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

FIx bljesak crawler

Browse Source
This commit is contained in:
Senad Uka
2022-02-17 07:16:35 +01:00
parent a748bcb303
commit f4ea31373d
2 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 := "" text := ""
e.ForEach("div.col-article-content, div.intro, div.s-main-content", func(_ int, el *colly.HTMLElement) { e.ForEach("article.b-article-detail, div#infiniteLoadBreakpoint, div.col-article-content, div.intro, div.s-main-content", func(_ int, el *colly.HTMLElement) {
text += extractJustText(el.DOM) text += extractJustText(el.DOM)
}) })

2
internal/scraper/srpskainfo.go
View File

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