Scales Xml 0.5.0 Overview

Scales Xml, in addition to immutability, aims to focus on these separate concerns:

Correctness

• XmlItems such as Text, Comments, PIs and CData are not containing nodes, they cannot have children and they cannot contain invalid characters
• QNames are a fundamental piece of XML and are treated as first class citizens
• Elements are a QName, the attributes they contain, the namespaces they declare and also do not directly have children
• Attributes are combinations of either a PrefixedQName or NoNamespaceQName and a value

Structure and manipulation

• Structure of XML (Tree) is separated from the contents of the Xml, allowing maximal re-use
• Operations on XML (Path - a Zipper over Tree) is separated from both the contents and the structure
• Changing any update to individual XML Element or Items will not cause cascading changes in the rest of the structure (function of Path)
• Transformations on XML Trees (via Path) should be composable and simple to follow.
• The data model is the same for both Push and Pull parsing (with the addition of EndElem for XmlPull)

XML Comprehensions - XPath

• Navigation through XML is either through Path directly or the inbuilt XPath syntax
• Syntax closely resembles normal XPath leveraging, for example, Paths ability to navigate parents
• Syntax should separate the navigation of elements from available predicates as far as possible (niceties such as \*QName remain)
• Syntax should not rely on string interpretation but on types (QName objects used directly)
  • String based XPath 1.0 querying based on Jaxen is however available
• XPaths are used as a basis for Tree transformation
• The full XPath axe are available (with the exception of the namespace axis)

Pull API

• Based on javax.xml.stream, allowing optimisations from jaxp/stax upgrades
• Uses the same types as the SAX based parser
• Leverages Scalaz Iteratees to provide processing of the event stream
  • You can choose how to process an xml stream, should your code control the flow, the enumerator or a mixture?
  • Immutable transformations on streams
• Provides an Iteratee to process XPath like locations via a

  List[QName]

• Provides a combinator (onDone) to iterate over ResumableIters allowing simplified XML parsing
• Non blocking IO based parsing via Aalto
• Scales adds the concept of resumable iteration:

  // think resumable folds without cps plugin
  type ResumableIter[E,A] = IterV[E, (A, IterV[E,_])]

Comprehensive Serialisation Support

• SerializerFactory / Serializer Typeclasses allow plugable serialisation schemes
• SerializeableXml Typeclass allows serializing of all XML through a single consistent interface
  • Lazily create XML using EphemeralStreams and Iterator[PullType]
• Developers can customise the serialisation process while maintaining correctness
  • Use XHTML style tag ends <el /> for empty elements
  • Serialise your attributes in an order of your choosing
  • Or choose to abandon certain correctness checks in favour of speed when you know that your XML inputs and outputs are always correct

JAXP Support

• XSLT transformation support
• Conversion to other XML DOMs
• Use of JAXP serialisation - e.g. pretty printing
• javax.xml.validation support
• Aims to hide all JAXP implementation inconsistencies

Xml Equality

• Compare xml items against each other
• Full control over QName handling
• Full control over QName values (attributes or text)
• Provides information about what is different and where
• Provides a simple conversion to scalaz.Equal

Performance

• Up to 20% faster than scala.xml
• Lower memory usage than both scala.xml and JAXP/DOM
• Parsing memory usage can be optimised for given usage