results, keep label text concise and unambiguous.

Can be one of the following values:

The 'Kind' property specifies the layout or structure of the label as it appears in the document.

• Layout options include: Simple (whole words), Tabbed (across segments), Substring (within words), Boxed (inside a box), Wrapped (across lines), and Ruled (vertical wrapping).
• Grooper attempts to detect the layout automatically, but you may override it if the default is incorrect.
• The layout affects how the label is matched and extracted, especially in complex or non-standard document formats.

Adjust this property if label matching is unreliable due to unusual formatting.

The 'Parent' property allows you to define a hierarchical relationship between labels, requiring this label to occur inside the specified parent label.

• Use this to model nested structures, such as fields within table headers or grouped sections.
• Custom labels defined on the parent container can be selected as parent labels.
• Table headers that are peers of the field may also serve as parent labels, enabling extraction from Nested Tables.

To add custom parent labels, select the parent container, right-click the tab headers, and use the "Add Custom Label" command.

The 'Padding' property specifies additional space to include around the label's bounding box when matching it in the document.

• Use this to fine-tune label detection in cases where the label may be surrounded by whitespace, lines, or other elements.
• Padding can be set independently for each edge (top, bottom, left, right).
• Adjust padding if label matching is too strict or too loose due to document formatting.

This is especially useful for labels inside boxes or with variable spacing.

The 'Volatile' property indicates that this label is not required for extraction or classification.

• When set to true, the label will be ignored for classification purposes, preventing false negatives if it is missing.
• Use this for labels that are only present in some document variants or are not essential for identifying the Data Element.
• Volatile labels are still used for extraction if present, but their absence does not affect classification.

Mark labels as volatile to improve robustness in variable or optional layouts.

The 'Alternate Versions' property allows you to specify additional label strings that are considered equivalent to the primary label.

• Use this to support common synonyms, abbreviations, or alternate spellings that may appear in different document layouts.
• Each alternate can have its own text, layout, and padding settings.
• All alternates are considered when matching labels during extraction and classification.

This property is essential for handling real-world document variability and maximizing extraction reliability.