go-pkg-xmlx/document.go

281 lines
6.8 KiB
Go
Raw Normal View History

// This work is subject to the CC0 1.0 Universal (CC0 1.0) Public Domain Dedication
// license. Its contents can be found at:
// http://creativecommons.org/publicdomain/zero/1.0/
/*
This package wraps the standard XML library and uses it to build a node tree of
any document you load. This allows you to look up nodes forwards and backwards,
as well as perform simple search queries.
Nodes now simply become collections and don't require you to read them in the
order in which the xml.Parser finds them.
The Document currently implements 2 search functions which allow you to
look for specific nodes.
2010-05-06 03:36:48 +00:00
*xmlx.Document.SelectNode(namespace, name string) *Node;
*xmlx.Document.SelectNodes(namespace, name string) []*Node;
2010-05-06 03:36:48 +00:00
SelectNode() returns the first, single node it finds matching the given name
and namespace. SelectNodes() returns a slice containing all the matching nodes.
2010-05-06 03:36:48 +00:00
Note that these search functions can be invoked on individual nodes as well.
This allows you to search only a subset of the entire document.
*/
package xmlx
import (
"os"
"io"
"io/ioutil"
"path"
"strings"
"xml"
"fmt"
"http"
iconv "github.com/sloonz/go-iconv/src"
)
type Document struct {
2010-05-06 03:36:48 +00:00
Version string
Encoding string
StandAlone string
SaveDocType bool
Root *Node
Entity map[string]string
Verbose bool
}
func New() *Document {
return &Document{
2010-05-06 03:36:48 +00:00
Version: "1.0",
Encoding: "utf-8",
StandAlone: "yes",
SaveDocType: true,
2010-05-06 03:36:48 +00:00
Entity: make(map[string]string),
Verbose: false,
}
}
// This loads a rather massive table of non-conventional xml escape sequences.
// Needed to make the parser map them to characters properly. It is advised to
// set only those entities needed manually using the document.Entity map, but
// if need be, this method can be called to fill the map with the entire set
// defined on http://www.w3.org/TR/html4/sgml/entities.html
func (this *Document) LoadExtendedEntityMap() { loadNonStandardEntities(this.Entity) }
func (this *Document) String() string {
2010-05-06 03:36:48 +00:00
s, _ := this.SaveString()
return s
}
func (this *Document) SelectNode(namespace, name string) *Node {
2010-05-06 03:36:48 +00:00
return this.Root.SelectNode(namespace, name)
}
func (this *Document) SelectNodes(namespace, name string) []*Node {
2010-05-06 03:36:48 +00:00
return this.Root.SelectNodes(namespace, name)
}
// *****************************************************************************
// *** Satisfy ILoader interface
// *****************************************************************************
func (this *Document) LoadString(s string) (err os.Error) {
// Ensure we are passing UTF-8 encoding content to the XML tokenizer.
if s, err = this.correctEncoding(s); err != nil {
return
}
// tokenize data
2010-05-06 03:36:48 +00:00
xp := xml.NewParser(strings.NewReader(s))
xp.Entity = this.Entity
2009-11-23 16:50:29 +00:00
2010-05-06 03:36:48 +00:00
this.Root = NewNode(NT_ROOT)
ct := this.Root
var tok xml.Token
var t *Node
var i int
var doctype string
var v xml.Attr
for {
if tok, err = xp.Token(); err != nil {
if err == os.EOF {
return nil
}
if this.Verbose {
fmt.Fprintf(os.Stderr, "Xml Error: %s\n", err)
}
return err
}
switch tt := tok.(type) {
case xml.SyntaxError:
return os.NewError(tt.String())
case xml.CharData:
ct.Value = strings.TrimSpace(string([]byte(tt)))
case xml.Comment:
2010-05-06 03:36:48 +00:00
t := NewNode(NT_COMMENT)
t.Value = strings.TrimSpace(string([]byte(tt)))
2010-05-06 03:36:48 +00:00
ct.AddChild(t)
case xml.Directive:
t = NewNode(NT_DIRECTIVE)
t.Value = strings.TrimSpace(string([]byte(tt)))
2010-05-06 03:36:48 +00:00
ct.AddChild(t)
case xml.StartElement:
t = NewNode(NT_ELEMENT)
t.Name = tt.Name
t.Attributes = make([]*Attr, len(tt.Attr))
for i, v = range tt.Attr {
t.Attributes[i] = new(Attr)
2010-05-06 03:36:48 +00:00
t.Attributes[i].Name = v.Name
t.Attributes[i].Value = v.Value
}
2010-05-06 03:36:48 +00:00
ct.AddChild(t)
ct = t
case xml.ProcInst:
if tt.Target == "xml" { // xml doctype
doctype = strings.TrimSpace(string(tt.Inst))
if i = strings.Index(doctype, `standalone="`); i > -1 {
this.StandAlone = doctype[i+len(`standalone="`) : len(doctype)]
i = strings.Index(this.StandAlone, `"`)
this.StandAlone = this.StandAlone[0:i]
}
} else {
t = NewNode(NT_PROCINST)
t.Target = strings.TrimSpace(tt.Target)
t.Value = strings.TrimSpace(string(tt.Inst))
2010-05-06 03:36:48 +00:00
ct.AddChild(t)
}
case xml.EndElement:
if ct = ct.Parent; ct == nil {
return
}
}
}
2010-05-06 03:36:48 +00:00
return
}
func (this *Document) LoadFile(filename string) (err os.Error) {
var data []byte
if data, err = ioutil.ReadFile(path.Clean(filename)); err != nil {
return
}
return this.LoadString(string(data))
}
func (this *Document) LoadUri(uri string) (err os.Error) {
2010-12-16 20:08:01 +00:00
var r *http.Response
if r, _, err = http.Get(uri); err != nil {
return
}
2010-05-06 03:36:48 +00:00
defer r.Body.Close()
var b []byte
if b, err = ioutil.ReadAll(r.Body); err != nil {
return
}
2010-12-16 20:08:01 +00:00
return this.LoadString(string(b))
}
2010-10-10 18:04:58 +00:00
func (this *Document) LoadStream(r io.Reader) (err os.Error) {
var b []byte
if b, err = ioutil.ReadAll(r); err != nil {
return
}
return this.LoadString(string(b))
}
// *****************************************************************************
// *** Satisfy ISaver interface
// *****************************************************************************
func (this *Document) SaveFile(path string) (err os.Error) {
var data string
if data, err = this.SaveString(); err != nil {
return
}
return ioutil.WriteFile(path, []byte(data), 0600)
}
func (this *Document) SaveString() (s string, err os.Error) {
if this.SaveDocType {
s = fmt.Sprintf(`<?xml version="%s" encoding="%s" standalone="%s"?>`,
this.Version, this.Encoding, this.StandAlone)
}
2010-05-06 03:36:48 +00:00
s += this.Root.String()
return
}
2010-10-10 18:04:58 +00:00
func (this *Document) SaveStream(w io.Writer) (err os.Error) {
2010-12-16 20:08:01 +00:00
var s string
if s, err = this.SaveString(); err != nil {
return
}
2010-12-16 20:08:01 +00:00
_, err = w.Write([]byte(s))
2010-05-06 03:36:48 +00:00
return
}
// Use libiconv to ensure we get UTF-8 encoded data. The Go Xml tokenizer will
// throw a tantrum if we give it anything else.
func (this *Document) correctEncoding(data string) (ret string, err os.Error) {
var cd *iconv.Iconv
var tok xml.Token
enc := "utf-8"
xp := xml.NewParser(strings.NewReader(data))
xp.Entity = this.Entity
loop:
for {
if tok, err = xp.Token(); err != nil {
if err == os.EOF {
break loop
}
return "", err
}
switch tt := tok.(type) {
case xml.ProcInst:
if tt.Target == "xml" { // xml doctype
var pair []string
var entry string
list := strings.Split(string(tt.Inst), " ", -1)
for _, entry = range list {
if pair = strings.Split(entry, "=", -1); len(pair) < 2 {
continue
}
switch pair[0] {
case "encoding":
enc = pair[1][1 : len(pair[1])-1]
break loop
}
}
}
}
}
if strings.ToLower(enc) == "utf-8" {
// Data already in utf-8 format. Nothing to do here.
return data, nil
}
if cd, err = iconv.Open("utf-8", enc); err != nil {
return
}
defer cd.Close()
return cd.Conv(data)
}