senaduka/old-svevijesti
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Fix scrapers

Browse Source
This commit is contained in:
Senad Uka
2022-02-22 05:18:39 +01:00
parent 4b863369b6
commit 0e9956a227
3 changed files with 2 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
internal/scraper/blijesak.go
View File

@@ -80,7 +80,7 @@ func setupBljesakArticlePageCrawler(crArticlePage *colly.Collector) {
text := ""
e.ForEach("div.col-xs-12, article.b-article-detail, div.col-article-content, div.intro, div.s-main-content", func(_ int, el *colly.HTMLElement) {
e.ForEach("div.intro, div#infiniteLoadBreakpoint", func(_ int, el *colly.HTMLElement) {
text += extractJustText(el.DOM)
})

2
internal/scraper/srpskainfo.go
View File

@@ -79,7 +79,7 @@ func setupSiArticlePageCrawler(crArticlePage *colly.Collector) {
text := ""
e.ForEach("div.article__top-content, div.article__content, h4, h3, h2, div.article__content", func(_ int, el *colly.HTMLElement) {
e.ForEach("div.article__top-content, div.article__content", func(_ int, el *colly.HTMLElement) {
text += extractJustText(el.DOM)
})

BIN
spider
View File

Binary file not shown.