func (this *contentExtractor) getSiblingsContent(currentSibling *goquery.Selection, baselinescoreSiblingsPara float64) []*goquery.Selection {
ps := make([]*goquery.Selection, 0)
if currentSibling.Get(0).DataAtom.String() == "p" && len(currentSibling.Text()) > 0 {
ps = append(ps, currentSibling)
return ps
} else {
potentialParagraphs := currentSibling.Find("p")
potentialParagraphs.Each(func(i int, s *goquery.Selection) {
text := s.Text()
if len(text) > 0 {
ws := this.config.stopWords.stopWordsCount(this.config.targetLanguage, text)
paragraphScore := ws.stopWordCount
siblingBaselineScore := 0.30
highLinkDensity := this.isHighLinkDensity(s)
score := siblingBaselineScore * baselinescoreSiblingsPara
if score < float64(paragraphScore) && !highLinkDensity {
node := new(html.Node)
node.Type = html.TextNode
node.Data = text
node.DataAtom = atom.P
nodes := make([]*html.Node, 1)
nodes[0] = node
newSelection := new(goquery.Selection)
newSelection.Nodes = nodes
ps = append(ps, newSelection)
}
}
})
}
return ps
}
//a lot of times the first paragraph might be the caption under an image so we'll want to make sure if we're going to
//boost a parent node that it should be connected to other paragraphs, at least for the first n paragraphs
//so we'll want to make sure that the next sibling is a paragraph and has at least some substatial weight to it
func (this *contentExtractor) isBoostable(node *goquery.Selection) bool {
stepsAway := 0
next := node.Next()
for next != nil && stepsAway < node.Siblings().Length() {
currentNodeTag := node.Get(0).DataAtom.String()
if currentNodeTag == "p" {
if stepsAway >= 3 {
if this.config.debug {
log.Println("Next paragraph is too far away, not boosting")
}
return false
}
paraText := node.Text()
ws := this.config.stopWords.stopWordsCount(this.config.targetLanguage, paraText)
if ws.stopWordCount > 5 {
if this.config.debug {
log.Println("We're gonna boost this node, seems content")
}
return true
}
}
stepsAway++
next = next.Next()
}
return false
}
//checks the density of links within a node, is there not much text and most of it contains bad links?
//if so it's no good
func (this *contentExtractor) isHighLinkDensity(node *goquery.Selection) bool {
links := node.Find("a")
if links == nil || links.Size() == 0 {
return false
}
text := node.Text()
words := strings.Split(text, " ")
nwords := len(words)
sb := make([]string, 0)
links.Each(func(i int, s *goquery.Selection) {
linkText := s.Text()
sb = append(sb, linkText)
})
linkText := strings.Join(sb, "")
linkWords := strings.Split(linkText, " ")
nlinkWords := len(linkWords)
nlinks := links.Size()
linkDivisor := float64(nlinkWords) / float64(nwords)
score := linkDivisor * float64(nlinks)
if this.config.debug {
logText := ""
if len(node.Text()) >= 51 {
logText = node.Text()[0:50]
} else {
logText = node.Text()
}
log.Printf("Calculated link density score as %1.5f for node %s\n", score, logText)
}
if score > 1.0 {
return true
}
return false
}
func (ve *VideoExtractor) getEmbedCode(node *goquery.Selection) string {
return node.Text()
}