// Add street to result and find street information (to WikipediaMoscow.result)
func (parser *WikipediaMoscow) processLink(_ int, s *goquery.Selection, done chan<- *StreetInfo) {
name := strings.TrimSpace(s.Text())
if len(name) == 0 {
done <- parser.getEmptyInfo("")
return
}
href, exists := s.Attr("href")
if !exists {
done <- parser.getEmptyInfo("")
return
}
var (
class string
info *StreetInfo
)
class, exists = s.Attr("class")
if exists && class == "new" {
info = parser.getEmptyInfo(name)
} else if resp, err := http.Get(parser.baseURL + href); err != nil {
info = parser.getEmptyInfo(name)
} else {
streetparser := NewWikipediaStreetParser()
info, err = streetparser.ParseStreetInfo(name, resp.Body)
if err != nil {
info = parser.getEmptyInfo(name)
}
}
done <- info
}
// node returns a string representation of the selection.
func node(i int, s *goquery.Selection) string {
switch node := s.Get(0); {
case node.Data == "h1":
return fmt.Sprintf(" \033[%dm# %s\033[0m\n\n", blue, text(s))
case node.Data == "h2":
return fmt.Sprintf(" \033[%dm## %s\033[0m\n\n", blue, text(s))
case node.Data == "h3":
return fmt.Sprintf(" \033[%dm### %s\033[0m\n\n", blue, text(s))
case node.Data == "p":
return fmt.Sprintf("\033[%dm%s\033[0m\n\n", none, indent(text(s), 1))
case node.Data == "pre" || s.HasClass("highlight"):
return fmt.Sprintf("\033[1m%s\033[0m\n\n", indent(text(s), 2))
case node.Data == "a":
return fmt.Sprintf("%s (%s) ", s.Text(), s.AttrOr("href", "missing link"))
case node.Data == "li":
return fmt.Sprintf(" • %s\n", contents(s))
case node.Data == "ul":
return fmt.Sprintf("%s\n", nodes(s))
case node.Data == "code":
return fmt.Sprintf("\033[1m%s\033[0m ", s.Text())
case node.Type == html.TextNode:
return strings.TrimSpace(node.Data)
default:
return ""
}
}
func ScrapeExamples(s *goquery.Selection) []string {
examples := []string{}
s.Find("span.h").Each(func(i int, s *goquery.Selection) {
examples = append(examples, s.Text())
})
return examples
}
func JoinNodesWithSpace(s *goquery.Selection) string {
texts := []string{}
s.Each(func(i int, s *goquery.Selection) {
texts = append(texts, s.Text())
})
return strings.Join(texts, " ")
}
func parseColors(s *goquery.Selection) string {
colors := ""
s.Each(func(i int, s *goquery.Selection) {
colors += s.Text()
})
return colors
}
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
}
func (rc *TwitterChecker) findSigInTweet(h SigHint, s *goquery.Selection) ProofError {
inside := s.Text()
html, err := s.Html()
checkText := h.checkText
if err != nil {
return NewProofError(keybase1.ProofStatus_CONTENT_FAILURE, "No HTML tweet found: %s", err)
}
G.Log.Debug("+ Checking tweet '%s' for signature '%s'", inside, checkText)
G.Log.Debug("| HTML is: %s", html)
rxx := regexp.MustCompile(`^(@[a-zA-Z0-9_-]+\s+)`)
for {
if m := rxx.FindStringSubmatchIndex(inside); m == nil {
break
} else {
prefix := inside[m[2]:m[3]]
inside = inside[m[3]:]
G.Log.Debug("| Stripping off @prefx: %s", prefix)
}
}
if strings.HasPrefix(inside, checkText) {
return nil
}
return NewProofError(keybase1.ProofStatus_DELETED, "Could not find '%s' in '%s'",
checkText, inside)
}
//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
}
func getDataFromDOM(s *gq.Selection, arr []string, code string) string {
var dt string
if arr[0] == "text" {
dt = s.Text()
} else {
dt, _ = s.Attr(arr[0])
}
return encode_string(dt, code)
}
func (e extractContent) noParasWithoutTable(s *goquery.Selection) bool {
s.FindMatcher(pTags).Each(func(i int, s *goquery.Selection) {
if len(s.Text()) < 25 {
s.Remove()
}
})
return s.FindMatcher(pTags).Length() == 0 && !nodeIs(s.Nodes[0], atom.Td)
}
func (d *Document) getLinkDensity(s *goquery.Selection) float32 {
linkLength := len(s.Find("a").Text())
textLength := len(s.Text())
if textLength == 0 {
return 0
}
return float32(linkLength) / float32(textLength)
}
func parseHeader(element *goquery.Selection, info *TrainInfo) {
element.Find("span").Each(func(i int, element *goquery.Selection) {
switch i {
case 0:
info.Category, info.Number, info.Name = parseTrainDenomination(element.Text())
case 2:
info.From, info.To = parseTrainRoute(element.Text())
}
})
}
func displayDetails(single *goquery.Selection) {
text := strings.TrimSpace(single.Text())
href, _ := single.Attr("href")
length := utf8.RuneCountInString(text)
if ((length > 5) && wordExists(text, "keywords")) || ((length > 5) && wordExists(href, "keywords")) {
if wordExists(text, "products") {
fmt.Println("Link", single.Text(), "--->", href)
}
}
}
func parseResource(s *goquery.Selection) (_production, _stored, _capacity int) {
productionStr, _ := s.Attr("title")
production, _ := strconv.Atoi(productionStr)
status := s.Text()
split := strings.Split(status, "/")
stored, _ := strconv.Atoi(split[0])
capacity, _ := strconv.Atoi(split[1])
return production, stored, capacity
}
func addIngredient(ingredients []data.Ingredient, a *goquery.Selection) []data.Ingredient {
if href, ok := a.Attr("href"); ok {
glog.V(2).Info(" href: " + href)
id, err := strconv.Atoi(strings.Split(href, "/")[2])
if err != nil {
glog.Errorf("Failed to extract id from %s: %v", href, err)
} else {
ingredients = append(ingredients, data.Ingredient{Name: a.Text(), Id: id})
}
}
return ingredients
}
func describeSentences(s *goquery.Selection) TextDescription {
var d TextDescription
var text string
// get text of this node and then split for sentences
if s.Children().Length() > 0 {
text = getTextFromHtml(s)
} else {
text = s.Text()
}
sentences := tokenizer.Tokenize(text)
d.CountSentences = len(sentences)
//fmt.Println("==============================================")
for _, s := range sentences {
sentence := s.Text
if len(sentence) == 0 {
continue
}
c := len(get_words_from(sentence))
//fmt.Println(sentence)
d.AverageWords += c
if c > 3 {
// presume normal sentence usually has more 3 words
d.CountLongSentences++
if c < 25 {
// but a sentence should not have nore 25 words. We will not
// consider such sentence as a good one
d.CountGoodSentences++
}
lastsymbol := sentence[len(sentence)-1:]
if strings.ContainsAny(lastsymbol, ".?!") {
d.CountCorrectSentences++
}
}
}
if d.CountSentences > 0 {
d.AverageWords = int(d.AverageWords / d.CountSentences)
}
return d
}
//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 IsAllCode(bodyContent *goquery.Selection) bool {
preDom := bodyContent.Find("pre")
if preDom != nil {
for index := 0; index < preDom.Length(); index++ {
dom := preDom.Eq(index)
dom.Remove()
}
}
conStr := bodyContent.Text()
conStr = strings.TrimSpace(conStr)
if len(conStr) < 100 {
return true
}
return false
}
func getText(s *goquery.Selection, includeDecendents bool) string {
if s.Length() == 0 {
return ""
}
if includeDecendents {
return strings.TrimSpace(s.Text())
}
var buff []string
for node := s.First().Nodes[0].FirstChild; node != nil; node = node.NextSibling {
if node.Type == html.TextNode {
buff = append(buff, node.Data)
}
}
return strings.TrimSpace(strings.Join(buff, ""))
}
func (s *StateMachine) ProcessSelection(sl *goquery.Selection) {
if strings.HasPrefix(sl.Text(), "備註") {
s.State = InFootnote
}
if strings.HasPrefix(sl.Text(), "資訊更新日期") {
s.State = NotInFootnote
}
if s.State == InFootnote {
if sl.Size() != 1 {
panic("element size is not 1")
}
s.ProcessNode(sl.Nodes[0])
}
}
// return the chapter number and the line number for the sentence
func getLineNumber(sel *goquery.Selection) (chapterNumber int, lineNumber int) {
rawtext := sel.Text()
textArray := strings.Split(rawtext, ":")
var err error = nil
chapterNumber, err = strconv.Atoi(textArray[0])
if err != nil {
log.Fatal(err)
}
lineNumber, err = strconv.Atoi(textArray[1])
if err != nil {
log.Fatal(err)
}
return
}
func addEntryType(typename string, s *goquery.Selection) {
linkLabel := s.Find(".memItemRight").Find("a")
linkLabel.Each(func(i int, s *goquery.Selection) {
structName := s.Text()
structName = strings.Trim(structName, "\r\n ")
link, _ := s.Attr("href")
link = strings.Trim(link, "\r\n ")
if len(structName) != 0 && len(link) != 0 {
_, err := db.Exec("insert or ignore into searchIndex(name,type,path) VALUES('" + structName + "','" + typename + "','" + link + "')")
if err != nil {
log.Fatal("Insert " + typename + " " + structName + "Failed!")
return
}
log.Print("Insert " + typename + structName)
}
})
}
func ScrapeFileLink(s *goquery.Selection) {
i := Images{Created_at: time.Now(), Updated_at: time.Now()}
i.Name = s.Text()
href, _ := s.Attr("href")
isImage := i.IsImageType()
if isImage {
i.Source = fmt.Sprintf("http://rghost.ru%s/image.png", href)
downloaded := i.DownloadImage()
if downloaded {
i.Uploaded_to = "yes"
i.Archived = false
err := i.InsertImage()
if err != nil {
log.Fatal(err)
}
}
}
}
func processUl(ul *goquery.Selection, depth int) {
ul.Find("li").Each(func(_ int, li *goquery.Selection) {
li.Find("ul").Each(func(_ int, childUl *goquery.Selection) {
processUl(childUl, depth+1)
})
lines := StringToLines(li.Text())
var indentedLines []string
for i, line := range lines {
if i == 0 {
liMarkIndex := depth % 2
mark := liMark[liMarkIndex]
indentedLines = append(indentedLines, "\n"+mark+" "+line)
} else {
indentedLines = append(indentedLines, " "+line)
}
}
li.ReplaceWithHtml(strings.Join(indentedLines, "\n"))
})
ul.ReplaceWithHtml(ul.Text())
}
func (e extractContent) getSiblingContent(
a *Article,
s *goquery.Selection,
baseScore uint) []*html.Node {
var ret []*html.Node
if nodeIs(s.Nodes[0], atom.P) && len(s.Text()) > 0 {
return s.Nodes
}
ps := s.FindMatcher(pTags)
for _, n := range ps.Nodes {
cc := a.getCCache(n)
if len(cc.text) > 0 {
if cc.stopwords > baseScore && !cc.highLinkDensity {
ret = append(ret, createNode(atom.P, "p", cc.text))
}
}
}
return ret
}
func printSelectionTextWithTitle(title string, sel *goquery.Selection) {
Debug("%v selection: %v", title, sel.Text())
}
func printSelectionText(sel *goquery.Selection) {
Debug("selection: %v", sel.Text())
}
func UpdatePageSummary(docName, PageFilePath, PreviewFilePath string) error {
var pageSummaryData PageSummaryData
html, err := readPageAsHtml(docName, PageFilePath)
if err != nil {
return err
}
htmlreader := bytes.NewReader(html)
//htmlString := string(html)
//log.Info("HtmlString " + htmlString)
doc, err := gq.NewDocumentFromReader(htmlreader)
if err != nil {
return err
}
var SelectedNodes *gq.Selection
//======= work out the document heading ========
DocTitle := ""
if DocTitle == "" {
SelectedNodes = doc.Find("h1, h2, h3, h4").First()
if len(SelectedNodes.Nodes) == 1 {
DocTitle = strings.TrimSpace(SelectedNodes.Text())
}
}
if DocTitle == "" {
DocTitle = docName
}
//DocTitle = base64.StdEncoding.EncodeToString([]byte(DocTitle))
DocTitle = malkovich.FolderNameToDocName(DocTitle)
pageSummaryData.PageTitle = DocTitle
//======== look for an image =========
DocImage := ""
SelectedNodes = doc.Find("img").First()
if len(SelectedNodes.Nodes) == 1 {
for _, nodeAttr := range SelectedNodes.Nodes[0].Attr {
if nodeAttr.Key == "src" {
DocImage = nodeAttr.Val
break
}
}
}
//DocImage = base64.StdEncoding.EncodeToString([]byte(DocImage))
pageSummaryData.FirstImage = DocImage
//======== look for the first paragraph =========
FirstParagraph := ""
SelectedNodes = doc.Find("p").First()
if len(SelectedNodes.Nodes) == 1 {
FirstParagraph = strings.TrimSpace(SelectedNodes.Text())
}
//TODO:HIGH Maybe limit to a set number of charactors here.
//FirstParagraph = base64.StdEncoding.EncodeToString([]byte(FirstParagraph))
pageSummaryData.FirstParagraph = FirstParagraph
jsonData, err := json.Marshal(pageSummaryData)
if err != nil {
return err
}
// TODO:MED what would the best file permissions be here?
err = ioutil.WriteFile(PreviewFilePath, jsonData, os.FileMode(0644))
if err != nil {
panic(err.Error())
}
return nil
}
// return the sentence pure content without comment number
func getPureContent(sel *goquery.Selection) string {
pureText := sel.Text()
return pureText
}
func (ve *VideoExtractor) getEmbedCode(node *goquery.Selection) string {
return node.Text()
}
