Module io.annot8.core

This module contains the core interfaces for use within Annot8. It does not contain any implementations of the interfaces described.

The basic premise of how Annot8 works is as follows:

You have an Item, which is your container and holds information about the data you wish to process. This Item contains one or more Content objects, which are essentially views on the data - for instance, if the data you wished to process was a HTML page, you might have a plain text 'view' of the data as well as a formatted view.

On to these Content objects, you create Annotations. Annotations indicate that a certain part of the Content has some special meaning (e.g. it is a mention of a person). These Annotations are stored in an AnnotationStore, which is held by each Content object. An Annotation has Bounds associated with it, which indicate the extent of the Annotation. The type of Bounds will vary depending on the Content.

You can link Annotations together with Groups. When adding an Annotation to a Group, you must assign it a role. This allows Groups to be used for a wide range of use cases, such as creating relationships or event, and indicating co-reference between Annotations. Groups are stored in a GroupStore, which is held on the Item object so that a Group may contain Annotations from multiple Content objects.

Most of the objects above support Properties, which allow you to attach additional information to an object. Many objects also have a type and an ID.

There are three types of 'component' which are used to create and manipulate the above objects. A Source will create new Items from some external source, for example a file system or a database. The Items created by a Source will be passed to one or more Processors, which will manipulate the Items by adding new Content objects to it, manipulating the Annotations within these Content objects, and creating Groups of Annotations.

Both the Sources and Processors can use the third type of component - a Resource - to access common or external resources (e.g. a database connection, or reference dataset).

It is envisaged that Sources, Processors and Resources will generally be corralled together to form a processing pipeline which will read data in from one or more sources and do some processing on it. However, the pipeline concept is not defined in this module so as to leave users free to combine the components in any manner they wish.