A system for text processing and information extraction
This object was generated by sbt-buildinfo.
This code shows how a custom Eidos reader can be created by configuring the standard one and then adjusting the components that it uses.
This code shows how a custom Eidos reader can be created by configuring the standard one and then adjusting the components that it uses. This is an alternative to modifying its behavior by inheritance and method overrides. It probably works best when it is entire components that switched on and off. Some of the components are turned on and off at build time via the conf file. The fate of others is controlled at object creation time by manipulation of components.
A system for text processing and information extraction
Eidos accepts text, uses the processors project (via a modified EidosProcessor) to tokenize and annotate it into a Document, tasks an engine from the odin project to extract Mentions from it with the help of local Actions and Finders (now called FinderRefiners), repeatedly refines the odin mentions with OdinRefiners, converts them into EidosMentions, refines those with EidosRefiners, and finally incorporates them into an AnnotatedDocument. For good measure there are also ProcessorRefiners that modify the initial Document.
OdinRefiners is a collection of functions that each refines (convert, transform, modifies) a Seq[Mention] (odin Mentions) to another Option[Seq[Mention]] which will be fed into the next refiner.
EidosRefiners do the same, but with EidosMentions, although they are hidden in an AnnotatedDocument: AnnotatedDocument => Option[AnnotatedDocument].
In the meantime there are also ProcessorRefiners which work on the Documents. See the refiners package.
The collections of refiners form a pipeline which can be configured at runtime or even be supplied to Eidos from elsewhere.