newXMLDoc {XML}

Create internal XML node or document object

Description

These are used to create internal ‘libxml’ nodes and top-level document objects that are used to write XML trees. While the functions are available, their direct use is not encouraged. Instead, use xmlTree as the functions need to be used within a strict regime to avoid corrupting C level structures.

xmlDoc creates a new XMLInternalDocument object by copying the given node and all of its descendants and putting them into a new document. This is useful when we want to work with sub-trees with general tools that work on documents, e.g. XPath queries.

newXMLDoc allows one to create a regular XML node with a name and attributes. One can provide new namespace definitions via namespaceDefinitions. While these might also be given in the attributes in the slightly more verbose form of c('xmlns:prefix' = 'http://...'), the result is that the XML node does not interpret that as a namespace definition but merely an attribute with a name 'xmlns:prefix'. Instead, one should specify the namespace definitions via the namespaceDefinitions parameter.

In addition to namespace definitions, a node name can also have a namespace definition. This can be specified in the name argument as prefix:name and newXMLDoc will do the right thing in separating this into the namespace and regular name. Alternatively, one can specify a namespace separately via the namespace argument. This can be either a simple name or an internal namespace object defined earlier.

How do we define a default namespace?

Usage

xmlDoc(node, addFinalizer = TRUE)
newXMLDoc(dtd = "", namespaces=NULL, addFinalizer = TRUE, 
           name = character(), node = NULL, isHTML = FALSE)
newHTMLDoc(dtd = "loose", addFinalizer = TRUE, name = character(), 
            node = newXMLNode("html",
                               newXMLNode("head", addFinalizer = FALSE), 
                               newXMLNode("body", addFinalizer = FALSE),
                              addFinalizer = FALSE)) 
newXMLNode(name, ..., attrs = NULL, namespace = character(),
            namespaceDefinitions = character(),
             doc = NULL, .children = list(...), parent = NULL,
	     at = NA, cdata = FALSE,
             suppressNamespaceWarning =
                 getOption("suppressXMLNamespaceWarning", FALSE),
             sibling = NULL, addFinalizer = NA,
              noNamespace = length(namespace) == 0 && !missing(namespace),
               fixNamespaces = c(dummy = TRUE, default = TRUE))
newXMLTextNode(text, parent = NULL, doc = NULL, cdata = FALSE, 
                escapeEntities = is(text, "AsIs"), addFinalizer = NA)
newXMLCDataNode(text, parent = NULL, doc = NULL, at = NA, sep = "\n",
                   addFinalizer = NA)
newXMLCommentNode(text, parent = NULL, doc = NULL, at = NA, addFinalizer = NA)
newXMLPINode(name, text, parent = NULL, doc = NULL, at = NA, addFinalizer = NA)
newXMLDTDNode(nodeName, externalID = character(),
              systemID = character(), doc = NULL, addFinalizer = NA)  

Arguments

node

a XMLInternalNode object that will be copied to create a subtree for a new document.

dtd

the name of the DTD to use for the XML document. Currently ignored!

namespaces

a named character vector with each element specifying a name space identifier and the corresponding URI for that namespace that are to be declared and used in the XML document, \ e.g. c(shelp = "https://www.omegahat.net/XML/SHelp")

addFinalizer

a logical value indicating whether the default finalizer routine should be registered to free the internal xmlDoc when R no longer has a reference to this external pointer object. This can also be the name of a C routine or a reference to a C routine retrieved using getNativeSymbolInfo.

name

the tag/element name for the XML node and the for a Processing Instruction (PI) node, this is the "target", e.g. the identifier for the system for whose attention this PI node is intended.

...

the children of this node. These can be other nodes created earlier or R strings that are converted to text nodes and added as children to this newly created node.

attrs

a named list of name-value pairs to be used as attributes for the XML node. One should not use this argument to define namespaces, i.e. attributes of the form xmlns:prefix='http:/...'. Instead, such definitions should be specified ideally via the namespaceDefinitions argument, or even the namespace argument. The reason is that namespace definitions are special attributes that are shared across nodes wherease regular attributes are particular to a node. So a namespace needs to be explicitly defined so that the XML representation can recognize it as such.

namespace

a character vector specifying the namespace for this new node. Typically this is used to specify i) the prefix of the namespace to use, or ii) one or more namespace definitions, or iii) a combination of both. If this is a character vector with a) one element and b) with an empty names attribute and c) whose value does not start with http:/ or ftp:/, then it is assumed that the value is a namespace prefix for a namespace defined in an ancestor node. To be able to resolve this prefix to a namespace definition, parent must be specified so that we can traverse the chain of ancestor nodes. However, if c) does not hold, i.e. the string starts with http:/ or ftp:/, then we take this single element to be a namespace definition and the since it has no name b), this is the definition for the default namespace for this new node, i.e. corresponding to xmlns='http:/...'. It is cumbersome to specify "" as a name for an element in a character vector (as c('' = 'value') gives an unnecessary error!. Elements with names are expanded to namespace definitions with the name as the prefix and the value as the namespace URI.

doc

the XMLInternalDocument object created with newXMLDoc that is used to root the node.

.children

a list containing XML node elements or content. This is an alternative form of specifying the child nodes than ... which is useful for programmatic interaction when the "sub"-content is already in a list rather than a loose collection of values.

text

the text content for the new XML node

nodeName

the name of the node to put in the DOCTYPE element that will appear as the top-most node in the XML document.

externalID

the PUBLIC identifier for the document type. This is a string of the form A//B//C//D. A is either + or -; B identifies the person or insitution that defined the format (i.e. the "creator"); C is the name of the format; and language is an encoding for the language that comes from the ISO 639 document.

systemID

the SYSTEM identifier for the DTD for the document. This is a URI

namespaceDefinitions

a character vector or a list with each element being a string. These give the URIs identifying the namespaces uniquely. The elements should have names which are used as prefixes. A default namespace has "" as the name. This argument can be used to remove any ambiguity that arises when specifying a single string with no names attribute as the value for namespace. The values here are used only for defining new namespaces and not for determining the namespace to use for this particular node.

parent

the node which will act as the parent of this newly created node. This need not be specified and one can add the new node to another node in a separate operation via addChildren.

sibling

if this is specified (rather than parent) this should be an XMLInternalNode and the new node is added as a sibling of this node, after this node, i.e. to the right. This is just a convenient form of parent = xmlParent(node).

cdata

a logical value indicating whether to enclose the text within a CDATA node (TRUE) or not (FALSE). This is a convenience mechanism to avoid having to create the text node and then the CDATA node. If one is not certain what characters are in the text, it is useful to use TRUE to ensure that they are “escaped”.

It is an argument for newXMLNode as the child nodes can be given as simple strings and are converted to text nodes. This cdata value is passed to the calls to create these text nodes and so controls whether they are enclosed within CDATA nodes.

suppressNamespaceWarning

see addChildren

at

this allows one to control the position in the list of children at which the node should be added. The default means at the end and this can be any position from 0 to the current number of children.

sep

when adding text nodes, this is used as an additional separator text to insert between the specified strings.

escapeEntities

a logical value indicating whether to mark the internal text node in such a way that protects characters in its contents from being escaped as entities when being serialized via saveXML

noNamespace

a logical value that allows the caller to specify that the new node has no namespace. This can avoid searching parent and ancestor nodes up the tree for the default namespace.

isHTML

a logical value that indicates whether the XML document being created is HTML or generic XML. This helps to create an object that is identified as an HTML document.

fixNamespaces

a logical vector controlling how namespaces in child nodes are to be processed. The two entries should be named dummy and default. The dummy element controls whether we process child nodes that have a namespace which was not defined when the node was created. These are created as “dummy” namespaces and can be resolved now that the parent node is defined and the name space may be defined. When we know it is not yet defined, but will be defined in an ancestor node, we can turn off this processing with a value of FALSE.

The default element controls how we process the child nodes and give them the default name space defined in the parent or ancestor nodes.

Details

These create internal C level objects/structure instances that can be added to a libxml DOM and subsequently inserted into other document objects or “serialized” to textual form.

Value

Each function returns an R object that points to the C-level structure instance. These are of class XMLInternalDocument and XMLInternalNode, respectively

Note

These functions are used to build up an internal XML tree. This can be used in the Sxslt package (https://www.omegahat.net/Sxslt/) when creating content in R that is to be dynamically inserted into an XML document.

Author(s)

Duncan Temple Lang

References

https://www.w3.org/XML/, http://www.xmlsoft.org, https://www.omegahat.net

See Also

xmlTree saveXML

Examples

doc = newXMLDoc()

 # Simple creation of an XML tree using these functions
top = newXMLNode("a")
newXMLNode("b", attrs = c(x = 1, y = 'abc'), parent = top)
newXMLNode("c", "With some text", parent = top)
d = newXMLNode("d", newXMLTextNode("With text as an explicit node"), parent = top)
newXMLCDataNode("x <- 1\n x > 2", parent = d)

newXMLPINode("R", "library(XML)", top)
newXMLCommentNode("This is a comment", parent = top)

o = newXMLNode("ol", parent = top)

kids = lapply(letters[1:3],
               function(x)
                  newXMLNode("li", x))
addChildren(o, kids)

cat(saveXML(top))


x = newXMLNode("block", "xyz", attrs = c(id = "bob"),
                      namespace = "fo",
                      namespaceDefinitions = c("fo" = "http://www.fo.org"))

xmlName(x, TRUE) == "fo"

  # a short cut to define a name space and make it the prefix for the
  # node, thus avoiding repeating the prefix via the namespace argument.
x = newXMLNode("block", "xyz", attrs = c(id = "bob"),
                      namespace = c("fo" = "http://www.fo.org"))


 # name space on the attribute
x = newXMLNode("block", attrs = c("fo:id" = "bob"),
                      namespaceDefinitions = c("fo" = "http://www.fo.org"))




x = summary(rnorm(1000))
d = xmlTree()
d$addNode("table", close = FALSE)

d$addNode("tr", .children = sapply(names(x), function(x) d$addNode("th", x)))
d$addNode("tr", .children = sapply(x, function(x) d$addNode("td", format(x))))

d$closeNode()


# Just doctype
z = xmlTree("people", dtd = "people")
# no public element
z = xmlTree("people", dtd = c("people", "", "https://www.omegahat.net/XML/types.dtd"))
# public and system
z = xmlTree("people", dtd = c("people", "//a//b//c//d", "https://www.omegahat.net/XML/types.dtd"))

# Using a DTD node directly.
dtd = newXMLDTDNode(c("people", "", "https://www.omegahat.net/XML/types.dtd"))
z = xmlTree("people", dtd = dtd)


x = rnorm(3)
z = xmlTree("r:data", namespaces = c(r = "http://www.r-project.org"))
z$addNode("numeric", attrs = c("r:length" = length(x)), close = FALSE)
lapply(x, function(v) z$addNode("el", x))
z$closeNode()
# should give   <r:data><numeric r:length="3"/></r:data>


# shows namespace prefix on an attribute, and different from the one on the node.
z = xmlTree()
z$addNode("r:data",
         namespace = c(r = "http://www.r-project.org",
                       omg = "https://www.omegahat.net"),
         close = FALSE)
x = rnorm(3)
z$addNode("r:numeric", attrs = c("omg:length" = length(x)))


z = xmlTree("people", namespaces = list(r = "http://www.r-project.org"))
z$setNamespace("r")

z$addNode("person", attrs = c(id = "123"), close = FALSE)
z$addNode("firstname", "Duncan")
z$addNode("surname", "Temple Lang")
z$addNode("title", "Associate Professor")
z$addNode("expertize", close = FALSE)
z$addNode("topic", "Data Technologies")
z$addNode("topic", "Programming Language Design")
z$addNode("topic", "Parallel Computing")
z$addNode("topic", "Data Visualization")
z$closeTag()
z$addNode("address", "4210 Mathematical Sciences Building, UC Davis")



   # 
txt = newXMLTextNode("x &lt; 1")
txt # okay
saveXML(txt) # x &amp;lt; 1

   # By escaping the text, we ensure the entities don't
   # get expanded, i.e. &lt; doesn't become &amp;lt;
txt = newXMLTextNode(I("x &lt; 1"))
txt # okay
saveXML(txt) # x &lt; 1


newXMLNode("r:expr", newXMLTextNode(I("x < 1")),
            namespaceDefinitions = c(r = "http://www.r-project.org"))