How is validate parsing and processing XML, Schema, and Schematron.

[jordanpadams]

Some technical deep dive questions:

1. How does validate read in XML? What parser is it using?
2. Is it using the same parser to read schema, schematron, and label?
3. What errors is it detecting and how?
4. What Java library is "applying those schema/schematron"?
5. How does the caching of the schemas and schematrons work?
6. Do we transform the schematron with the stylesheet with every file that is processed? Or is the stylesheet cached?

[jordanpadams]

1. How does validate read in XML? What parser is it using?

NASA-PDS/validate uses Java's standard XML parsing libraries:

• Label XML files are parsed using both SAX (Simple API for XML) and DOM (Document Object Model) parsers.
  • The code creates a SAXParserFactory and an XMLReader for parsing.
  • It also uses a DocumentBuilder to build a DOM document from the input stream.
• Example:

  SAXParserFactory saxParserFactory = SAXParserFactory.newInstance();
  saxParserFactory.setNamespaceAware(true);
  saxParserFactory.setXIncludeAware(Utility.supportXincludes());
  saxParserFactory.setValidating(false);
  XMLReader reader = saxParserFactory.newSAXParser().getXMLReader();

• Error Handling is performed by custom error handlers (LabelErrorHandler, XMLErrorListener) that convert exceptions into validation problems.

---

2. Is it using the same parser to read schema, schematron, and label?

• Labels: Parsed with SAX/DOM parsers.
• Schemas (XSD): Validated using Java's SchemaFactory (usually Apache Xerces under the hood) for XML Schema.
• Schematrons: Read as XML, but then processed using XSLT transformations through a TransformerFactory (typically Saxon).

File Type	Parser/Processor Used
Label XML	SAXParser/DOMBuilder (Java standard library)
Schema XSD	SchemaFactory (JAXP, Xerces)
Schematron	XSLT Transformer (usually Saxon via JAXP API)

---

3. What errors is it detecting and how?

• XML Syntax Errors: Detected during parsing (well-formedness).
• Schema Validation Errors: Detected using SchemaFactory and ValidatorHandler, including structural and type errors.
• Schematron Validation Errors: Detected as assertion/report failures after XSLT transformation and application.
• Semantic Errors: Custom checks, e.g., missing xml-model PIs, missing or incorrect schematypens, version mismatches.

All errors are converted into ValidationProblem objects, which include severity, type, and location, and are reported via a ProblemHandler.

---

4. What Java library is "applying those schema/schematron"?

• Schema (XSD): Java's SchemaFactory for XMLConstants.W3C_XML_SCHEMA_NS_URI (Xerces).
• Schematron:
  • Uses an XSLT process:
    1. Transforms the .sch file into an XSLT stylesheet using the ISO Schematron skeleton (iso_svrl_for_xslt2.xsl).
    2. Applies the stylesheet to the XML via a TransformerFactory (Saxon is used by default).

The key Java class for schematron transformation is SchematronTransformer.

---

5. How does the caching of the schemas and schematrons work?

• Schema Caching:
  • Schemas are read once and cached as StreamSource instances.
  • Cached in memory for reuse (fields like userSchemaFiles, cachedLSResolver).
• Schematron Caching:
  • Transformed schematron XSLT stylesheets are cached in fields like cachedSchematron or userSchematronTransformers.
  • Only re-transformed if a new, previously unseen schematron file is encountered.

This avoids repeated I/O and transformation overhead for each validation operation.

---

6. Do we transform the schematron with the stylesheet with every file that is processed? Or is the stylesheet cached?

• The transformed XSLT stylesheet is cached and reused.
  • After the first transformation of a schematron file, the resulting XSLT is stored in memory and reused for subsequent validations.
  • Transformation is only repeated if a new schematron file is provided or user-specified transformers are supplied.

---

References

• LabelValidator.java
• SchemaValidator.java
• SchematronTransformer.java
• User guide: Operating validate

---

[jordanpadams]

warning: answers brought to you by GitHub co-pilot and may not be 100% accurate in all instances.