How can I validate documents against Schematron schemas in Java?

Question

As far as I can tell, JAXP by default supports W3C XML Schema and RelaxNG from Java 6 .

I can see a few APIs, mostly experimental or incomplete, on the schematron.com links page .

Is there an approach on validating schematron in Java that's complete, efficient and can be used with the JAXP API?

Answer 1

Jing supports pre-ISO Schematron validation (note that Jing's implementation is based also on XSLT).

There are also XSLT implementations that can be very easily invoked from Java. You need to decide what version of Schematron you are interested in and then get the corresponding stylesheet - all of them should be available from schematron.com. The process is very simple simple, involving basically 2 steps:

• apply the skeleton XSLT on your Schematron schema to get a new XSLT stylesheet that represents your Schematron schema in XSLT
• apply the obtained XSLT on your instance document or documents to validate them

JAXP is just an API and it does not require support for Relax NG from an implementation. You need to check the specific implementation that you use to see if that supports Relax NG or not.

Answer 2

纯Java Schematron实现位于https://github.com/phax/ph-schematron/它支持XSLT方法和纯Java方法。

Answer 3

You can check out SchematronAssert (disclosure: my code). It is meant primarily for unit testing, but you may use it for normal code too. It is implemented using XSLT.

Unit testing example:

ValidationOutput result = in(booksDocument)
    .forEvery("book")
    .check("author")
    .validate();
assertThat(result).hasNoErrors();

Standalone validation example:

StreamSource schemaSource = new StreamSource(... your schematron schema ...);
StreamSource xmlSource = new StreamSource(... your xml document ... );
StreamResult output = ... here your SVRL will be saved ... 
// validation 
validator.validate(xmlSource, schemaSource, output);

Work with an object representation of SVRL:

ValidationOutput output = validator.validate(xmlSource, schemaSource);
// look at the output
output.getFailures() ... 
output.getReports() ...