xmltree/graph.go

package graph

import (
	"errors"
	"log"
)

var attributeNameLaw map[rune]struct{} = make(map[rune]struct{})

func init() {
	for _, c := range "abcdefghijklmnopqrstuvwsyzABCDEFGHIJKLMNOPQRSTUVWSYZ-_" {
		attributeNameLaw[c] = struct{}{}
	}
}

type Range struct {
	Start int
	End   int
}

type ElementTag struct {
	Range
}

type Element struct {
	Range
	Tag        *ElementTag
	Attributes []*Attribute
	Source     []rune
	Children   []*Element
}

type Attribute struct {
	Range
	Key   Range
	Value Range
}

func (e *Element) GetTagName() string {
	return string(e.Source[e.Tag.Start:e.Tag.End])
}

func GetTag(source []rune, cur int) (*ElementTag, int) {

	for i := cur; i < len(source); i++ {
		c := source[i]
		if c == ' ' {
			return &ElementTag{Range{cur, i - 1}}, i
		}
	}

	return nil, -1
}

func SkipAttribute(source []rune, cur int) int {
	for i := cur; i < len(source); i++ {
		c := source[i]
		if c == ' ' || c == '>' {
			return i - 1
		}
	}
	panic(errors.New("over the range of the next attribute"))
}

func GetAttribute(source []rune, cur int) *Range {

}

func GetAttributes(source []rune, cur int) []*Attribute {

	var attrs []*Attribute

LOOP_TOP:
	for i := cur; i < len(source); i++ {

		c := source[i]
		switch c {
		case '>':
			break LOOP_TOP
		case '/':
			if source[i+1] == '>' {
				break LOOP_TOP
			}
		case ' ':
			continue
		default:

			attr := &Attribute{}
			attr.Range.Start = i

			for ; i < len(source); i++ {
				c = source[i]
				if _, ok := attributeNameLaw[c]; !ok {
					i = SkipAttribute(source, i)
					break
				} else {
					if c == ' ' || c == '>' {
						attr.Range.End = i
						break
					}
				}
			}

		}

	}

	return attrs
}

func ParseChild(parent *Element, cur int) *Element {

	start := cur
	end := start

	child := &Element{}
	child.Tag, cur = GetTag(parent.Source, start)
	GetAttributes(parent.Source, cur)

	for i := cur; i < len(parent.Source); i++ {

		c := parent.Source[i]
		if c == '>' {
			end = i - 1
			return child
		}

	}

	return nil
}

func GetChildren(parent *Element) {
	// GetChild
	for i := parent.Range.Start; i < parent.Range.End; i++ {
		c := parent.Source[i]

		if c == '<' {
			tag := ParseChild(parent, i)
			log.Println(tag)
		}
	}
}

func ParseXML(data []byte) *Element {
	source := []rune(string(data))

	root := &Element{tag: nil, Source: source, Pos: Range{0, len(source)}}

	return root
}