Comparison Pipeline Architecture

The DeltaNova comparison pipelines are built around a central comparator engine that processes XML inputs through a series of configurable filter stages. Understanding this architecture helps you integrate DeltaNova effectively and extend its behaviour for your specific use case.

Pipeline Structure

DeltaNova's has three main comparators:

Input filters - transform each input document before comparison.
Comparison engine - aligns and annotates differences between inputs.
Output filters - transform the delta result into the required format.

Each phase supports extension points where you can insert custom XSLT or Java filters.

Comparator Classes

Three comparator classes are available, each building on the same core pipeline.

Comparators	Best for	Configuration format
`Document Comparator`	Text-centric XML documents	DCP (Document Comparator Pipeline)
`DataComparator`	Data-centric XML files	DTCP (Data Comparator Pipeline)
`Pipelined Comparator`	Maximum flexibility	DXP (DeltaXML Pipeline)

Extension Points

The Document and Data Comparators exposes named extension points where filters can be inserted.

inputPreFlatteningPoint - before element flattening, for both inputs.
inputAExtensionPoints / inputBExtensionPoints - after element flattening, per input.
inputExtensionPoints - after element flattening, both inputs.
postTablePoint / preTablePoint - around table processing.
preAttributePoint - before attribute processing.
outputExtensionPoints / finalPoint - on the output stream.