merch-tracker-v2/task-processor
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

extended price search

Browse source
This commit is contained in:
nquidox 2025-10-13 20:54:12 +03:00
parent 392400bd44
commit 4b53a1d395
1 changed files with 30 additions and 17 deletions
Download patch file Download diff file Expand all files Collapse all files

25
internal/parsers/mandarake.go
View file

@ -235,16 +235,32 @@ func (s *MandarakeParser) findNode(doc *html.Node, params price) *html.Node {
} }
func (s *MandarakeParser) findData(doc *html.Node, params price) []string { func (s *MandarakeParser) findData(doc *html.Node, params price) []string {
if doc != nil { if doc == nil {
return nil
}
var ( var (
crawler func(*html.Node) crawler func(*html.Node)
values []string values []string
getText func(*html.Node) string
) )
getText = func(n *html.Node) string {
if n.Type == html.TextNode {
return n.Data
}
var result strings.Builder
for c := n.FirstChild; c != nil; c = c.NextSibling {
result.WriteString(getText(c))
}
return result.String()
}
crawler = func(node *html.Node) { crawler = func(node *html.Node) {
if node.Type == html.ElementNode && node.Data == params.subTag { if node.Type == html.ElementNode && node.Data == params.subTag {
if strings.Contains(node.FirstChild.Data, params.substring) { text := strings.TrimSpace(getText(node))
values = append(values, node.FirstChild.Data) if strings.Contains(text, params.substring) {
values = append(values, text)
} }
} }
for child := node.FirstChild; child != nil; child = child.NextSibling { for child := node.FirstChild; child != nil; child = child.NextSibling {
@ -252,11 +268,8 @@ func (s *MandarakeParser) findData(doc *html.Node, params price) []string {
} }
} }
crawler(doc) crawler(doc)
return values return values
} }
return nil
}
func (s *MandarakeParser) getPrice(rawStr []string) int { func (s *MandarakeParser) getPrice(rawStr []string) int {
re := regexp.MustCompile(`\([^)]*?([0-9,]+)[^)]*?\)`) re := regexp.MustCompile(`\([^)]*?([0-9,]+)[^)]*?\)`)