misc-mirrors/article_scraper
1
0
Fork 0
mirror of https://gitlab.com/news-flash/article_scraper.git synced 2025-07-09 17:05:30 +02:00
Code Releases Activity

tmp: dont strip scripts

Browse source
This commit is contained in:
Jan Lukas Gernert 2020-01-27 16:36:32 +01:00
parent 23514aff9e
commit 98348b7e59
1 changed files with 20 additions and 22 deletions
Download patch file Download diff file Expand all files Collapse all files

42
src/lib.rs
View file

@ -121,10 +121,10 @@ impl ArticleScraper {
return Err(error); return Err(error);
} }
if let Err(error) = ArticleScraper::eliminate_noscrip_tag(&context) { // if let Err(error) = ArticleScraper::eliminate_noscript_tag(&context) {
error!("Eliminating <noscript> tag failed - '{}'", error); // error!("Eliminating <noscript> tag failed - {}", error);
return Err(error); // return Err(error)
} // }
if download_images { if download_images {
if let Err(error) = self if let Err(error) = self
@ -646,7 +646,7 @@ impl ArticleScraper {
); );
// strip all scripts // strip all scripts
let _ = ArticleScraper::strip_node(&context, &String::from("//script")); //let _ = ArticleScraper::strip_node(&context, &String::from("//script"));
// strip all comments // strip all comments
let _ = ArticleScraper::strip_node(&context, &String::from("//comment()")); let _ = ArticleScraper::strip_node(&context, &String::from("//comment()"));
@ -782,23 +782,21 @@ impl ArticleScraper {
Ok(()) Ok(())
} }
fn eliminate_noscrip_tag(context: &Context) -> Result<(), ScraperError> { // fn eliminate_noscript_tag(context: &Context) -> Result<(), ScraperError> {
let xpath = "//noscript"; // let xpath = "//noscript";
let node_vec = Self::evaluate_xpath(context, xpath, false)?; // let node_vec = Self::evaluate_xpath(context, xpath, false)?;
// for mut node in node_vec {
for mut node in node_vec { // if let Some(mut parent) = node.get_parent() {
if let Some(mut parent) = node.get_parent() { // node.unlink();
node.unlink(); // let children = node.get_child_nodes();
let children = node.get_child_nodes(); // for mut child in children {
for mut child in children { // child.unlink();
child.unlink(); // let _ = parent.add_child(&mut child);
let _ = parent.add_child(&mut child); // }
} // }
} // }
} // Ok(())
// }
Ok(())
}
} }
#[cfg(test)] #[cfg(test)]