TRSX overview

TRSX is short for “TRaining Set XML.”

TRSX provides developers flexibility in managing their model data. With TRSX, you can manage an entire model in a single file outside of Mix and import the model into a Mix.nlu project. You can also manage training data in separate TRSX files and import them individually.

TRSX is designed to be a universal file format. Nuance will maintain and update the format with additional features and also add new functionality to Mix.nlu for handing TRSX.

The TRSX specification is defined, owned, and maintained by Nuance Communications, Inc.

A TRSX file has the following main sets of data:

  • Metadata
    • Entry nodes with key-value pairs
  • Sources
    • List of sources used to label data
  • Ontology
    • Intents
    • Entity
    • Intent/Entity relationships, known as Links
    • Entity/Entity relationships, defined as Relations
  • Dictionaries
    • List of entities included in a List type entity
  • Samples (the training set)
    • Training samples with annotations for entities

The TRSX specification allows for representing a complete Mix.nlu project. Keeping all aspects of the project in a single file is possible and may be simpler in some situations. However, if you’re dealing with larger projects, it might be worth keeping the ontology separate (and shared across languages, for example, ontology.trsx.xml) as well as having separate TRSX files for dictionaries (music-list.trsx.xml) and samples (samples-group-a.trsx.xml, samples-group-b.trsx.xml).

This section describes TRSX version 2.6, which is currently deployed in the Mix environment.