This commit is contained in:
2023-12-18 16:51:47 +01:00
commit 88741b2303
36 changed files with 1490 additions and 0 deletions

View File

@@ -0,0 +1,26 @@
package scraper
import (
"github.com/PuerkitoBio/goquery"
)
func extractJustText(el *goquery.Selection) string {
textPart := ""
htmlPart, _ := el.Html()
if len(el.Nodes) == 0 {
return ""
}
//fmt.Println("Checking: ", htmlPart, "Duzina: ", strconv.Itoa(len(el.Nodes)), " Type je ", el.Nodes[0].Type, " jednakost ", el.Text() == htmlPart)
if el.Text() == htmlPart {
return el.Text() + "\n"
}
el.Children().Each(func(_ int, el2 *goquery.Selection) {
if el2.Is("div, p, span, a") {
textPart += extractJustText(el2)
}
})
return textPart
}