Initial Release
This demo application can be used to demonstrate how you can leverage classic iKnow matching to build composite “rules” that represent particular patterns to be extracted from natural language text. Dictionary items can be assembled into sequences of which some elements may be optional and others aren’t, and take into account the negation context of matched elements.
The application also includes code to “parse” the paragraph structure of texts with various numbering patterns, which might be usable for other scenarios, independent of the rules building functionalities. The paragraph numbering detection allows for several typical OCR issues (”11.” rather than “ii.”). See the code in Demo.RulesBuilder.Paragraph:ParseString() for more details.
For a detailed description of the GUI, check this article on the InterSystems Developer Community: https://community.intersystems.com/post/iknow-demo-apps-part-4-rules-builder-demo