Overview of Document Preprocessor

This component provides a CdaCasInitializer SECTION annotator that transforms a Clinical Document Architecture (CDA) document into plain text, provided the CDA document conforms to the DTD schema.

As part of the conversion to plain text, section (segment) markers are inserted into the text and hyphens are inserted into words that should be hyphenated. The resulting text is stored in a new View, which has its own Sofa.

Sections are detected and Segment (also called "section") annotations are added to the CAS. Document level data is extracted and stored in the CAS as Property annotations.

This does not handle all CDA documents. The CDA document must conform to the DTD resources/cda/NotesIIST_RTF.DTD.

Analysis engines (annotators)


The file cTAKESdesc/docpredesc/AggregateAE.xml defines a pipeline for preprocessing documents. The pipeline is a simple pipeline with only one delegate analysis engine (one annotator), the CdaCasInitializer, and is included for testing. Typically the CdaCasInitializer.xml descriptor is included in a more complete pipeline rather than using the AggregateAE.xml descriptor that is in this project.


The CdaCasInitializer descriptor defines the analysis engine (annotator) for preprocessing documents. It creates a plain text view from a CDA view. The plain text view can then be annotated, using other components, for tokens, parts of speech, chunks, etc.


  • No labels