misc-mirrors/article_scraper
1
0
Fork 0
mirror of https://gitlab.com/news-flash/article_scraper.git synced 2025-07-07 16:15:32 +02:00
Code Releases Activity

fmt & clippy

Browse source
This commit is contained in:
Jan Lukas Gernert 2023-03-19 23:37:42 +01:00
parent 280c516cbe
commit f7fa696921
2 changed files with 4 additions and 8 deletions
Download patch file Download diff file Expand all files Collapse all files

4
src/constants.rs
View file

@ -73,8 +73,6 @@ pub const UNLIKELY_ROLES: &[&str] = &[
pub const DEFAULT_TAGS_TO_SCORE: &[&str] = pub const DEFAULT_TAGS_TO_SCORE: &[&str] =
&["SECTION", "H2", "H3", "H4", "H5", "H6", "P", "TD", "PRE"]; &["SECTION", "H2", "H3", "H4", "H5", "H6", "P", "TD", "PRE"];
pub const DEPRECATED_SIZE_ATTRIBUTE_ELEMS: Lazy<HashSet<&str>> =
Lazy::new(|| HashSet::from(["TABLE", "TH", "TD", "HR", "PRE"]));
pub const PRESENTATIONAL_ATTRIBUTES: &[&str] = &[ pub const PRESENTATIONAL_ATTRIBUTES: &[&str] = &[
"align", "align",
"background", "background",
@ -89,6 +87,8 @@ pub const PRESENTATIONAL_ATTRIBUTES: &[&str] = &[
"valign", "valign",
"vspace", "vspace",
]; ];
pub static DEPRECATED_SIZE_ATTRIBUTE_ELEMS: Lazy<HashSet<&str>> =
Lazy::new(|| HashSet::from(["TABLE", "TH", "TD", "HR", "PRE"]));
pub static DIV_TO_P_ELEMS: Lazy<HashSet<&str>> = Lazy::new(|| { pub static DIV_TO_P_ELEMS: Lazy<HashSet<&str>> = Lazy::new(|| {
HashSet::from([ HashSet::from([
"BLOCKQUOTE", "BLOCKQUOTE",

8
src/full_text_parser/mod.rs
View file

@ -855,9 +855,7 @@ impl FullTextParser {
Ok(()) Ok(())
} }
pub(crate) fn post_process_document( pub(crate) fn post_process_document(document: &Document) -> Result<(), FullTextParserError> {
document: &Document
) -> Result<(), FullTextParserError> {
if let Some(mut root) = document.get_root_element() { if let Some(mut root) = document.get_root_element() {
Self::simplify_nested_elements(&mut root)?; Self::simplify_nested_elements(&mut root)?;
@ -869,9 +867,7 @@ impl FullTextParser {
Ok(()) Ok(())
} }
pub(crate) fn post_process_page( pub(crate) fn post_process_page(node: &mut Node) -> Result<(), FullTextParserError> {
node: &mut Node,
) -> Result<(), FullTextParserError> {
Util::clean_conditionally(node, "fieldset"); Util::clean_conditionally(node, "fieldset");
Util::clean_conditionally(node, "table"); Util::clean_conditionally(node, "table");
Util::clean_conditionally(node, "ul"); Util::clean_conditionally(node, "ul");