// ProcessHTML parses given html from Reader interface and fills up OpenGraph structure
func (og *OpenGraph) ProcessHTML(buffer io.Reader) error {
z := html.NewTokenizer(buffer)
for {
tt := z.Next()
switch tt {
case html.ErrorToken:
if z.Err() == io.EOF {
return nil
}
return z.Err()
case html.StartTagToken, html.SelfClosingTagToken, html.EndTagToken:
name, hasAttr := z.TagName()
if atom.Lookup(name) == atom.Body {
return nil // OpenGraph is only in head, so we don't need body
}
if atom.Lookup(name) != atom.Meta || !hasAttr {
continue
}
m := make(map[string]string)
var key, val []byte
for hasAttr {
key, val, hasAttr = z.TagAttr()
m[atom.String(key)] = string(val)
}
og.ProcessMeta(m)
}
}
return nil
}
// Token returns the next Token. The result's Data and Attr values remain valid
// after subsequent Next calls.
func (z *Tokenizer) Token() Token {
t := Token{Type: z.tt}
switch z.tt {
case TextToken, CommentToken, DoctypeToken:
t.Data = string(z.Text())
case StartTagToken, SelfClosingTagToken, EndTagToken:
name, moreAttr := z.TagName()
for moreAttr {
var key, val []byte
key, val, moreAttr = z.TagAttr()
t.Attr = append(t.Attr, Attribute{"", atom.String(key), string(val)})
}
if a := atom.Lookup(name); a != 0 {
t.DataAtom, t.Data = a, a.String()
} else {
t.DataAtom, t.Data = 0, string(name)
}
}
return t
}