vkryukov/remove_html_attribute.go

## remove_html_attribute.go
package main

import (
	"bytes"
	"fmt"
	"strings"

	"code.google.com/p/go.net/html"
	"code.google.com/p/go.net/html/atom"
)

// RemoveFirstHTMLAttribute removes an attribute from an HTML node and returns the result.
func RemoveFirstHTMLAttribute(node *html.Node, attr string) *html.Node {
	var attributes []html.Attribute
	for _, a := range node.Attr {
		if a.Key != attr {
			attributes = append(attributes, a)
		}
	}

	node.Attr = attributes
	return node
}

// Render is a helper that renders html.Node as a string.
func Render(h *html.Node) string {
	b := new(bytes.Buffer)
	html.Render(b, h)
	return b.String()
}

// ParseFragment parses a string and returns an html node with no <html><head></head><body>...</body></html> overhead.
func ParseFragment(s string) *html.Node {
	n, err := html.ParseFragment(strings.NewReader(s), &html.Node{
		Type:     html.ElementNode,
		Data:     "body",
		DataAtom: atom.Body,
	})
	if err != nil {
		return nil
	}
	return n[0]
}

// RemoveAttribute removes an attribute from an HTML string and returns the result.
func RemoveAttribute(s, attr string) string {
	n := ParseFragment(s)
	if n == nil {
		return s
	}
	return Render(RemoveFirstHTMLAttribute(n, attr))
}

func main() {
	const str = "<table style='small' id='this is my id'><tbody><tr><td id='123'>Text</td></tr></tbody></table>"

	fmt.Println(RemoveAttribute(str, "id"))
	fmt.Println(RemoveAttribute(str, "style"))
}
	package main

	import (
	"bytes"
	"fmt"
	"strings"

	"code.google.com/p/go.net/html"
	"code.google.com/p/go.net/html/atom"
	)

	// RemoveFirstHTMLAttribute removes an attribute from an HTML node and returns the result.
	func RemoveFirstHTMLAttribute(node html.Node, attr string) html.Node {
	var attributes []html.Attribute
	for _, a := range node.Attr {
	if a.Key != attr {
	attributes = append(attributes, a)
	}
	}

	node.Attr = attributes
	return node
	}

	// Render is a helper that renders html.Node as a string.
	func Render(h *html.Node) string {
	b := new(bytes.Buffer)
	html.Render(b, h)
	return b.String()
	}

	// ParseFragment parses a string and returns an html node with no <html><head></head><body>...</body></html> overhead.
	func ParseFragment(s string) *html.Node {
	n, err := html.ParseFragment(strings.NewReader(s), &html.Node{
	Type: html.ElementNode,
	Data: "body",
	DataAtom: atom.Body,
	})
	if err != nil {
	return nil
	}
	return n[0]
	}

	// RemoveAttribute removes an attribute from an HTML string and returns the result.
	func RemoveAttribute(s, attr string) string {
	n := ParseFragment(s)
	if n == nil {
	return s
	}
	return Render(RemoveFirstHTMLAttribute(n, attr))
	}

	func main() {
	const str = "<table style='small' id='this is my id'><tbody><tr><td id='123'>Text</td></tr></tbody></table>"

	fmt.Println(RemoveAttribute(str, "id"))
	fmt.Println(RemoveAttribute(str, "style"))
	}