Annot8

Annot8 is a framework for extracting structured information from unstructured data. As a simple example, this could be extracting e-mail addresses from a collection of text documents, or faces from a collection of images.

Terminology

• Annotation - Some information highlighted in Content, for example a span of text
• Content - A “view” on an Item, for example the text extracted from a Word document
• Item - A data object for processing, for example a Word document
• Processor - A processor is an Annot8 component that processes Content in some way, for example annotating e-mail addresses
• Property - A property is a key-value pair on an Annot8 object (such as an Annotation, an Item or Content) which provides additional meta-data, for example the author of a document
• Source - An Annot8 component which creates new Item objects, for example by finding files in a folder on your hard drive

Processor and Source are known as components, and components are generally formed of three separate classes:

• Processor or Source - the code that actually does the work
• Descriptor - code to describe and create new components
• Settings (optional) - configuration information for the component (generally used by the descriptor when creating new instances)

Core Repository

This is the core Annot8 repository, which contains the actual framework. This repository does not contain any implementations of components, these are in the Annot8 Components repository.

This repository contains 4 sub-projects

• annot8-api - The core interfaces (API) used by Annot8
• annot8-common - Common tools, classes and helpers
• annot8-implementations - The standard implementation of Annot8 classes
• annot8-testing - Implementations and utilities useful when testing

Licence

Code in this repository is licenced under the Apache Software Licence 2. See the NOTICE file for any additional restrictions.