kemitix/paperoni
Archived
1
0
Fork 0
Code Issues 4 Pull requests 3 Projects Releases Packages Wiki Activity

Minor fixes in moz_readability

Browse source 
- swap unwrap for if let statement in `get_article_metadata`
- add default when extracting the title from a possible `<title>` element
- fix extracting alternative titles from h1 tags
This commit is contained in:
Kenneth Gitere 2021-04-21 19:14:25 +03:00
parent dbac7c3b69
commit 960f114dc6
1 changed files with 4 additions and 5 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

9
src/moz_readability/mod.rs
View file

@ -429,8 +429,7 @@ impl Readability {
let mut matches = None; let mut matches = None;
if let Some(property) = node_attr.get("property") { if let Some(property) = node_attr.get("property") {
matches = regexes::PROPERTY_REGEX.captures(property); matches = regexes::PROPERTY_REGEX.captures(property);
if matches.is_some() { if let Some(captures) = &matches {
let captures = matches.as_ref().unwrap();
for capture in captures.iter() { for capture in captures.iter() {
let mut name = capture.unwrap().as_str().to_lowercase(); let mut name = capture.unwrap().as_str().to_lowercase();
name = regexes::REPLACE_WHITESPACE_REGEX name = regexes::REPLACE_WHITESPACE_REGEX
@ -564,7 +563,7 @@ impl Readability {
.root_node .root_node
.select_first("title") .select_first("title")
.map(|title| title.text_contents().trim().to_string()) .map(|title| title.text_contents().trim().to_string())
.expect("This file has no <title> tag to extract a title from"); .unwrap_or("".to_string());
let orig_title = cur_title.clone(); let orig_title = cur_title.clone();
let mut title_had_hierarchical_separators = false; let mut title_had_hierarchical_separators = false;
let word_count = |s: &str| -> usize { s.split_whitespace().count() }; let word_count = |s: &str| -> usize { s.split_whitespace().count() };
@ -598,8 +597,8 @@ impl Readability {
} }
} else if cur_title.len() > 150 || cur_title.len() < 15 { } else if cur_title.len() > 150 || cur_title.len() < 15 {
let mut h1_nodes = self.root_node.select("h1").unwrap(); let mut h1_nodes = self.root_node.select("h1").unwrap();
let (_, h1_count) = h1_nodes.size_hint(); let h1_count = self.root_node.select("h1").unwrap().count();
if Some(1) == h1_count { if h1_count == 1 {
cur_title = Self::get_inner_text(h1_nodes.next().unwrap().as_node(), None); cur_title = Self::get_inner_text(h1_nodes.next().unwrap().as_node(), None);
} }
} }