XML/XPath: Difference between revisions

{{libheader|libxml2}}

Nim standard library provides several modules to process XML, but none to process XPath requests. There is no third party library to do that either.

So we provide here a solution which uses the C "libxml2" library. Nim allows to interface with C with a great flexibility. Declaring the external function is quite easy. The major part of the work consists in declaring the data types. In "libxml2" there are a lot of structures and we have translated only what is needed to solve the task (or a bit more actually).

For the first request, we use <code>//section[1]/item[1]</code> instead of <code>//item[1]</code> (as the Python version does). The latter request works but returns the list of the first elements in each section which doesn’t seem to be what is expected.

<lang Nim>import sequtils, strutils

const LibXml = "libxml2.so"

type

XmlDocPtr = pointer

XmlXPathContextPtr = pointer

XmlElementKind = enum

xmlElementNode = 1

xmlAttributeNode = 2

xmlTextNode = 3

xmlCdataSectionNode = 4

xmlEntityRefNode = 5

xmlEntityNode = 6

xmlPiNode = 7

xmlCommentNode = 8

xmlDocumentNode = 9

xmlDocumentTypeNode = 10

xmlDocumentFragNode = 11

xmlNotationNode = 12

xmlHtmlDocumentNode = 13

xmlDtdNode = 14

xmlElementDecl = 15

xmlAttributeDecl = 16

xmlEntityDecl = 17

xmlNamespaceDecl = 18

xmlXincludeStart = 19

xmlXincludeEnd = 20

XmlNsKind = XmlElementKind

XmlNsPtr = ptr XmlNs

XmlNs = object

next: XmlNsPtr

kind: XmlNsKind

href: cstring

prefix: cstring

private: pointer

context: XmlDocPtr

XmlAttrPtr = pointer

XmlNodePtr = ptr XmlNode

XmlNode = object

private: pointer

kind: XmlElementKind

name: cstring

children: XmlNodePtr

last: XmlNodePtr

parent: XmlNodePtr

next: XmlNodePtr

prev: XmlNodePtr

doc: XmlDocPtr

ns: XmlNsPtr

content: cstring

properties: XmlAttrPtr

nsDef: XmlNsPtr

psvi: pointer

line: cushort

extra: cushort

XmlNodeSetPtr = ptr XmlNodeSet

XmlNodeSet = object

nodeNr: cint

nodeMax: cint

nodeTab: ptr UncheckedArray[XmlNodePtr]

XmlPathObjectKind = enum

xpathUndefined

xpathNodeset

xpathBoolean

xpathNumber

xpathString

xpathPoint

xpathRange

xpathLocationset

xpathUsers

xpathXsltTree

XmlXPathObjectPtr = ptr XmlXPathObject

XmlXPathObject = object

kind: XmlPathObjectKind

nodeSetVal: XmlNodeSetPtr

boolVal: cint

floatVal: cdouble

stringVal: cstring

user: pointer

index: cint

user2: pointer

index2: cint

XmlSaveCtxtPtr = pointer

XmlBufferPtr = pointer

# Declaration of needed "libxml2" procedures.

proc xmlParseFile(docName: cstring): XmlDocPtr

{.cdecl, dynlib: LibXml, importc: "xmlParseFile".}

proc xmlXPathNewContext(doc: XmlDocPtr): XmlXPathContextPtr

{.cdecl, dynlib: LibXml, importc: "xmlXPathNewContext".}

proc xmlXPathEvalExpression(str: cstring; ctxt: XmlXPathContextPtr): XmlXPathObjectPtr

{.cdecl, dynlib: LibXml, importc: "xmlXPathEvalExpression".}

proc xmlXPathFreeContext(ctxt: XmlXPathContextPtr)

{.cdecl, dynlib: LibXml, importc: "xmlXPathFreeContext".}

proc xmlXPathFreeObject(obj: XmlXPathObjectPtr)

{.cdecl, dynlib: LibXml, importc: "xmlXPathFreeObject".}

proc xmlSaveToBuffer(vuffer: XmlBufferPtr; encoding: cstring; options: cint): XmlSaveCtxtPtr

{.cdecl, dynlib: LibXml, importc: "xmlSaveToBuffer".}

proc xmlBufferCreate(): XmlBufferPtr

{.cdecl, dynlib: LibXml, importc: "xmlBufferCreate".}

proc xmlBufferFree(buf: XmlBufferPtr)

{.cdecl, dynlib: LibXml, importc: "xmlBufferCreate".}

proc xmlBufferContent(buf: XmlBufferPtr): cstring

{.cdecl, dynlib: LibXml, importc: "xmlBufferContent".}

proc xmlSaveTree(ctxt: XmlSaveCtxtPtr; cur: XmlNodePtr): clong

{.cdecl, dynlib: LibXml, importc: "xmlSaveTree".}

proc xmlSaveClose(ctxt: XmlSaveCtxtPtr)

{.cdecl, dynlib: LibXml, importc: "xmlSaveClose".}

proc `$`(node: XmlNodePtr): string =

## Return the representation of a node.

let buffer = xmlBufferCreate()

let saveContext = xmlSaveToBuffer(buffer, nil, 0)

discard saveContext.xmlSaveTree(node)

saveContext.xmlSaveClose()

result = $buffer.xmlBufferContent()

xmlBufferFree(buffer)

iterator nodes(xpath: string; context: XmlXPathContextPtr): XmlNodePtr =

## Yield the nodes which fit the XPath request.

let xpathObj = xmlXPathEvalExpression(xpath, context)

if xpathObj.isNil:

quit "Failed to evaluate XPath: " & xpath, QuitFailure

assert xpathObj.kind == xpathNodeset

let nodeSet = xpathObj.nodeSetVal

if not nodeSet.isNil:

for i in 0..<nodeSet.nodeNr:

yield nodeSet.nodeTab[i]

xmlXPathFreeObject(xpathObj)

# Load and parse XML file.

let doc = xmlParseFile("xpath_test.xml")

if doc.isNil:

quit "Unable to load and parse document", QuitFailure

# Create an XPath context.

let context = xmlXPathNewContext(doc)

if context.isNil:

quit "Failed to create XPath context", QuitFailure

var xpath = "//section[1]/item[1]"

echo "Request $#:".format(xpath)

for node in nodes(xpath, context):

echo node

echo()

xpath = "//price/text()"

echo "Request $#:".format(xpath)

for node in nodes(xpath, context):

echo node.content

echo()

xpath = "//name"

echo "Request $#:".format(xpath)

let names = toSeq(nodes(xpath, context)).mapIt(it.children.content)

echo names

xmlXPathFreeContext(context)</lang>

{{out}}

<pre>Request //section[1]/item[1]:

<item upc="123456789" stock="12">

<name>Invisibility Cream</name>

<price>14.50</price>

<description>Makes you invisible</description>

</item>

Request //price/text():

14.50

23.99

4.95

3.56

Request //name:

@["Invisibility Cream", "Levitation Salve", "Blork and Freen Instameal", "Grob winglets"]</pre>