Alembic Workbench User's Guide

9. Tagging File for Entities

Annotating versus Editing Data Files

Once a file is opened, it can be annotated; the user can add, delete, or manipulate the boundaries of markup. (Exception: markup that is write-protected cannot be edited.) The raw text itself, however, is protected against changes and cannot be modified. Editing, therefore, should be completed off-line.

Adding Tags

In order to apply a tag to text, the text must be selected. Selected text is highlighted in the color of the current tag. When a tag is applied to the selected text, the text itself is shown in the foreground color of the current tag, and remains highlighted in the background color. The user will notice in the tag line (at the bottom of the tool) the name of the tag that was applied. Note that the tool allows the addition of embedded and coextensive tags. For the case of a text string with nested tags, the tag is simply added to the lexeme selected by the user. For example,

<NUMEX TYPE=MONEY><ENAMEX TYPE=LOCATION>US</ENAMEX> $10 million</NUMEX>

For text tagged with coextensive tags, the text is highlighted in the most recently added tag. However, when the file is saved out in an SGML-encoded file, the tagset definition is consulted to determine the priority (position) of the new tag.

To change the current tag, click on the Tag menu and select a new tag. The current tag will reflect the chosen one, and any selected text will be highlighted in the color of the new current tag. Alternately, use the keyboard accelerator, e.g. press Control-l to change the current tag Name to Location, specified in the .awb-prefs file. Both of these methods can be applied to already selected text, changing the selection to the new current tag.

Deleting and Changing Boundaries of Tags

To delete a tag from text, the text should be selected by pressing Control-Shift mouse button 1, followed by either the Delete or BackSpace key. Once the tag is deleted, the colored highlighting disappears. The boundaries of already-existing tags can be extended or shortened by similarly selecting the highlighted text, and then pressing the appropriate combination of Meta-mouse button to change the boundaries of the selection.

When there are coextensive tags, the text will be displayed in the color of the outermost tag. For example, in the case of:

<ENAMEX TYPE=PERSON><LEX TYPE=NNP>Clinton</LEX></ENAMEX>

Clinton will be highlighted in the color designated for PERSON tags.

To delete any underlying tags, locate the cursor on the highlighted text and press tab. Delete accordingly.

**Table 1: Mouse Bindings**
Mouse Binding	Abbreviation	Description
mouse button 1	<B1>	To select a word or to extend selection left
mouse button 2	<B2>	To tag selection with current tag
mouse button 3	<B3>	To select a word or extend selection right
mouse button 1 and drag	<Drag-B1>	Drag out an exact selection
Shift-button 1	<S-B1>	Extend selection to here
Meta-button 1	<M-B1>	Shorten selection left
Control-Shift button 1	<C-S-B1>	Select a tag to change or delete
Tab		To "sift" through coextensive tags
Shift-button-2	<S-button-1>	To override Auto Tag's auto mode

Note: The Esc key can be used as an alternative to the Meta key.

NEXT: 10. Tagging Files for Coreference

Go to 6. Setting up and Editing the Tag Menu

Return to Alembic Workbench User's Guide Table of Contents