// 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
}
func htmlToText(r io.Reader) []byte {
t := html.NewTokenizer(r)
var out bytes.Buffer
var ignorescore int
for {
switch token := t.Next(); token {
case html.StartTagToken:
if _, ok := ignoretag[string(t.Token().Data)]; ok {
ignorescore++
}
case html.EndTagToken:
if _, ok := ignoretag[string(t.Token().Data)]; ok {
ignorescore--
}
case html.ErrorToken:
return out.Bytes()
case html.CommentToken:
continue
case html.TextToken:
if ignorescore == 0 {
html := strings.TrimSpace(t.Token().Data)
if len(html) > 0 {
fmt.Fprintln(&out, html)
}
}
}
}
}
// rewriteHTML scans the HTML for tags with url-valued attributes, and updates
// those values with the urlRewriter function. The updated HTML is output to the
// writer.
func rewriteHTML(reader io.Reader, writer io.Writer, urlRewriter func(string) string) error {
// Note: This assumes the content is UTF-8.
tokenizer := html.NewTokenizer(reader)
var err error
for err == nil {
tokenType := tokenizer.Next()
switch tokenType {
case html.ErrorToken:
err = tokenizer.Err()
case html.StartTagToken, html.SelfClosingTagToken:
token := tokenizer.Token()
if urlAttrs, ok := atomsToAttrs[token.DataAtom]; ok {
for i, attr := range token.Attr {
if urlAttrs.Has(attr.Key) {
token.Attr[i].Val = urlRewriter(attr.Val)
}
}
}
_, err = writer.Write([]byte(token.String()))
default:
_, err = writer.Write(tokenizer.Raw())
}
}
if err != io.EOF {
return err
}
return nil
}
func isHTML(content []byte) bool {
isHTML := false
if len(content) == 0 {
return isHTML
}
if len(content) > 1024 {
content = content[:1024]
}
z := html.NewTokenizer(bytes.NewReader(content))
isFinish := false
for !isFinish {
switch z.Next() {
case html.ErrorToken:
isFinish = true
case html.StartTagToken:
tagName, _ := z.TagName()
if bytes.Equal(tagName, []byte("html")) {
isHTML = true
isFinish = true
}
}
}
return isHTML
}
func scrape(r io.Reader) {
z := html.NewTokenizer(r)
buf := &bytes.Buffer{}
L:
for {
tt := z.Next()
tok := z.Token()
switch tt {
case html.StartTagToken:
// if you find a link, replace it with our stylesheet
if tok.DataAtom == atom.Tr {
// check for correct class attr and then switch to
// html.NewTokenizerFragment
}
break
case html.EndTagToken:
// once you reach the end of the head, flush everything left in
// the tokenizer to the buffer
if tok.String() == "</head>" {
buf.Write(z.Buffered())
break L
}
case html.ErrorToken:
// this is left in here for things like tracking pixels that have
// the HTML content type, so our code doesn't break
break L
}
}
}
func Autodiscover(b []byte) (string, error) {
r := bytes.NewReader(b)
z := html.NewTokenizer(r)
for {
if z.Next() == html.ErrorToken {
if err := z.Err(); err == io.EOF {
break
} else {
return "", ErrNoRssLink
}
}
t := z.Token()
switch t.DataAtom {
case atom.Link:
if t.Type == html.StartTagToken || t.Type == html.SelfClosingTagToken {
attrs := make(map[string]string)
for _, a := range t.Attr {
attrs[a.Key] = a.Val
}
if attrs["rel"] == "alternate" && attrs["href"] != "" &&
(attrs["type"] == "application/rss+xml" || attrs["type"] == "application/atom+xml") {
return attrs["href"], nil
}
}
}
}
return "", ErrNoRssLink
}
func ParseHtml(r io.Reader, url string) ([]byte, error) {
z := html.NewTokenizer(r)
var newHtml []byte
lastTag := ""
for {
tt := z.Next()
rawHtmlBytes := z.Raw()
switch tt {
case html.ErrorToken:
e := z.Err()
if e.Error() == "EOF" {
return newHtml, nil
} else {
return make([]byte, 0), z.Err()
}
case html.TextToken:
rawHtml := strings.TrimSpace(string(rawHtmlBytes[:]))
if len(rawHtml) > 0 && lastTag == "style" {
newCss := ParseCss(rawHtml, url)
newHtml = append(newHtml, []byte(newCss)...)
} else {
newHtml = append(newHtml, rawHtmlBytes...)
}
case html.DoctypeToken, html.CommentToken, html.EndTagToken:
newHtml = append(newHtml, rawHtmlBytes...)
case html.StartTagToken:
lastTag = flushTagToken(&newHtml, z, url)
case html.SelfClosingTagToken:
flushTagToken(&newHtml, z, url)
}
if tt != html.StartTagToken {
lastTag = ""
}
}
}
func GetPriceForBestBuy(url string) float64 {
resp, err := http.Get(url)
if err != nil {
log.Fatal(err)
}
defer resp.Body.Close()
z := html.NewTokenizer(resp.Body)
for {
tt := z.Next()
switch {
case tt == html.ErrorToken:
return 0.0
case tt == html.StartTagToken:
t := z.Token()
isSpan := t.Data == "meta"
if isSpan {
for _, attr := range t.Attr {
if attr.Key == "id" && strings.Contains(attr.Val, "schemaorg-offer-price") {
nxt := z.Next()
if nxt == html.TextToken {
t = z.Token()
return parseCurrency(t.Data)
}
}
}
}
}
}
}
func getLinks(u *url.URL) []*url.URL {
resp, err := http.Get(u.String())
if err != nil {
logs.Log(fmt.Sprintf("Couldn't crawl %s", u))
}
defer resp.Body.Close()
links := make([]*url.URL, 0)
tokenizer := html.NewTokenizer(resp.Body)
for {
tokenType := tokenizer.Next()
switch tokenType {
case html.ErrorToken:
return links
case html.StartTagToken, html.SelfClosingTagToken:
token := tokenizer.Token()
if link, ok := getURL(u, token); ok {
links = append(links, link)
}
}
}
return links
}
func parseTitle(resp io.Reader, fallback string) string {
r := io.LimitedReader{
R: resp,
N: 8192,
}
h := html.NewTokenizer(&r)
for {
tt := h.Next()
switch tt {
case html.ErrorToken:
return fallback
case html.StartTagToken:
tag, _ := h.TagName()
if string(tag) == "title" {
nt := h.Next()
switch nt {
case html.ErrorToken:
return "Failed to parse title"
case html.TextToken:
return h.Token().Data
}
}
}
}
return fallback
}
func (w *WebClient) obtainCsrf(b io.Reader) error {
var errorMessage error = nil
z := html.NewTokenizer(b)
for {
tt := z.Next()
switch {
case tt == html.ErrorToken:
return errorMessage
case tt == html.SelfClosingTagToken:
t := z.Token()
isMeta := t.Data == "meta"
if isMeta && len(t.Attr) > 0 {
if (t.Attr[1].Key == "name") && (t.Attr[1].Val == "csrf-token") {
w.csrf = t.Attr[0].Val
log.Debugf("Csrf Token: %s", w.csrf)
} else if (t.Attr[0].Key == "name") && (t.Attr[0].Val == "csrf-token") {
w.csrf = t.Attr[1].Val
log.Debugf("Csrf Token: %s", w.csrf)
}
}
case tt == html.StartTagToken:
t := z.Token()
if (t.Data == "div") && len(t.Attr) > 0 && (t.Attr[0].Key == "id") && (t.Attr[0].Val == "flash_alert") {
z.Next()
errorMessage = errors.New(z.Token().String())
}
}
}
}
// crawl the page
func Crawl(url string, ch chan string) {
resp, _ := http.Get(url_prefix + url)
tokenizer := html.NewTokenizer(resp.Body)
defer resp.Body.Close()
for {
token := tokenizer.Next()
switch {
case token == html.ErrorToken:
// End of page
ch <- "END!"
return
case token == html.StartTagToken:
start_tt := tokenizer.Token()
if start_tt.Data == "div" {
//fmt.Println("get a div! %v", num)
if isSummary(start_tt) {
getQ(*tokenizer, ch)
}
} else {
continue
}
}
}
}
/*
* avanza_get_sellprice
*
* Site: Avanza
* Gets the current sellprice from a given httpResponse
*/
func (this *Parse) avanza_get_sellprice(resp *http.Response) float64 {
z := html.NewTokenizer(resp.Body)
for {
tt := z.Next()
switch {
case tt == html.ErrorToken:
return 0.0
case tt == html.StartTagToken:
t := z.Token()
if isCatch := t.Data == "span"; isCatch {
for _, attr := range t.Attr {
if strings.Contains(attr.Val, "sellPrice") {
z.Next()
tt := z.Token()
strval := strings.Replace(tt.String(), ",", ".", -1)
value, _ := strconv.ParseFloat(strval, 64)
return value
}
}
}
}
}
}
func GetPriceForWalmart(url string) float64 {
resp, err := http.Get(url)
if err != nil {
log.Fatal(err)
}
defer resp.Body.Close()
z := html.NewTokenizer(resp.Body)
for {
tt := z.Next()
switch {
case tt == html.ErrorToken:
return 0.0
case tt == html.StartTagToken:
t := z.Token()
isSpan := t.Data == "script"
if isSpan {
for _, attr := range t.Attr {
if attr.Key == "id" && strings.Contains(attr.Val, "tb-djs-wml-base") {
nxt := z.Next()
if nxt == html.TextToken {
return parseJson(z.Token().Data)
}
}
}
}
}
}
}
// scrapePageWorker -- this is the function that does most of the work in parsing the HTML
func scrapePageWorker(page *io.ReadCloser, out chan [2]string, chFinished chan bool) {
defer func() {
chFinished <- true
}()
z := html.NewTokenizer(*page)
// infinite loop to toss state tokens into a url map
for {
var result [2]string
tt := z.Next()
switch {
case tt == html.ErrorToken:
return
case tt == html.StartTagToken:
t := z.Token()
isAnchor := t.Data == "a"
if !isAnchor {
continue
}
if isAnchor {
for _, attr := range t.Attr {
if attr.Key == "id" {
result[0] = attr.Val
}
if attr.Key == "data-href" {
result[1] = attr.Val
out <- result
}
}
}
}
} // end for
}
func getHTMLContent(r io.Reader, tag []byte) (result string, err error) {
z := html.NewTokenizer(r)
result = ""
valid := 0
cacheLen := len(tag)
for {
tt := z.Next()
switch tt {
case html.ErrorToken:
err = z.Err()
return
case html.TextToken:
if valid == 1 {
return string(z.Text()), nil
}
case html.StartTagToken, html.EndTagToken:
tn, _ := z.TagName()
if len(tn) == cacheLen && bytes.Equal(tn[0:cacheLen], tag) {
if tt == html.StartTagToken {
valid = 1
} else {
valid = 0
}
}
}
}
}
// Find all children links on a page and the title of the page from an HTTP response
func (w Web) findAllLinks(httpBody io.Reader, baseURL string) (links []string, title string) {
page := html.NewTokenizer(httpBody)
for {
tokenType := page.Next()
// End of the page, we are done
if tokenType == html.ErrorToken {
return
}
token := page.Token()
// Extract the page title
// React uses <title> tags also, but they have got special attributes
if tokenType == html.StartTagToken && token.DataAtom.String() == "title" && len(token.Attr) == 0 {
page.Next()
title = page.Token().Data
}
// Parse a link
if tokenType == html.StartTagToken && token.DataAtom.String() == "a" {
href, hasLink := w.extractLink(token)
if hasLink && w.ShouldCrawl(baseURL, href) {
links = append(links, w.unifyURL(href, baseURL))
}
}
}
}
func (a *HttpWebotsArchive) load() error {
resp, err := http.Get(a.baseurl)
if err != nil {
return err
}
tokenizer := html.NewTokenizer(resp.Body)
nameRx := regexp.MustCompile(fmt.Sprintf(`^webots-(.*)-%s.tar.bz2$`, a.arch))
for {
t := tokenizer.Next()
if t == html.ErrorToken {
err := tokenizer.Err()
if err == io.EOF {
break
}
return err
}
if t != html.StartTagToken {
continue
}
tName, hasAttrib := tokenizer.TagName()
if string(tName) != "a" {
continue
}
if hasAttrib == false {
continue
}
stopped := false
for stopped == false {
key, val, next := tokenizer.TagAttr()
if string(key) != "href" {
continue
}
stopped = !next
// we got a link, test if it has the right prefix
matches := nameRx.FindStringSubmatch(string(val))
if matches == nil {
continue
}
v, err := ParseWebotsVersion(matches[1])
if err != nil {
return err
}
a.versions = append(a.versions, v)
}
}
sort.Sort(&a.versions)
return nil
}
// Returns the href attribute of a <link rel="shortcut icon"> tag or error if not found.
func FindIcon(b []byte) (string, error) {
r := bytes.NewReader(b)
z := html.NewTokenizer(r)
for {
if z.Next() == html.ErrorToken {
if err := z.Err(); err == io.EOF {
break
} else {
return "", ErrNoIcon
}
}
t := z.Token()
switch t.DataAtom {
case atom.Link:
if t.Type == html.StartTagToken || t.Type == html.SelfClosingTagToken {
attrs := make(map[string]string)
for _, a := range t.Attr {
attrs[a.Key] = a.Val
}
if attrs["rel"] == "shortcut icon" && attrs["href"] != "" {
return attrs["href"], nil
}
}
}
}
return "", ErrNoIcon
}
func extractLinkUrls(page string) []string {
z := html.NewTokenizer(strings.NewReader(page))
hrefs := make([]string, 10)
for {
tt := z.Next()
switch {
case tt == html.ErrorToken:
// End of the document, we're done
return hrefs
case tt == html.StartTagToken:
t := z.Token()
isAnchor := t.Data == "a"
if isAnchor {
// we found a link
attributes := t.Attr
for _, attr := range attributes {
if attr.Key == "href" {
href := attr.Val
hrefs = append(hrefs, href)
}
}
}
}
}
}
// Preprocess escapes disallowed tags in a cleaner way, but does not fix
// nesting problems. Use with Clean.
func Preprocess(config *Config, fragment string) string {
if config == nil {
config = DefaultConfig
}
var buf bytes.Buffer
write := func(raw string) {
_, err := buf.WriteString(raw)
// The only possible error is running out of memory.
expectError(err, nil)
}
t := html.NewTokenizer(strings.NewReader(fragment))
for {
switch tok := t.Next(); tok {
case html.ErrorToken:
err := t.Err()
// The only possible errors are from the Reader or from
// the buffer capacity being exceeded. Neither can
// happen with strings.NewReader as the string must
// already fit into memory.
expectError(err, io.EOF)
if err == io.EOF {
write(html.EscapeString(string(t.Raw())))
return buf.String()
}
case html.TextToken:
write(string(t.Raw()))
case html.StartTagToken, html.EndTagToken, html.SelfClosingTagToken:
raw := string(t.Raw())
tagName, _ := t.TagName()
allowed := false
if tag := atom.Lookup(tagName); tag != 0 {
if _, ok := config.elem[tag]; ok {
allowed = true
}
}
if !allowed {
if _, ok := config.elemCustom[string(tagName)]; ok {
allowed = true
}
}
if !allowed {
raw = html.EscapeString(raw)
}
write(raw)
case html.CommentToken:
raw := string(t.Raw())
if config.EscapeComments || !strings.HasPrefix(raw, "<!--") || !strings.HasSuffix(raw, "-->") {
raw = html.EscapeString(raw)
}
write(raw)
default:
write(html.EscapeString(string(t.Raw())))
}
}
}
// ScrapeArticle fetches and parses the article.
// article should be provided as a *Article.
func ScrapeArticle(article Article) error {
cookies := NewCookieJar()
client := &http.Client{Jar: cookies}
// build request
req, err := http.NewRequest("GET", article.GetLink(), nil) //create http request
err = buildArticleHeader(req)
if err != nil {
log.Error("could not build article request:", err)
return err
}
//send http request
resp, err := client.Do(req)
if err != nil {
log.Error("error sending article request:", err)
return err
}
defer resp.Body.Close()
// TODO: check resp.Header to see if X-Article-Template is [full]
// parse request
parser := html.NewTokenizer(resp.Body)
err = article.DoParse(parser) //parse the html body
if err != nil {
log.Error("error building article request:", err)
return err
}
return nil
}
/**
* Download a given url and return all urls found on that page.
*/
func retrieveLinks(url string) []string {
resp, err := http.Get(url)
links := make([]string, 0)
if err != nil {
logger.Error("Detected broken url", url)
return links
}
defer resp.Body.Close()
page := html.NewTokenizer(resp.Body)
for {
tokenType := page.Next()
if tokenType == html.ErrorToken {
return links
}
token := page.Token()
if tokenType == html.StartTagToken && token.DataAtom.String() == "a" {
for _, attr := range token.Attr {
if attr.Key == "href" {
links = append(links, attr.Val)
}
}
}
}
}
// GetTitle gets the title token of a HTML page
func GetTitle(resp *http.Response, url string) string {
fURL := resp.Request.URL.String()
z := html.NewTokenizer(resp.Body)
for {
tt := z.Next()
switch tt {
case html.ErrorToken:
return ""
case html.StartTagToken:
t := z.Token()
if t.Data == "title" {
tt = z.Next()
t = z.Token()
d := t.Data
if len(d) > 450 {
d = d[:450]
}
d = strings.TrimSpace(strings.Replace(d, "\n", " ", -1))
if fURL != url {
return fmt.Sprintf("%v (%v)", d, fURL)
}
return d
}
}
}
}
func ParseHtml(r io.Reader) map[string]RefType {
refs := make(map[string]RefType)
tokenizer := html.NewTokenizer(r)
for {
tt := tokenizer.Next()
switch {
case tt == html.ErrorToken:
// tokenizer telling us its reached the end
return refs
case tt == html.StartTagToken:
token := tokenizer.Token()
element, ok := knownElements[token.Data]
if !ok {
continue
}
for _, attr := range token.Attr {
refType, ok := element[attr.Key]
if !ok {
continue
}
refs[attr.Val] = refType
break
}
}
}
return refs
}
func (p *PickerAttr) Pick(r io.Reader) (data []string, err error) {
z := html.NewTokenizer(r)
for {
tt := z.Next()
switch tt {
case html.ErrorToken:
if z.Err() == io.EOF {
return data, nil
}
case html.StartTagToken:
tag_name, attr := z.TagName()
if string(tag_name) != p.TagName {
continue
}
var key, value []byte
for attr {
key, value, attr = z.TagAttr()
if string(key) == p.Attr {
data = append(data, string(value))
}
}
}
}
return data, z.Err()
}
func (parser *HtmlContentParser) Parse(c *MemoryContent, in io.Reader) error {
z := html.NewTokenizer(in)
for {
tt := z.Next()
switch {
case tt == html.ErrorToken:
if z.Err() == io.EOF {
return nil
}
return z.Err()
case tt == html.StartTagToken:
tag, _ := z.TagName()
switch string(tag) {
case "head":
if err := parser.parseHead(z, c); err != nil {
return err
}
case "body":
if err := parser.parseBody(z, c); err != nil {
return err
}
}
}
}
}
// Scrape scrapes a web page and populates the course struct.
func Scrape(r io.Reader) *Course {
z := html.NewTokenizer(r)
buf := &bytes.Buffer{}
L:
for {
tt := z.Next()
tok := z.Token()
//if tok.DataAtom != atom.Link {
// for anything that isn't a link, just write the raw data to the
// buffer
buf.Write(z.Raw())
//}
switch tt {
case html.StartTagToken:
// if you find a link, replace it with our stylesheet
if tok.DataAtom == atom.Link {
// this isn't the correct stylesheet, just a placeholder
//buf.Write([]byte(`<link rel="stylesheet" href="/cxo.css">`))
}
break
case html.EndTagToken:
// once you reach the end of the head, flush everything left in
// the tokenizer to the buffer
if tok.String() == "</head>" {
buf.Write(z.Buffered())
break L
}
case html.ErrorToken:
// this is left in here for things like tracking pixels that have
// the HTML content type, so our code doesn't break
break L
}
}
}
// Search for
// <head>
// <meta http-equiv="X-XRDS-Location" content="....">
func findMetaXrdsLocation(input io.Reader) (location string, err error) {
tokenizer := html.NewTokenizer(input)
inHead := false
for {
tt := tokenizer.Next()
switch tt {
case html.ErrorToken:
return "", tokenizer.Err()
case html.StartTagToken, html.EndTagToken:
tk := tokenizer.Token()
if tk.Data == "head" {
if tt == html.StartTagToken {
inHead = true
} else {
return "", errors.New("Meta X-XRDS-Location not found")
}
} else if inHead && tk.Data == "meta" {
ok := false
content := ""
for _, attr := range tk.Attr {
if attr.Key == "http-equiv" &&
strings.ToLower(attr.Val) == "x-xrds-location" {
ok = true
} else if attr.Key == "content" {
content = attr.Val
}
}
if ok && len(content) > 0 {
return content, nil
}
}
}
}
return "", errors.New("Meta X-XRDS-Location not found")
}
func fetchHyperLink(httpBody io.Reader) []string {
defer ioutil.ReadAll(httpBody)
links := make([]string, 0)
body := html.NewTokenizer(httpBody)
for {
tokenType := body.Next()
if tokenType == html.ErrorToken {
return links
}
token := body.Token()
if tokenType == html.StartTagToken {
if token.DataAtom.String() == "a" || token.DataAtom.String() == "link" {
for _, attribute := range token.Attr {
if attribute.Key == "href" {
links = append(links, attribute.Val)
}
}
} else if token.DataAtom.String() == "img" || token.DataAtom.String() == "script" {
for _, attribute := range token.Attr {
if attribute.Key == "src" {
links = append(links, attribute.Val)
}
}
}
}
}
return links
}
